William Lawrence Neuman - Understanding Research - Books A La Carte (2nd Edition) (2016, Pearson Education)

Understanding Research Second Edition

W. Lawrence Neuman University of Wisconsin at Whitewater

330 Hudson Street, NY, NY 10013

VP, Product Development: Dickson Musslewhite Director, Content Strategy and Development: Sharon Geary Editor in Chief: Ashley Dodge Managing Editor: Sutapa Mukherjee Sponsoring Editor: Tanimaa Mehra Content Manager: Carly Czech Editorial Project Manager: Lumina Datamatics, Inc. Development Editor: Angela Mallowes, iEnergizer Aptara®, Ltd. Instructional Designer: Monica Agarwal, iEnergizer Aptara®, Ltd. Asset Development Team: LearningMate Solutions, Ltd. VP, Director of Marketing: Maggie Moylan Director, Project Management Services: Etain O’Dea

Project Team Lead: Vamanan Namboodiri Project Manager: Pooja Aggarwal Director of Field Marketing: Jonathan Cottrell Senior Marketing Coordinator: Susan Osterlitz Operations Manager: Mary Fischer Operations Specialist: Mary Ann Gloriande Associate Director of Design: Blair Brown Interior Design: Kathryn Foot Cover Design: Lumina Datamatics, Inc. Cover Art: StudioSmart/Shutterstock Digital Studio Project Manager: Elissa Senra-Sargent Digital Studio Team Lead: Peggy Bliss Full-Service Project Management and Composition: Megha Bhardwaj, iEnergizer Aptara®, Ltd. Printer/Binder: Courier Kendallville Cover Printer: Phoenix Color

Acknowledgements of third party content appear on page 292–293, which constitutes an extension of this copyright page. Copyright © 2017, 2009 by Pearson Education, Inc., or its affiliates. All Rights Reserved. Printed in the United States of America. This publication is protected by copyright, and permission should be obtained from the publisher prior to any prohibited reproduction, storage in a retrieval system, or transmission in anyform or by any means, electronic, mechanical, photocopying, recording, or otherwise. For information regarding permissions, request forms and the appropriate contacts within the Pearson Education Global Rights & Permissions department, please visit www.pearsoned.com/permissions/. PEARSON, ALWAYS LEARNING, and REVEL are exclusive trademarks owned by Pearson Education, Inc., or its affiliates, in the U.S., and/or other countries. Unless otherwise indicated herein, any third-party trademarks that may appear in this work are the property of their respective owners and any references to third-party trademarks, logos or other trade dress are for demonstrative or descriptive purposes only. Such references are not intended to imply any sponsorship, endorsem*nt, authorization, or promotion of Pearson’s products by the owners of such marks, or any relationship between the owner and Pearson Education, Inc., or its affiliates, authors, licensees or distributors. Library of Congress Cataloging-in-Publication Data Names: Neuman, W. Lawrence (William Lawrence), 1950- author. Title: Understanding research/W. Lawrence Neuman, University of Wisconsin at Whitewater. Description: Second edition. | Boston : Pearson, [2017] | Includes bibliographical references and index. Identifiers: LCCN 2016015419| ISBN 9780205910380 (alk. paper) | ISBN 0205910386 (alk. paper) Subjects: LCSH: Social sciences–Research. | Social sciences—Methodology. | Sociology—Research. | Sociology—Methodology. Classification: LCC H62 .N3882 2017 | DDC 300.72—dc23 LC record available at https://lccn.loc.gov/2016015419

10 9 8 7 6 5 4 3 2 1 Books a la Carte ISBN-13: 978-0-20-591038-0 ISBN-10: 0-20-591038-6

Brief Contents 1

Why Do Research?

2 Planning a Study

Becoming an Ethical Researcher

Sampling

Measuring Social Life

The Survey

116

The Experiment

138

Research with Nonreactive Measures

160

Making Sense of the Numbers

182

10 Observing People in Natural Settings

11 Looking at the Past and across Cultures

12 Writing a Research Report

210 236 261

iii

Contents Preface ix

1 Why Do Research?

1.1: On What Basis Do We Make Decisions? 1.1.1: Alternatives to Research 1.1.2: Frustrations and Misunderstandings about Research

1.2: What Is Empirical Social Research and Why Is It Respected?

Learning from History: Research and Religion

1.2.1: What Are Critical Thinking Skills? Summary Review: Critical Thinking Skills 1.2.2: What Counts as Evidence? 1.2.3: Research Is a Process That Results in a Product 1.3: What Are the Types of Social Research? 1.3.1: Four Quantitative Data Collection Techniques 1.3.2: Two Qualitative Data Collection Techniques

3 4

8 9 9 10 10 10 11

Summary Review: Quantitative and Qualitative Data Collection Techniques

1.4: What Is the Purpose of the Research and How Is It Used? 1.4.1: What Is the Purpose of a Study?

11 11

Example Study: Exploratory Research

Example Study: Descriptive Research

Example Study: Explanatory Research

Example Study: Evaluation Study

1.4.2: How Will We Use the Research?

Summary Review: Purposes of Research

Example Study: Applied Research

1.5: What Are the Steps in the Research Process?

Summary: What You Learned about Doing Research

Quick Review

2 Planning a Study

2.1: How Do We Select a Topic to Study?

2.2: The What, Why, and How of a Literature Review 23 2.2.1: Where Do You Find the Research Literature? 25 2.2.2: Scholarly Journals 25

2.3.1: Refine the Topic 2.3.2: Design Your Search 2.3.3: Locate the Research Reports Example Study: Sexual Harassment Literature Search 2.3.4: Read and Take Notes on the Reports Found 2.3.5: Organize Notes, Synthesize, and Write the Review 2.3.6: Create the Reference List

2.5: How to Design a Study for a Research Proposal 2.5.1: When and How Do You Focus the Research Question? 2.5.2: To What Universe Can You Generalize from a Study’s Findings? 2.5.3: Will You Follow a Linear or Nonlinear Path When Doing Research? 2.5.4: Do You Examine Variables and Hypotheses or Cases and Contexts? 2.5.5: How Will You Analyze Patterns in the Data That You Gather? Summary Review: Quantitative versus Qualitative Research 2.5.6: What Type of Explanation Will Give Meaning to the Patterns in the Data? 2.5.7: What Are the Units of Analysis in Your Study? Learning from History: Night-Lights and Spuriousness Summary: What You Learned about Planning a Study

38 38 39 39 40 43 44 44 45 47 47

3 Becoming an Ethical Researcher

3.1: What Is the Ethical Imperative? 3.1.1: Scientific Misconduct

51 52

Example Study: Scientific Misconduct and the Miracle Study

3.1.2: Unethical but Legal

2.2.3: Sources Other Than Scholarly Journals

28 29

Learning from History: Nazi Doctors

34 35

Quick Review

Summary Review: Different Types of Periodicals

2.3: Six-Step Literature Review Process

37 2.4: How Do We Focus the Research Question? 2.4.1: Narrowing a Topic into a Research Question 37 2.4.2: Comparing Good and Not-So-Good 38 Research Questions

3.2: The Ethical Issues Involved in Using People as Research Participants 3.2.1: The Origin of Ethical Principles for Research with Humans

Example Study: Are U.S. Colleges the World’s Best?

29 30 30

3.2.2: Protecting Research Participants from Harm

52 53 53 54 54

Contents v

Example Study: Zimbardo Prison Study

Example Study: Milgram Obedience Study

Example Study: Tearoom Trade

3.2.3: Participation Must Be Voluntary and Informed Learning from History: Unethical Research by Facebook 3.2.4: Limits to Using Deception in Research 3.2.5: Privacy, Anonymity, and Confidentiality Learning from History: Not Breaking the Confidence Guarantee

57 58 58 59 61

3.2.6: Extra Protections for Special Populations 61 3.2.7: Formal Protections for Research Participants 62 Ethics Code of the American Association for Public Opinion Research

Summary Review: Basic Principles of Ethical Research

3.3: How Do Sponsors Affect Research Ethics?

Tips for the Wise Consumer: Who Paid for a Study

3.3.1: Arriving at Particular Findings 3.3.2: Limits on How to Conduct Studies 3.3.3: Suppressing Findings

65 65 65

Tips for the Wise Researcher: Drawing Systematic Samples 4.5.2: The Stratified Sample

81 82

Example Study: Stratified Sample

Learning from History: General Social Survey Oversample

4.5.3: A Cluster Sample

Example Study: Cluster Sample

4.6: Sampling in Difficult or Specialized Situations 4.6.1: Random-Digit Dialing 4.6.2: Within-Household Sampling 4.6.3: Sampling Hidden Populations

85 85 86 86

Example Study: Hidden Populations

Summary Review: Types of Samples

4.7: Making Inferences from a Sample to Population 4.7.1: Sampling Errors 4.7.2: Sample Size 4.7.3: Confidence Intervals

87 87 88 88

Summary: What You Learned about Sampling

Quick Review

5 Measuring Social Life

Learning from History: Why So Little Research on Gun Violence or Medical Marijuana?

5.1: Why Measure? 5.1.1: Who Is Poor?

93 93

3.4: Politics and Social Research

Learning from History: Who Is Poor

Example Study: Political Influence on Crime Research 67

5.2: How Do We Make the Social World Visible?

3.5: “Value-Free” and “Objective” Research

5.3: Do We Measure with Numbers or Words? 5.3.1: Two Parts of the Measurement Process 5.3.2: Conceptualization and Operationalization

95 95 96

Learning from History: Measuring Social Distance

Tips for the Wise Consumer: Creating a Good Measure

Summary: What You Learned about Becoming an Ethical Researcher?

Quick Review

4 Sampling

67 68

4.1: Why Do We Use Samples?

4.2: Types and Applications of Nonrandom Samples 4.2.1: Convenience Sampling 4.2.2: Quota Sampling 4.2.3: Purposive Sampling

71 71 72 73

Example Study: Purposive Sample

4.2.4: Snowball Sampling

5.3.3: Quantitative Conceptualization and Operationalization 5.3.4: Qualitative Conceptualization and Operationalization Example Study: Operationalizing Social Ties

97 98 98

Summary Review: Steps in Quantitative and Qualitative Conceptualization and Operationalization 99

Example Study: Snowball Sample

4.3: The Terminology Used to Discuss Random Sampling

5.4: How Can We Create Good Measures? 5.4.1: Relationship of Reliability and Validity 5.4.2: Three Types of Measurement Validity

100 101 102

Summary Review: Measurement Validity Types

102

Learning from History: The Famous Literary Digest Mistake

Example Study: Content Validity and the Contact Hypothesis

102

4.4: Producing a Simple Random Sample 4.4.1: The Sampling Distribution

76 77

Summary Review: Twelve Terms in Random Sampling

5.5: What Are the Principles of Qualitative Measurement?

103

5.6: What Are the Principles of Quantitative Measurement?

103

4.5: Types and Uses of Random Samples 4.5.1: Systematic Sample

80 80

Summary Review: Characteristics of the Four Levels of Measurement

104

vi Contents 5.7: How Do We Construct and Use Indexes and Scales? 5.7.1: Mutually Exclusive, Exhaustive, and Unidimensional 5.7.2: Index Construction Learning from History: Index of Dissimilarity 5.7.3: Two Index Construction Issues 5.7.4: Scale Construction 5.7.5: Commonly Used Scales

105 106 106 107 108 108 108

Example Study: Social Distance and Immigration Views

110

Example Study: Semantic Differential and Wine

110

Example Study: Guttman Scaling and Neighborhood Preference

112

Summary Review: Four Major Scales

113

Summary: What You Learned about Measuring Social Life

6.6.2: The Interview Stages 6.6.3: Training Interviewers 6.6.4: Using Probes in Interviews 6.6.5: Interviewer Bias Example Study: Interviewer Race Effects Are Subtle and Pervasive 6.6.6: Computer-Assisted Telephone Interviewing 6.7: How Can You Be Ethical in Survey Research? Summary: What You Learned about Surveys

113

Quick Review

113

6 The Survey

116

6.1: What Is a Social Survey? 6.1.1: How Does an Opinion Poll Differ from a Social Survey?

117

Example Study: Views on Same-Sex Marriage

118

117

6.1.2: Survey Data and Cause-Effect Explanations 119

Quick Review

7 The Experiment

133 134 134 135 135 135 136 137

137

138

7.1: When Are Experiments Most Useful? 7.1.1: Questions You Can Answer with the Experimental Method 7.1.2: Limitations of Experimental Research

139

7.2: Why Assign People Randomly? 7.2.1: Random Assignment and Random Sampling

140

140 140

141

7.3: How Do We Use Variables in an Experiment?

142

Example Study: Was It a Gun or a Tool?

142

7.3.1: Planning an Experiment

143

Summary Review: Steps in Conducting an Experiment

143

6.2: How Do We Conduct a Survey? 6.2.1: Start-Up Stage 6.2.2: Implementation Stage 6.2.3: Data Analysis Stage

120 120 121 121

7.4: How Do We Combine Parts into Experimental Designs? 7.4.1: Types of Experimental Design 7.4.2: True Experimental Designs

144 144 144

6.3: Writing Good Survey Questions 6.3.1: What Are Leading Questions?

121 123

Example Study: Communal Values and Assumed Similarity

147

Summary Review: Survey Question Writing Pitfalls

123

6.3.2: Getting Answers to Survey Questions

124

7.4.3: Pre-Experimental Designs 7.4.4: Quasi-Experimental Designs

148 149

Summary Review: Open- and Closed-Ended Response Formats

125

Summary Review: A Comparison of Experimental Designs

Example Study: Questionnaire Items from the 2009 Pew Research Center Survey

126

Example Study: Anti-Marijuana Television Ads

151

Learning from History: The Power of Words

128

7.5: What Is Experimental Validity and Why Is It Important? 7.5.1: Internal Validity 7.5.2: External Validity

152 152 154

Summary Review: Threats to Internal and External Validity

154

7.4.5: Design Notation

6.4: How Can You Design an Effective Questionnaire? 6.4.1: Length of Survey or Questionnaire 6.4.2: Question Sequence

129 129 129

Example Study: Question Order Effects

130

6.5: The Advantages and Disadvantages of Different Survey Formats 6.5.1: Mail and Self-Administered Questionnaires 6.5.2: Telephone Interviews 6.5.3: Face-to-Face Interviews 6.5.4: Web Surveys 6.6: How Do You Interview in Survey Research? 6.6.1: The Interviewer’s Role

131

7.5.3: Field Experiments Learning from History: The Hawthorne Effect 7.5.4: Natural Experiments

150 151

154 154 155

131 132 132 132

Example Study: Labeling Fast Food Menus and Calories Consumed

155

7.6: How Can We Learn from Making Comparisons of Experimental Results?

156

132 133

Tips for the Wise Consumer: Examining How an Experiment Was Conducted

157

Contents vii

7.7: How Can You Be Ethical in Conducting Experiments? Summary: What You Learned about the Experiment

Quick Review

8 Research with Nonreactive Measures

8.1: What Makes a Study Nonreactive?

157

8.3: How Can You Find Content within Communication Messages? 8.3.1: How to Measure and Code in Content Analysis? Example Study: Content Analysis and Opposition to Aid for the Poor

184 186

157

9.2: How Do You Describe Quantitative Results?

186

158

9.3: Univariate Statistics 9.3.1: Where Is the Middle?

186 187

Summary Review: Three Measures of Central Tendency

188

160 161

8.2: How Can You Use Physical Evidence in Research? 161 8.2.1: Limitations of Physical Evidence 162 Example Study: Finding Data in a Graveyard

9.1.2: Prepare a Codebook 9.1.3: Clean the Data

162 163 163

9.3.2: What Is the Spread? 189 9.3.3: Three Common Ways to Measure Variation 189 Learning from History: Teen Birthrate Decline 9.3.4: Z-scores 9.3.5: Alternative Ways to Display Information on One Variable

191 191 192

9.4: Bivariate Statistics 9.4.1: Scattergrams

194 194

Example Study: Cohabitation and Gender Equality

194

9.4.2: Bivariate Tables 9.4.3: Measures of Association 9.4.4: Results with More than Two Variables

197 200 201

166

Tips for the Wise Consumer: Noticing Statistical Significance

202

167 169

Example Study: Is the Cohabitation–Divorce Relationship Spurious

203

8.4: How Can You Mine Existing Statistical Sources to Answer New Questions? 8.4.1: Locating Data

170 172

Example Study: Existing Statistics: Becoming a “Wet” or “Dry” County

9.5: How Do You Go beyond Description to Inference? 204 9.5.1: Statistical Significance 204 9.5.2: Levels of Significance 204

173

8.3.2: Coding, Validity, and Reliability 8.3.3: Content Analysis with Visual Material Example Study: Magazine Covers and Cultural Messages 8.3.4: Content Analysis Research Steps 8.3.5: Limitations of Content Analysis

164 165 165

9.4.5: Multiple Regression Analysis

8.4.2: Verifying Data Quality

173

Tips for the Wise Consumer: Using Data from Existing Statistical Sources

176

8.4.3: Thinking Creatively about Variables of Interest 8.4.4: Standardizing the Data 8.5: Using Secondary Sources to Answer Research Questions? 8.5.1: Limitations of Secondary Data Sources Example Study: Secondary Data and the Immigration-Welfare Question Summary Review: Review Strengths and Limitations of Nonreactive Research 8.6: How Do You Conduct Ethical Nonreactive Research? Summary: What You Learned about Research with Nonreactive Measures

Quick Review

9 Making Sense of the Numbers

176 176

Summary: What You Learned about Making Sense of the Numbers

Quick Review

10 Observing People

in Natural Settings

203

207

210

10.1: What Is Field Research? 10.1.1: Ethnography

211 211

Example Study: Ethnographic Inference in a Hotel

213

10.2: How Do You Study People in the Field?

214

179

10.3: How Do You Begin to Conduct Field Research?

214

179

Example Study: Two Field Research Studies of Work in Nursing Homes

214

177 178

180 180

180

182

9.1: How Do You Prepare Data for Statistical Analysis? 183 9.1.1: Organize the Raw Data into a Machine-Readable Format 183

10.3.1: Preparing for a Field Study 10.3.2: Starting the Research Project

216 216

Example Study: Negotiating with Gatekeepers

217

10.4: What Should You Do in the Field Site?

218

Learning from History: Taxis, Customers, and Tips

218

10.4.1: Being in the Field 10.4.2: Strategies for Success in the Field 10.5: How Do You Collect Data in the Field? 10.5.1: Observing and Taking Field Notes

219 220 222 222

viii Contents Example Study: Noticing Details 10.5.2: Interviewing in Field Research

222 227

11.4.3: Galton’s Problem 11.4.4: Gathering Comparative Data

253 254

10.6: How Do You Conclude? 10.6.1: Leaving the Field 10.6.2: Writing the Field Research Report

229 229 230

Example Study: Cross-National Study on Work Schedules

255

Summary Review: Data in H-C Research

256

Tips for the Wise Consumer: Field Research Reports

230

11.4.5: The Issue of Equivalence

256

11.5: How Can You Be an Ethical H-C Researcher?

10.7: How Can You Be an Ethical Field Researcher?

230

10.8: What Are Focus Groups and How Do You Use Them? Summary: What You Learned about Observing People in Natural Settings

Quick Review

11 Looking at the Past and across Cultures

“The Transformation of America’s Penal Order” by Campbell and Schoenfeld

231

Quick Review

233

236 237

“The Transformation of Prison Regimes in Late Capitalist Societies” by Sutton

238

11.1: Historical-Comparative Research

238

Example Study: Ethnic Cleansing

239

11.1.1: How Are Field Research and H-C Research Alike? 11.1.2: What Is Unique about H-C Research?

239 240

Learning from History: Conditions in Medieval Western Europe

241

Summary Review: A Comparison of Approaches to Research

242

Summary Review: Features of a Distinct H-C Approach to Doing Research

242

11.2: How Do You Conduct a Historical-Comparative Research Study? 11.2.1: Acquire the Necessary Background 11.2.2: Conceptualize and Begin to Focus 11.2.3: Locate and Evaluate the Evidence 11.2.4: Organize the Evidence 11.2.5: Synthesize and Develop Concepts 11.2.6: Write the Report

243 243 243 243 244 244 244

11.3: How Do You Conduct Research on the Past? 11.3.1: Types of Historical Evidence

244 245

Example Study: Women of the Ku Klux Klan

248

11.4: How Do You Conduct Research That Compares Cultures? 11.4.1: Looking across Cultures to See a Wider Range Example Study: Abortion Politics in the United States and Germany 11.4.2: Can You Really Compare? Example Study: Immigrants in Two European Countries

Summary: What You Learned about Looking at the Past and across Cultures

250 250 250 251 252

Writing a Research Report

258 259

259

261

12.1: Why Write a Report?

262

12.2: How Do You Proceed with the Writing Process? 12.2.1: Know Your Audience 12.2.2: Pick a Style and Tone 12.2.3: Organize Your Thoughts

262 262 263 264

Tips for the Wise Researcher: Outlining

265

12.2.4: Go Back to the Library 12.2.5: Engage in Prewriting Activities 12.2.6: Rewrite Your Report

265 265 266

Summary Review: Steps in the Writing Process

267

12.3: How Do You Write about Cause-Effect Relations?

267

12.4: How Do You Write a Quantitative Research Report? 12.4.1: Abstract or Executive Summary 12.4.2: Presentation of the Problem 12.4.3: Description of the Method 12.4.4: Results and Tables or Charts 12.4.5: Discussion Section 12.4.6: Conclusion or Summary

268 268 269 269 269 269 270

12.5: How Do You Write a Qualitative Research Report? 12.5.1: Report on Field Research

270 271

Learning from History: Boys in White

271

12.5.2: Report on Historical-Comparative Research 12.6: How Do You Prepare a Research Proposal? 12.6.1: Proposals for Research Grants Summary: What You Learned about Writing a Research Report

273 273 273 277

Quick Review

277

Appendix A Sample Research Proposals

279

Appendix B Data and Literature Research

285

References

288

Credits292 Index

294

Preface To the Student Welcome to Understanding Research, 2nd Edition. Learning how to do social research can be fun, but many students believe they know little about the topic and are intimidated by it. You have already encountered the results of social research studies. They are in course materials, newspapers, Internet sources, and news programs. Most professional work settings, businesses, and agencies regularly use research results. While you may have encountered the results of research, this course looks behind the scene to examine the processes of doing research that produced the results. Here are four basic ideas about social research to keep in mind as you read this text: • Social research is a process that produces a product, namely research results. • The research process and its results have relevance for individuals, organizations, communities, and nations. • The research process grew out of the combined wisdom and experience of thousands of people across many decades. • You can master the fundamentals of doing research. Social research is a process or an ongoing activity that takes place over time. People do research; it does not just happen. They engage in a series of actions to produce the product, i.e., research results. As a process or activity, real people make decisions, take risks, engage in various steps, write things down, and think seriously. The principles, techniques, and stages of this process are outcome of thousands researchers who worked over many decades to iron out difficulties and seek the best ways to learn about the social world. The purpose of a course on research is to help you learn about doing research and understand the process by which we acquire knowledge about the social life around us. At times, the research process and some of its results seem obscure or esoteric, but most of the time studies have real consequences. They are relevant to daily life and to being a citizen, a friend, a parent, a professional, an employee, or a business owner. It is not always immediately apparent, but most research studies can have practical consequences for how we make decisions. Some people, out of ignorance, say it is “just research” or “only a study.” This occurs when they cannot see the connections of research to their lives and to the lives of people, organizations, and events around them. Many newcomers feel intimidated by the social research enterprise. Yes, college professors, high-powered

research scientists and others with years of advanced schooling and training conduct most research studies. This does not mean that research is beyond a beginning student. To conduct research study only requires an ability to think, to collect evidence, and to examine connections or implications. A beginning student may not grasp very complex, advanced results or be able to conduct a highly sophisticated research study, but grasping the basic principles, key procedures, and overall process is possible with an investment of modest amounts of time and effort. Once you grasp the fundamentals of the research process, it is a short distance to move to doing small-scale studies of your own. Understanding and doing research can open an entire world of studies, findings, and new insights.

To the Instructor Few students approach a course on research methods with excitement and positive expectations. It is often a required course in the curriculum and tends to generate unnecessary angst and anxiety among many students. Yet, learning how to do research does not have to be unpleasant, difficult, or stressful. Conducting a study can be fun and exciting. After all, by doing social research students explore and learn new things, probe into diverse areas of social life, and feel empowered by creating new knowledge. Conducting a study does require self-awareness, rigor, and discipline—but students acknowledge the need for self-awareness, rigor, and discipline for the interests about which they are motivated, such as athletic competition, a hobby such as video gaming, spectator sports, fashion or music, or a volunteer activity. My approach to teaching about social research comes from personal experience. I feel genuine joy when I see students learn—and specifically, watch them learn about processes of discovery and knowledge creation. Over the past three decades, I taught social research methods to undergraduates and graduate students, reflecting, adjusting, learning, and improving over that period. My goal has been to identify what students need to know and present it in a manner that they can easily grasp. This meant reaching to the fundamentals of social science research ideas and techniques, creating a transparent structure to organize material, and providing students with both everyday relevant examples and academic studies that build basic knowledge. My goal has been to make the essentials of doing high-quality research accessible to students in ways that they can become excited about the

x Preface research process. In short, I seek to distill the core principles, process, and procedures of research and present them in a manner that students will want to learn them. Many professions, applied fields, and academic disciplines use the findings and techniques of social science research. My own background has been as an eclectic and wide-ranging sociologist. I am committed to a broad, ecumenical approach to social scientific inquiry. The scientificresearch community has produced diverse approaches and techniques for conducting social scientific research. I believe it is a serious error to fixate on a single research approach or technique—be it the experiment or survey, quantitative methods in general, or qualitative ethnographic research. It is a serious error because it limits our understanding of a complex, changing social world, and because it misinforms students about the scope and promise of research. By being inclusive with regard to diverse forms that social science research can take, we gain much and lose little. To me, it is unwise to disengage the concrete and technical aspects of conducting research from the broader epistemological issues and the ethical-political dimensions of the social science enterprise. I view social research as an accomplishment by human actors that takes place in specific social-historical contexts. Removing human agency and context from how we think about the research process only introduces distortion and diminishes understanding. I believe applying the principles, process, and results of social research is consequential for the choices and decisions we make in our daily lives and in organizational settings. When we apply an open-ended understanding of the logic and results of social research, better choices and decisions in our organizations, communities, and lives frequently follow. Students can improve their lives and the life conditions of the people around them when they understand the research process. A corollary of this point is that the failure to understand research will likely condemn the next generation to fall behind and make many unwise decisions. Few students who learn about social research become full-time professional research scientists, but most will become parents, friends, colleagues, employees, citizen-voters, and community members. I believe having a sound understanding of the processes and principles of social research is likely to improve how they will fulfill those life-long roles.

New to the Edition The second edition of Understanding Research seeks to expand upon the core principles of this course: to make the essentials of doing high quality research accessible to students in ways that they can become excited about the research process. As you have come to expect with this course, core principles, processes, and procedures of

research are distilled and presented in a manner that students will want to learn them. It is a stress-free and enjoyable approach to the research methods course by providing salient real world examples throughout. • Each chapter opens with a vignette/case that has real world relevance and connections to the content material of the chapter. • Data and examples throughout the course represent what is happening right now in research methods with most examples drawn from studies published recently in academic journals. • The course offers new tools and methods for applying classic concepts like expanded strategies for approaching a literature review and ways to think critically and creatively about nonreactive research techniques. • In addition to offering guidance on the fundamentals of writing a research report, this edition offers a stronger emphasis on the seriousness of preparing a research report and the importance of communicating findings clearly and efficiently. • In addition to end of chapter summaries and quizzes, each chapter now has periodic reviews of major points and student self-tests for regular feedback.

REVELTM Educational technology designed for the way today’s students read, think, and learn When students are engaged deeply, they learn more effectively and perform better in their courses. This simple fact inspired the creation of REVEL: an immersive learning experience designed for the way today’s students read, think, and learn. Built in collaboration with educators and students nationwide, REVEL is the newest, fully digital way to deliver respected Pearson content. REVEL enlivens course content with media interactives and assessments—integrated directly within the authors’ narrative—that provide opportunities for students to read about and practice course material in tandem. This immersive educational technology boosts student engagement, which leads to better understanding of concepts and improved performance throughout the course. Learn more about REVEL: www.pearsonhighered.com/ revel

Features Students need to have both the cognitive and affective dimensions of the learning process addressed to learn and understand material fully. The pedagogical features in this

Preface xi

course guide a student’s travels through each chapter’s content, stimulate their interest, and enhance both their content learning and their engagement with the material. Designed to move beyond the primary cognitive objective of content mastery, the features enhance affective objectives as well. Presenting accurate content in a well-organized manner is insufficient. Students struggle to stay focused when they believe material to be irrelevant to their lives and the world around them, and if they feel overwhelmed by and insecure about the material. This makes stimulating student interest and actively engaging them essential to facilitate their learning. The pedagogical features work together to accomplish these tasks in four ways. Increasing students’ motivation to learn by demonstrating that the material has “real world” relevance and connections to other issues and knowledge areas. Students tend to engage with material that they see as being relevant, consequential, and interconnected. Each chapter opens with a research topic, drawn from a variety of fields. Some of these topics are fast food advertising directed at children (Chapter 1), people who are “rednecks” (Chapter 5), and occupations that require “emotional work” (Chapter 10). In addition, each chapter has “boxed features.” Some are case studies of published research on a topic that students may find of interest, or historical events related to the material. As students learn specific methods of doing research, they also see how the methods can reveal new insights about issues in the social world. Two other features, Making It Practical and Tips for the Wise Consumer, emphasize how students can apply the material in a chapter. Examples include Using Article Search Tools (Chapter 2), Improving Unclear Questions (Chapter 6), and Recommendations for Taking Field Notes (Chapter 10). Stimulating student interest in and engagement with the material by arousing their curiosity. Student interest and excitement tend to grow when they can see how material offers a pathway to discovering what was previously unknown or contains aspects of a puzzle or mystery that they can solve. The interactive hide/reveal features in the book’s electronic version use curiosity to increase engagement. Addressing student feelings of anxiety and insecurity about the material by promoting a sense of accomplishment. Periodical feedback that signals success can circumvent student feelings of uncertainty, disappointment, and defeat when facing material the student finds to be challenging and complex at first. Once a student recognizes that he or she is capable of learning complex material, positive feelings about self and about the material often develop. The student also gains the confidence needed to move onto higher-level material. As a student moves through each chapter, he or she will encounter periodic Summary Review tables of major points, and journaling opportunities for regular instructor feedback. In addition,

key terms have links to their definition when first introduced, and again when the term reappears and is used with material in subsequent chapters. This repetition reinforces learning as it builds student confidence. Sparking critical thinking by including unresolved dilemmas, moral-political dimensions of material, and the reasons why researchers use certain procedures. Interest often fades when a student encounters a research procedure without a rationale for its use or closedended material, i.e., material for which all issues have been resolved and uncertainty removed. Interest can grow when a student confronts areas of disagreement and debate, and is able to see the rationale for using particular procedures. In addition to providing key historical context, the Learning from History feature presents students with an opportunity to analyze and reflect upon past issues and compare them to more recent research and events. Each chapter concludes with a Shared Writing exercise, which allows the student to write open/ended reflections, opinions, and ideas about issues in the material that he or she can share with other students.

Chapter Content Each chapter of the text has a similar format and mixes the practical-applied aspects of research with the foundational principles and techniques of doing a study. After a brief opening study to stimulate interest, students learn about a specific aspect of the research process anchored with learning objectives for each module. Chapter 1 outlines the basics of what social research entails. It explains why a student will find it beneficial to understand the research process. There are updates to the explanations on the meaning and importance of critical thinking as well as on the idea of having “standards of evidence” for data in social science research. They see the steps in conducting a research study and learn about some of the purposes for doing a research study. In addition, new examples from the recent research literature are used to illustrate the variety of types of social research. In Chapter 2 readers learn about the process of moving from a broad topic of interest to specific study design issues, including how to conduct a literature review. The opening issue of tattoos is carried into the chapter. A new feature is the organization of topics in a question-answer format, including the practical design issues for conducting both qualitative and quantitative research. The discussion on variables and hypotheses shows readers how to move from having general ideas about a cause-effect relationship to drawing a diagram a causal explanation. Chapter 3 considers both traditional issues in research ethics and some of the social-political concerns of doing social research. The current edition adds an entry on the

xii Preface controversy a year ago surrounding Facebook mining user data, and how this event raised issues of consent and privacy. The chapter also asks readers to consider the role of a whistle-blower in research and pressures on researchers from funding organizations or government bodies. Chapter 4 begins with a study of teens in large urban areas of the United States, especially those with high concentrations of crime and poverty, carrying illegal guns or other weapons to raise the question of how a researcher might go about drawing a sample of such a group. Throughout the chapter, readers learn to distinguish representative from non-representative sampling, and the importance of different forms of random samples. The emphasis is on mastering fundamental sampling concepts, although sampling in specialized situations is also discussed. This includes cluster sampling, and the chapter uses a recent article on whether local beliefs about the law and law enforcement influence levels of violence in U.S. urban neighborhoods to illustrate how to conduct a cluster sample. Chapter 5 provides readers with both the general ideas of social science measurement and several specific illustrations of central idea that researchers make aspects of the social world visible and turn it into research data. Readers also learn how the processes of conceptualization and operationalization jointly make this possible. Most of the illustrative examples in the chapter are from recent studies of U.S. racial-ethnic relations. In addition to general discussions of reliability and validity, and a review of major scales or indexes, readers see how qualitative researchers measure. The 2012 study by Desmond on evictions of inner-city people in Milwaukee illustrates how he used various quantitative and qualitative measures. The example of quantitative measurement includes examples of the 2010 index of dissimilarity scores for U.S. cities (in which Milwaukee has the highest black–white segregation). Readers learn about the Guttman scaling pattern and then see an example by Xie and Zhou (2012) who measured whites’ preference for housing in a racially mixed neighborhood for several large U.S. cities. A 2009 study on white Anglo’s social distance from Latinos and its connection with their immigration views illustrates the Borgadus Social Distance scale. Chapter 6 is a condensed introduction to survey research. This edition opens with an example of a recent survey on same-sex marriage. The chapter uses other recent examples from the research literature as the readers learn about various survey issues, such as survey question wording, question order effect, and social desirability in survey interviewing. The chapter intentionally avoids treating the survey technique in isolation from related methodology issues. The discussion of the survey is connected to a discussion of correlational versus experimental research approaches, the general process of operationalizing variables, preparing data for analysis, and ethics involved in reporting survey results.

In Chapter 7 readers learn about the great power of experimental research for demonstrating causality as well as the many specifics of experimental design. The chapter makes clear the similarities and differences between random sampling from a large population and randomization in research participant selection into experimental groups. The chapter highlights the contrast between a highly controlled laboratory setting and the value of a natural and field experiment where control by the experimenter is difficult. Chapter 8 covers an array of non-reactive research techniques, including content analysis and using existing statistical sources. This edition includes a new emphasis that encourages readers to use their creativity and powers of observation and to consider how they might unobtrusively observe, document, and analyze data on an issue. As an illustration, the chapter has an example study that measured the relative “walkability” of urban areas through the careful observation and the documentation of specific physical features. At the same time, the text asks readers to consider issues of possibility violating privacy in such studies. Chapter 9 is a very elementary introduction to statistical ideas and techniques used in social research. It begins by explaining how to organize and manage quantitative data, and closes with a review of several common statistical tests and ways to interpret their results. The emphasis of the chapter is to provide readers with a conceptual understanding and to develop their quantitative reasoning skills, more than to have them engage in computation. The chapter illustrates several concepts or techniques using actual data or examples from recent studies on the issues of cohabitation and teen pregnancy. Chapter 10 introduces readers to ethnographic field research. The sequence of chapter topics follows the chronology of what a person does as he/she conducts a field research study. The chapter opens with a study on the concept of “emotion work.” Readers see this concept continued in other chapter examples. The goal is to show readers not only the specific techniques to use when conducting a field research study, but also the ways by which researchers can develop or elaborate upon a concept, or engage grounded theorizing, during the process of doing field research. In Chapter 11 readers learn the value of the historical comparative approach for answering the “big questions” as they used it to study and reveal a great deal about several topic through examples taken from the recent literature on incarnation rates, workplace flexibility, and acts of genocide. Since many student readers find this approach difficult, the chapter makes it accessible and manageable providing them with a step-by-step guide on to how one would conduct a historical-comparative study. Chapter 12 is an overview of how to write research reports and proposals. As in the first edition, readers learn many tips about the writing process, but also that writing is serious, time-consuming work. Readers see the differences

Preface xiii

in the writing tasks between a qualitative study report, a quantitative study report, and a research proposal. New to this edition is illustrating how to write a qualitative data study by showing it in an article that was introduced in the chapter on field research. Readers not only learn about the process of conducting the research, but also learn ways to write up findings in a formal report.

Acknowledgments

Available Instructor Resources

Nina Coppens University of Massachusetts

The following resources are available for instructors. These can be downloaded at http://www.pearsonhighered. com/irc. Login required. • PowerPoint—provides a core template of the content covered throughout the text. Can easily be added to customize for your classroom. • Instructor’s Manual—includes chapter summaries and outlines; learning objectives; key terms with definitions; online resources; suggested readings, class exercises and activities; and video resources. • Test Bank—includes additional questions beyond the REVEL in multiple choice and open-ended—short and essay response—formats. • MyTest—an electronic format of the Test Bank to customize in-class tests or quizzes. Visit: http://www. pearsonhighered.com/mytest.

Many thanks go out to the reviewers who gave thoughtful feedback on all or parts of this text: Ken Baker Gardner-Webb University Karen Benton Urbana University

Elizabeth Easter University of Kentucky Molly George University of California, Santa Barbara Phyllis Kuehn California State University, Fresno John Lewis University of Southern Mississippi Angus McCartney Troy University Andrew Supple University of North Carolina, Greensboro Annette Taylor University of San Diego I dedicate this text to Diane, for all her patience and support. W. Lawrence Neuman

This page intentionally left blank

Chapter 1

Why Do Research?

Learning Objectives 1.1 Describe the role of research as one of

thefoundations of decision making 1.2 Explain why empirical social research is more

than gathering information and the role of critical thinking in the research process

1.4 Examine how the purpose of research

depends upon the outcome that the researcher is trying to accomplish 1.5 List the seven steps in the research

process

1.3 Summarize quantitative and qualitative

data collection techniques Perhaps you have young children, or siblings or nieces/nephews, or maybe you will have children of your own in the future. Any parent will tell you that a child’s eating habits are a major concern. In the United States, a rapid increase in childhood obesity and diabetes over the past decade has become a serious public health issue. Advertising greatly influences children, and one-half of all advertising targeted at

children is for food. A study (Gantz et al., 2007) reported that children ages 2 through 7 see an average of 12 food ads per day on TV (30 hours per year), whereas those ages 8 to 12 see an average of 21 food ads per day (50 hours per year). Most food advertising is for high calorie, low nutrition snacks, candy, soda, or fast food. Only 4 percent of ads are for dairy products, 1 percent is for fruit juices, and no ads are for fresh

2 Chapter 1

Figure 1.1 Number of Televisions in Households as a Moderator of Taste Preferences Total preference scores may range from –1 (preferred the unbranded food in all comparisons) to 1 (preferred the McDonald’s branded food in all comparisons).

4+ Number of Televisions in Household

fruits or vegetables. This is why several nations ban television advertising that targets children. Norway, Quebec (Canada), and Sweden ban all advertisem*nts during children’s television programs, and 30 countries put limits on advertising. Like other companies, fast food chains use branding to attract children (or their parents) as customers. “Branding” is when a company attaches its name and logo to products or services and promotes them in ways to create a strong mental and emotional connection within potential consumers. An example of branding is McDonald’s, which has more than 31,000 restaurants serving 60 million people worldwide each day. Researchers (Robinson et al., 2007) wanted to learn whether McDonald’s branding affected the food choices of young children (ages 3 to 5). They placed two sets of food items in front of the children—one food item (milk, French fries, hamburger, chicken nuggets, and baby carrots) was in a McDonald’s wrapper, and the other was not. They asked, “Can you tell me which is from McDonald’s?” to make certain the children saw the difference. They next asked the children to take one bite and taste each food item, then tell the researchers whether the food was the same or one tasted better than the other did. In fact, the food was identical. They also asked parents about television viewing habits and fast food restaurant visits. Results (see Figures 1.1 and 1.2) showed that more children said that the item in the McDonald’s package tasted better for all five food items.

0.3

0.4 0.5 Preference Score

Figure 1.2 Frequency of Eating at McDonald’s as a Moderator of Taste Preferences Total preference scores may range from –1 (preferred the unbranded food in all comparisons) to 1 (preferred the McDonald’s branded food in all comparisons).

Eating from McDonald’s

2– 3 per Week

1 per Week

1– 2 per Month

Under 1 time per Month

0.1

0.2

0.3 0.4 Preference Score

0.5

0.6

Why Do Research? 3

Note that McDonald’s did not sell baby carrots at the time of this study. In addition, children whose parents had taken them to McDonald’s were most likely to prefer McDonald’s. The researchers concluded that by the age of five the children had internalized the McDonald’s brand as indicating superior food taste. Studies such as these help parents, citizens, and public officials understand how young children make food choices. We do not have to send children to a research laboratory to study food choices. At times we can study changes in a local setting. Studies (Bernhardt et al., 2013) found fast food advertising that targets at children emphasizes giveaways, games, and fun to entice the children. In 2010, Santa Clara County, California, banned restaurants from giving away toys with fast food meals if the meals failed to meet certain nutritional standards. Before the ban, only five of 120 children’s meal combinations (4 percent) at fast food restaurants met the standards. Under the ban, a restaurant could continue to give away toys as a sales promotion, but only for healthy meals that met the standards. A team of researchers (Otten et al., 2012) compared pre-ban versus post-ban meal sales at four restaurants within the ban area and at four restaurants several miles outside of it. They found dramatic changes in the ban area, with a two- to three-fold increase in the sales of healthy meals over preban period, but no change in healthy meal sales outside the ban area. Other communities have required fast food restaurants to include calorie and nutritional information on menus, but studies found that changes to menu labels alone do not increase healthy eating.1 Researchers conduct studies on topics such as children’s food choices or fast food sales to understand what is actually occurring in the social world. We can use the study’s findings to make decisions that affect our own lives, our families, and our communities. Once you learn how researchers conduct studies, you will be in a better position to make decisions and choices.

for personal and family life, the community, or a career. Most of us make decisions using a mix of common sense; advice from experts, friends, and family; past experiences; cultural literacy; and school knowledge. Some of us also use religious faith, personal prejudice and values, horoscopes and lucky numbers, guesswork, or folklore. Few of us turn to research, so you might ask, how will a book on social research help me make decisions? Social research does not offer all the answers, but it can help you do the following: • Make better decisions about your daily life (What type of person should I marry to reduce the odds a divorce? How should I raise my child to be healthy and happy? Which college major will help me advance in a career and earn a good income 10 years from now?). • Understand events in the larger world around you (Why are many people getting divorced? Do racially integrated schools really reduce inter-racial tension? What community development plans will reduce traffic congestion and improve the environment in the future?). • Decide professional issues (Which product is likely to sell the most? How do I find out whether my employees are happy? Which children’s reading program is most effective?). You undoubtedly have heard of people doing research, or you may have read research findings or done some research. We use the word “research” in multiple ways. It is a process in which we apply specific principles and techniques to create high-quality knowledge; it is a body of knowledge (i.e., information, ideas, or theories) built up over time, and it is an orientation or frame of mind to apply when looking at and thinking about information, questions, or issues. What do you think of when you hear the word “research”? Compare Your Thoughts

Table 1.1 Responses to the Question: What Is Research?

1.1: On What Basis Do WeMake Decisions? 1.1 Describe the role of research as one of the foundations of decision making Each day you make thousands of choices and decisions. Most are trivial, such as what to have for breakfast; whether to text a friend; whether to purchase pink, white, or blue facial tissue; and which TV show to watch. Others are important, such as monitoring your child’s TV watching and eating habits. Many decisions have real consequences 1

On fast food restaurant labels and healthy eating, see Elbel, Gyamfi, and Kersh (2011), Stutts et al. (2011), and Tandon et al. (2011).

Response to question

Evaluation of response

It is fun and exciting.

Usually, research allows you to exercise your creativity and discover new things that are almost always fun.

It is difficult and mysterious.

Often, research has some difficult parts that may seem mysterious until you learn about them, but most aspects of research are easy to grasp.

It is practical and relevant.

Usually, it yields results you can use to make a difference in real life.

It is valuable and rewarding.

Most of the time. Properly conducted research can have a big payoff for life decisions you make, workplace effectiveness, and job prospects.

It is a waste of time and effort.

Not exactly. Research can take time and effort, but it is rarely a waste if done properly.

It is always correct.

No, it is not always right. Nonetheless, it is more likely to be right than the alternatives, such as relying on tradition, authority, or your personal experience. It is also useful to distinguish between better and worse research.

4 Chapter 1

1.1.1: Alternatives to Research We make most everyday decisions without doing research, checking on research findings, or using a research orientation. Most of the time, this works just fine, especially for trivial decisions. Unfortunately, studies suggest that few of us are great decision makers. Often we rely on distorted thinking and are unaware of misjudgments or bias. This is where a reliance on research helps; it reduces misjudgment, bias, and distorted thinking. Even though research produces valuable information and expands understanding, it is not 100 percent foolproof. It does not guarantee perfect results every time or yield “absolute truth.” Don’t be discouraged. In a head-to-head comparison with all the other ways we reach decisions, research wins easily. This is why professional organizations, well-educated people, and most leaders rely on research when they make important decisions. Centuries ago, people went to oracles, looked at the leaves at bottom of a teacup, or consulted the stars to make major decisions. Today, people in all fields—medicine, business, education, law enforcement, public policy—look to research publications or study research findings for guidance. Relying on social research for decisions is not always simple. You may have heard the dozens of contradictory and confusing research-based recommendations about health and diet in the mass media. What is so great about research if there is so much disagreement? Compare Your Thoughts A lot of what fills the mass media using the terms research or scientific does not actually involve scientific research. Unfortunately, people use the word “research” in the media when no real research backs a statement. Some of what you hear may be research backed, but it could be selective or incomplete, overstated, or distorted. The media “noise

Do you use “performance-enhancing” drinks, footwear, supplements or use special devices clothing and devices like wrist bands and compression stockings that guarantee health or improved sports performance?

machine” jumbles together many different types of statements. It is little wonder that many people are skeptical of research. Media distortion of research or social issues creates confusion. You may hear of a terrible social problem in the mass media, but closer inspection may reveal that it was seriously distorted. Real knowledge grows from the outcomes of many well-conducted research studies, not a single isolated study. A study by Heneghan et al. (2012) examined the advertising claims of 100 such products of this multi-billion dollar industry in magazines and on the web. The authors viewed 1,035 web pages and 615 advertisem*nts for such products in 110 magazines in the United States and the United Kingdom in March 2012. When the advertisem*nts provided no solid evidence for claims, study authors contacted the companies requesting any research to back the claims. They concluded, “There is a striking lack of evidence to support the vast majority of sportsrelated products that make claims related to enhanced performance or recovery, including drinks, supplements and footwear.” They found no evidence at all for over onehalf the products, and of those with evidence, for one-half it was too vague and uninformative to allow an evaluation. This left 74 studies to back product claims. The authors concluded that of the 74 studies that did contain sufficient information to allow an appraisal, only 3 percent to 4 percent, were quality, serious studies. Thus, little solid evidence supports claims of most products to enhance health or sports performance that they actually can do what they promise. Do You Have Research Phobia? By the time you

read this, you have heard about research and science as you sat in classrooms, did homework, and read textbooks. Unless you had talented, enthusiastic teachers, you might have had an unpleasant experience or developed “research phobia.” Maybe it was smelly science labs or

Some people become intimidated by scientific research, or fear being labeled a geek or nerd, or imagine the fictional mad scientist from fiction stories or horror movies.

Why Do Research? 5

challenging math tests without enough preparation time. In many schools, only about 10 percent of students really “get into” research. Many others see it as being irrelevant or strange at best. Some people believe that only college professors, people with medical or Ph.D. degrees, and high-powered professional scientists can do research. Maybe you watched a famous researcher being interviewed on television or have picked up an obscure research publication filled with incomprehensible jargon, statistics, and exotic formulas. You may feel that research is beyond you and has little relevance for your daily life or career. Don’t be turned off by research, after just one class on social research, millions of students improved their decision making by using the techniques, insights, and information-gathering skills of research. One purpose of this book is to demonstrate that empirical social research is not frightening and beyond your ability, nor is it irrelevant. I won’t lie, doing research can be hard work. It allows little room for being sloppy, lazy, “spaced out,” or careless. Research takes concentration, serious thinking, rigor, and self-discipline. In this respect, it is not special. Making quality music or art, cooking fantastic food, growing an outstanding garden, starting up a new business, being a star athlete, or repairing complex machinery also take concentration, serious thinking, rigor, and self-discipline. However, doing research is also creative, exciting, and fun.

applying the knowledge and thinking skills acquired in their school years later in their daily life or job decisions. Another part of the answer is a simple matter of numbers. Imagine that 25,000 educated people want to be better informed. They read a book written by an expert who has researched a topic for six years. At the same time, 100 million people, who are just a little lazy, go out to watch an entertaining, glitzy 90-minute movie on the same topic. Like most movies, it contains a lot of inaccurate and distorted information. More people’s views and thinking will be shaped by the inaccurate and distorted information from the movie than an expert’s book. It is easy to be swayed if a large number of people agree. Just because most people believe something to be true does not make it true.

Learning to Do Research and Using It in “Real” Life Why don’t more people learn to do research

1.1.2: Frustrations and Misunderstandings about Research

and use it in their lives? A simple answer is ignorance. If you do not know it, you cannot use it. However, some people reject the results or method of research not from ignorance, but because the results contradict an intensely held belief, a traditional way of doing things, or because it runs counter to “what everybody knows.” Large numbers of people, even in the United States in the twenty-first century, 2 continue to believe in things that research repeatedly demonstrates to be false, such as the following: • UFOs and ESP (extrasensory perception) • Horoscopes and astrology • Unscientific thinking about age of the earth or basic forces of nature • Goblins, demons, witches, evil spirits, and devils Although the average level of schooling has risen, many people cling to magical-fantasy thinking. National averages in reading comprehension, critical thinking skills, basic social-geographic knowledge, and understanding of scientific research have changed little over the past decades.3 Part of the answer is that many people stop practicing and

WRITING PROMPT Making Decisions What approach do you use to make a major purchase or life decision (e.g., which college to attend)? What factors do you rely on to inform your decision? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

Three features of social science research can create frustration. Some people turn away from research because they misunderstand the significance of such features. One feature is that research does not deliver immediate answers; rather, research knowledge accumulates slowly over time. What is the use of doing research if it cannot give me an answer right now? It may give an answer, but a provisional one. We rely on research because it is better than the alternatives, and its answers tend to improve over time. What we, as a society or all humanity, knew last year may differ little from this year; however, what research tells us today differs significantly from what we knew 20 years ago regarding many issues. A second feature that causes frustration is that research findings are not fixed and unchanging. Its statements or theories4 about events change if there is new or better evidence. Research-based statements, findings, or theories are provisional, not 100 percent absolute and fixed for all time.

4 2

See Harris Interactive (2009). King Burton, Hicks, and Drigotav (2007). 3 See Pew Research Report (2007, 2009).

Theories are systematic, abstract, general explanations about why events occur or how some aspect of the world operates. They contain assumptions, arguments, and a set of interconnected ideas.

6 Chapter 1 They only stand as long as most of the evidence backs them. Moreover, the quality and amount of evidence determine the amount of confidence we place in a statement, idea, or theory. A related third feature that causes frustration is that research findings are usually in the form of chance-like statements, such as the following statements, “teacher attention is likely to affect a child’s learning,” “providing flexible benefits or work hours increases the likelihood of high employee satisfaction,” or “being overweight increases the chances you will develop diabetes.” Few of us accurately evaluate the risks and probabilities of things we do every day: the risk of being injured or killed, the probability of winning the lottery, the likelihood that we will make a profit in the stock market, and so forth, and we are slow to learn from our misjudgments (Mills and Keil, 2004). By contrast, most research carefully estimates the odds that events will occur. So, while people like 100 percent certainty, research findings are stated as the chances or odds of being true. Despite the comfort of simple, fixed answers, you probably already learned that such answers are very rare for complex issues in the real world. Research is an ongoing process of searching and working toward the truth. Its knowledge accumulates over time. Research may not offer us 100percent certainty, but it is the best available human knowledge, better than the alternatives; plus, it improves over time. In the media, in public debates over issues, and in faceto-face discussions about how things are now, why events or behaviors occur as they do, and what might be the best way to resolve an issue, people “make arguments.” The word “argument” as it is being used here does not mean the type where someone gets angry and shouts down another person. Rather, it means building up a set of logically connected statements that start simple and move toward a clear general conclusion that pulls everything together. My students and much of the public confuse two types of argument—the scientific or research-based argument and arguments that rely on a moral position, religious doctrine, or ideological belief. The latter builds on moral, religious, or ideological forms of reasoning. By contrast, research arguments use critical thinking that rests on systematic empirical evidence. We examine the bases of research arguments in the next section.

WRITING PROMPT Frustrations with the Research Process Explain your experience with one of the frustrations of research. Describe the level of frustration you felt when trying to make decisions. The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

1.2: What Is Empirical Social Research and Why Is It Respected? 1.2 Explain why empirical social research is more than gathering information and the role of critical thinking in the research process Research takes many forms. Let us say, I want to purchase a new car so I do “research” by reading websites about car features, by visiting showrooms and test driving cars, by examining reports on crash test, safety and mechanical reliability, and by comparing specifications such as legroom length or tire size. Doing empirical social research also involves gathering information, but it is much more than that. We use the word “research” in the multiple ways, such as the following four: 1. Research is gathering existing reports or findings from respected sources (i.e., academic journals or official government reports) and making sense of it. You first search for and collect information, next evaluate what you found, and finally synthesize the findings. In the process, you may weigh some evidence (your friend says a car looks cool) differently than other evidence (test driving the car or reading about its mechanical failure rate). 2. Research is examining and carefully inspecting a “body of evidence” or data. The evidence may be in the form of numbers, charts, tables, graphs, and statistics, or it can be specific documents (such as interview transcripts, religious texts such as the Bible or Koran, legal texts such as the Constitution or court decisions, historical texts, or even paintings or music scores). Doing research, a person carefully examines, evaluates, and re-examines the evidence to attain a deeper understanding, to reveal patterns and themes, or to find the “truth.” 3. Research is a process of applying accepted techniques and principles. The process is to ask questions in certain ways, gather information systematically, observe in detail, measure precisely, draw a sample, analyze using statistics, or perform experiments. 4. Research is a way of thinking and making arguments. It means applying critical thinking (discussed next) and adopting a critical orientation and skeptical perspective. You examine assumptions, consider alternatives, and do not accept what you see at face value. You reflect on how you and others arrived at decisions. Empirical social research can involve all four of these activities. It is also an ongoing process of accumulating

Why Do Research? 7

information, with results stated in terms of likelihood or probabilities and not as fixed absolutes. Because it is evidence based, findings can change as the accumulated evidence reveals new insights and understandings. The findings from empirical research might differ from what some religious doctrines or leaders say.

Learning from History Research and Religion

Galileo Galilei, Italian astronomer (1564–1642)

Research usually wins out over horoscopes, lucky numbers, or superstition, but organized religion is a more sensitive subject. Most people, especially in the United States, profess a belief in God and feel a deep attachment to a specific religion (usually Judeo-Christian). Scientific research and religion, including non-Judeo-Christian religious beliefs, like Islam and Buddhism, can disagree. This is nothing new. It has been going on at least since Galileo Galilei (1564–1642). Although he remained loyal to his religion, Galileo was also committed to scientific research, honesty, and truth. He rejected a blind allegiance to philosophical and religious authority. Based on careful research and systematic empirical evidence, his views on how the world works changed. He insisted the earth went around the sun and not vice versa, as held by leading religious authorities at that time. Because he opposed established religious doctrine, authorities placed him under house arrest and banned his books. Today we recognize that Galileo was right, and in time, religious authorities relaxed their views and altered their position. The science versus religion conflict is easily overdrawn. At one extreme are some highly devout people who reject all science and believe exclusively in a specific religious faith. Whether it is gravity, age of the earth, medical care, or causes of crime, they believe that religion alone has the only true answers. At the other extreme are nonreligious people who think all religion is false and put their faith in science alone. Whether it is social justice, moral decisions of right or wrong, or life after death, they feel that science offers all the correct answers. Most people, including most scientists, fall in between the extremes. Religious extremists of any religion (Christianity, Islam, Hinduism, or others) and science extremists each believe they have all the answers. Most people find value in both sides and see each side addressing certain types of issues better than other types.

Science-based research may be better than religion at providing answers for some types of questions. Science cannot tell moral right from wrong, whether there is a God, whether we have a soul, or what happens to us after death. In the past, religious authorities made rather foolish statements about astronomy, biology, and many social issues (such as supporting slavery). The line between what belongs to religion or to science is always shifting. At one time, science only dealt with the physical world (planets, chemistry, or plants) and religious thinking dominated all social issues. As research techniques and scientific thinking advanced, people applied them to social issues—such as why crime rates rise and whether children raised in certain ways are better adjusted. Over time, people turned from the old answers, such as the devil or evil spirits caused errant behavior, and increasingly looked to researchbased that identified causes such as increased economic distress, child rearing without strong, clear values, or a weakening of community ties. Research can answer many questions that overlap with religious–moral issues but cannot answer questions, such as: Should I marry John? Is abortion immoral? Is the death penalty right or wrong? Will prayer cure my mother’s cancer? Nevertheless, research can answer related questions, such as: If one marries someone with a background of emotional instability and excessive alcohol or illegal drug use, is one likely to experience physical abuse and divorce? Do women who have an abortion experience less or more social, educational, and economic success? Does having a death penalty lower murder rates? Does praying for a person reduce the spread of cancer? Answers to these questions may help us make decisions about moral and religious concerns even if they do not provide simple, fixed answers. For example, suppose research showed that the death penalty has no effect on reducing murder rates. You may still favor it for other reasons (e.g., lower costs, revenge, or religious beliefs). At least if you know the research findings, you can consciously make judgments based on the facts or something else (e.g., moral or religious belief). Suppose studies showed that praying for a seriously ill person has no impact on recovery rates. You may still want to pray for other reasons. Perhaps it makes you feel better and gives your life a focus, or it gives the ill person a feeling of hope and reduces his or her distress. In short, research and moral–religious reasoning differ, but they can work together and be compatible.

WRITING PROMPT Questions That Research Can Help Answer Research can help provide answers to a great many questions, but not every possible question. What are 2–3 questions that social research can provide answers to, and 2–3 that are outside the scope of research? Explain the key feature separating the answerable versus unanswerable questions. The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

8 Chapter 1

1.2.1: What Are Critical Thinking Skills? One good reason for you to learn to think critically and to conduct research properly is so you do not simply follow what most other people think when it is wrong. Some people are confused about critical thinking because the word “critical” has three meanings: 1. being very serious, or being an urgent need, such as being critically ill; 2. being highly negative or antagonistic and looking for flaws, building on the verb to criticize; and 3. being very aware, judging carefully, and questioning by not accepting anything that happens to come along. The meaning of “critical” in critical thinking refers to the last meaning, although the first one sometimes also fits. Critical thinking is a way to think and see things. Psychologists and others who study how we think have cataloged a long list of common misperceptions or logical fallacies. Just as we can be misled when we look into a distorted mirror, we can fall for these fallacies.

WRITING PROMPT Thinking Critically Consider what you knew about “critical thinking” prior to completing this section. Compare your initial thoughts with how the text defines critical thinking. Describe the differences in the definitions. The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit The Importance of Thinking Critically Critical thinking helps us avoid common fallacies. It also discourages us from rushing to arrive at a fixed, closed, or set answer. Many people feel uncomfortable with ambiguity or an open-ended process of constant searching. They seek an absolute correct answer, here and now. Critical thinking warns us that one quick, simple answer is rarely correct and it should be distrusted. It also points to the value of looking at a question, issue, or evidence from more than one point of view. It tells us that adopting a single point of view often blinds us to important aspects of a question, issue, or problem. Finally, critical thinking leads us to uncover hidden assumptions. Assumptions—unstated premises or untested starting points—are necessary, and we use them all the time. There is nothing wrong with having assumptions. However, assumptions may block off certain avenues of inquiry while favoring others. Problems can arise when we fail to recognize or examine our assumptions. Critical thinking tells us to notice assumptions and see the

ways that they can limit choices. If we adopt alternative assumptions, the outcome may be very different. Often debates or disagreements reach an impasse because participants are using different assumptions. Revealing hidden assumptions can shift the discussion and allow a resolution or at least clarify the real issue. Example 1 Two executives, Mark and Susan, disagree over whether to advance product X or product Y. Product X costs less to produce, has a three-year life expectancy, and yields a $10 profit for each item sold. Product Y has slightly higher production costs but is of higher quality. It lasts six years and yields an $8 profit per item. Mark favors item X. He projects that it will sell 10,000 units to produce $100,000 in total profits. Susan favors item Y. She projects that it will sell 15,000 units for $120,000 in total profits. Their disagreement is not over cost or profits, but over time horizon and customer loyalty. Mark assumes a two-year timeframe and that retaining customers and building brand loyalty are minor concerns. Susan assumes a longer timeframe and that building brand loyalty among customers is important. Example 2 A school district creates two charter schools. New principals of each school want to create a high-quality learning environment. After the schools have been operational for 10 years, 80 percent of students from both schools enter a 4-year college and equal numbers go on to have very successful careers. However, the two principals recruited students differently and attracted very different student bodies based on their assumptions. Principal A assumed that a small sector of society has the talents required to succeed and that academics alone are important. Principal B assumed that most students are capable of success if given a chance and that a socially diverse classroom and student cooperation skills are equally important as academic test scores. Principal A recruited based on academic test scores alone, taking top-scoring students. Principal B recruited one-third of the students based on academics and two-thirds based on other talents (art, athletics, drama, music, or volunteering) or strong motivation. At Principal A’s school, 85 percent of students come from one section of town, Eliteville. It is where the high-income, well-educated residents live. Only 25 percent of students at Principal B’s charter school come from Eliteville; 75 percent come from all over the community. Despite having the identical outcomes in terms of student success, different assumptions about students by the principals created different student bodies and had divergent social consequences. One feature of critical thinking is to unveil hidden assumptions and reveal how assumptions influence outcomes. Besides critical thinking, social scientific research uses a particular form of argument, one that

Why Do Research? 9

builds logically from explicit assumptions and past studies, then rests on empirical evidence (discussed in next section).

Summary Review

Critical Thinking Skills • Avoid logical fallacies; practice careful thinking using “cold, hard logic.” • Maintain an open mind and look at all aspects of an issue; be cautious about simple, fast, and easy solutions offered for serious, complex issues. • Do not get locked into a single point of view; constantly look at issues from multiple perspectives. • Examine hidden assumptions; be aware of assumptions and their implications.

WRITING PROMPT Finding Hidden Assumptions Finding hidden assumptions is the most difficult part of critical thinking. Briefly outline a personnel policy or physical arrangement in a workplace that contains a hidden assumption about people fulfilling traditional gender roles. [Assume the workplace employs an equal number of men and women.] The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

1.2.2: What Counts as Evidence? Social research builds on systematic empirical evidence. We use evidence all the time and evaluate it in everyday life. Would you accept a third-hand rumor as better evidence than a written statement witnessed and co-signed by two neutral observers? Courts of law have rules of evidence that outline what is acceptable and what is not. In many other areas of life, we also adopt standards of what counts as solid, legitimate evidence, and use rules about how to interpret or assign meaning to evidence. Without shared standards and rules, different people might look at the same evidence but arrive at different conclusions. For this reason, social research provides rules of evidence—how to collect it, what counts as good evidence, and how to interpret it. Our confidence in research grows when we have a lot of strong evidence versus scanty or very weak evidence. Research requires you to look at empirical evidence. Moreover, we must collect the empirical evidence carefully and systematically according to rules or standards. The standards of evidence are the key.

Suppose a person is in a boat going down the Colorado River in the Grand Canyon. An artist looks at the rock formations and sees beauty based on aesthetic standards. An environmentalist sees serious erosion based on standards of water flow and soil/rock removal. A geologist sees evidence of ancient geological shifts or volcanic action using standards from the field of geology. A Native American sees evidence of messages from the Great Spirit based on standards from religious beliefs and folklore. Someone else sees evidence of UFO visits based on standards from reading lots of science fiction or fantasy literature. Besides evidence we need standards for specific types of evidence.

Where do the standards for evidence come from? Compare Your Thoughts Over several decades, thousands of researchers conducted tens of thousands of studies, and other people examined the studies searching for flaws. Today’s standards developed slowly as people repeatedly evaluated, critiqued, and suggested areas for improvement in studies. A shorthand way to summarize the standards-creating process is to say that it comes from the scientific community. Many standards for research evidence follow common sense. Suppose you have two samples of all the students at a large university. One sample contains 100 students and another contains 1,000. Everything else about the samples is identical. Common sense says you should use the one with 1,000 students. Suppose you want to find out a person’s attitude about a subject. You can ask that person once in one way with a single question, or you can ask him or her in more than one way, using multiple questions and at multiple time points. Common sense tells you that asking with multiple ways, questions, and times is the better way to learn a person’s true attitude. Other standards in research can get more technical or differ from common-sense thinking.

10 Chapter 1 Evidence in social research takes two forms: quantitative data and qualitative data. Some people confuse the idea of strong evidence with the form of evidence. In other words, they believe that quantitative data are always strong and qualitative data are always weak when, in fact, standards determine whether the evidence is strong or not. Strong evidence is determined by how carefully and systematically a researcher gathered the evidence.

1.2.3: Research Is a Process That Results in a Product Research is an ongoing process. The process has multiple parts or steps and adheres to guidelines. The parts come together and ultimately yield a product, knowledge, or information. It can be knowledge that confirms and reinforces something we already know, or it can be new knowledge that pushes the frontier of inquiry forward. Another way to think of the product of research is that the study results answer a question. Good research also stimulates new thinking and raises new questions that you would not have imagined until doing the research. There are many types of research, and specifics of the research process vary by topic area, form of evidence, and type of research question.

1.3: What Are the Types of Social Research? 1.3 Summarize quantitative and qualitative data collection techniques To conduct a study, you need to collect evidence or data using one or more specific techniques. This section is a brief overview of the major data gathering techniques. Research techniques are divided based on whether they provide quantitative or qualitative data. Quantitative data collection techniques include experiments, surveys, content analysis, and review of existing statistical sources. Qualitative data collection techniques include ethnographic field research and historical comparative research. It is important to note that these can be blended as needed to meet the unique needs of a given study. Some techniques are more effective when addressing specific types of questions or topics. Unfortunately, some people get locked into using one technique (such as questionnaire) and use it all the time, even when it’s not the most effective technique for a situation. A good researcher must be aware of the full range of techniques as well as their strengths and limitations. It takes skill, practice, and creativity to match a research question to an appropriate data collection technique.

1.3.1: Four Quantitative Data Collection Techniques Experiments Experimental research closely follows the logic and principles found in natural science research. To conduct an experiment, you create a situation and examine its effects on study participants, in laboratories or in real life. Experiments require a well-focused research question, a controlled setting, and small numbers of participants. In the typical experiment, you divide the people being studied into two or more groups. Next, you treat both groups identically, except you do something with one group (but not the other): the “treatment.” You measure the reactions of both groups precisely. By controlling the setting for both groups and giving only one group the treatment, you can conclude that any differences in the group reactions are due to the treatment alone. Recall the chapter opening study of young children presented with food in McDonald’s packaging versus the same food presented without the packaging. This study used the experimental technique. Instead of treating the children differently, the researchers altered the food packaging. In the study on banning toy sales promotions for fast food in Santa Clara County, the ban was a natural experiment that outlawed toys for unhealthy meals in some restaurants while the same toys were still available in restaurants outside the ban area. Surveys In survey research, you ask people questions. Questions can be presented in a written questionnaire (electronically or hard copy) or during an interview with the interviewer recording their answers. Unlike the experiment, you do not manipulate a situation or condition. You simply ask a group of people numerous questions in a short period of time. Typically, you later summarize answers to questions in percentages, tables, or graphs. Surveys give you a picture of what many people think or report doing. Often, a sample or a smaller group of selected people (e.g., 150 students) is used in survey research. If properly conducted, you can generalize results to a larger group (e.g., 5,000 students) from which the smaller group was selected. Content Analysis Content analysis is a technique for

examining information in written or symbolic material (e.g., books, newspapers, pictures, movies, song lyrics, etc.). In content analysis, you first identify a body of material to analyze. Next, you create a system for recording specific aspects of it. The system might include counting how often certain words or themes occur. Finally, you record what you found in the material. Often you measure information in the content as numbers and present it as tables or graphs. This technique lets you discover features in the content of large amounts of material that might otherwise go unnoticed. Existing Statistical Sources In existing statistics

research, you first locate previously collected information.

Why Do Research? 11

Often it is in the form of public documents, government reports, or previously conducted surveys. Next, you reorganize the information in new ways to address a research question. Locating sources can be very time-consuming. You may not even know whether the information for your research question is available when you start a study. When you reexamine existing quantitative information, you may use various statistical procedures.

1.3.2: Two Qualitative Data Collection Techniques Ethnographic Field Research To conduct a field

research study, you closely observe a small group of people over a length of time (e.g., weeks, months, years). Usually you begin with a loosely formulated idea or topic, not a fixed theory or hypothesis. You select a social group or natural setting for study, gain access and adopt a social role in the setting, and observe in great detail. You will get to know the people being studied personally and may conduct open-ended and informal interviews. In field research, it is important to take very detailed notes on a daily basis. After leaving the field site, you will reread the notes and prepare a written report.

WRITING PROMPT The Six Research Techniques List the six types of research techniques. For each either explain how you used it and the results or if you haven’t used the technique, describe how it could improve your research. The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

1.4: What Is the Purpose of the Research and How Is It Used? 1.4 Examine how the purpose of research depends upon the outcome that the researcher is trying to accomplish

Summary Review

Newcomers to research struggle to connect a specific research question with a particular research technique. Often, you will need to reformulate or rephrase the question. Even then, it is not easy to match your question to a type of research. First, you should clarify the purpose of the research. People conduct research with different goals or purposes. If you ask someone why he or she is conducting a study, you might get a range of responses: “My boss told me to”; “It was a class assignment”; “I was curious”; “My roommate thought it would be a good idea.” There are nearly as many reasons to do research as there are researchers. We can organize the purposes of research based on what we are trying to accomplish—explore a new topic, describe a social phenomenon, explain why something occurs, or evaluate an outcome. Studies may have multiple purposes (e.g., both to explore and to describe), but one major purpose is usually dominant.

Quantitative and Qualitative Data Collection Techniques

1.4.1: What Is the Purpose of a Study?

Historical-Comparative Research In historical-

comparative research, you examine aspects of social life in a past historical era or across different cultures. You may focus on one historical period or several, compare one or more cultures, or mix historical periods and cultures. As in field research, you combine theory building/testing with data collection. You begin with a loosely formulated question that you refine during the research process. In this type of study, you gather a wide array of evidence. Evidence often includes existing statistics and documents (e.g., novels, official reports, books, newspapers, diaries, photographs, and maps). In addition, you may make direct observations or conduct interviews.

Table 1.2 Quantitative and Qualitative Data Collection Techniques Quantitative Data Collection Techniques

Qualitative Data Collection Techniques

• • • •

• Ethnographic field research • Historical-comparative research

Experiments Surveys Content analyses Existing statistics

Exploring a Previously Unknown, Brand-new Issue In exploratory research, you examine an area that

no one has studied or see it in new ways. Exploratory research is often a first stage in a sequence of studies. You may need to conduct an exploratory study to find out enough to design and execute a second, more systematic and extensive study. Exploratory research addresses the “what?” question: “What is this social activity all about?” or “What is happening here?”

12 Chapter 1 Exploratory researchers tend to use qualitative data. They are not wedded to a specific theory or research question. The goal is to formulate questions for future research, adding more focus, not to produce definitive answers. If you conduct an exploratory study, you may get frustrated and feel it is difficult because there are few guidelines to follow. Everything is potentially important, the steps are ill defined, and the direction of inquiry changes frequently. You need to be creative, open minded, and flexible; adopt an investigative stance; and explore all sources of information.

Example Study Exploratory Research

Elijah Anderson (2012) conducted exploratory research in Philadelphia. Using ethnographic field research, he spent a great deal of time in many downtown locations (e.g., parks, restaurants, shopping malls, bars, food courts) and conducted many informal interviews. Based on his careful observations over an extended period of many natural settings and by analyzing those observations, he arrived at a new idea, the cosmopolitan canopy, to describe a type of urban social interaction he saw. It refers to physical places and times in which the people present engage in compatibility, acceptance, and civil behavior toward strangers from different racial-ethnic backgrounds and social classes. It emerges informally and spontaneously, and it differs from other, more guarded, behaviors that usually occur among people from different backgrounds or classes in urban settings.

Describing an Issue, Situation, or Relationship In-depth Perhaps you may have an idea or ques-

tion about a topic—a new marketing plan, a way to improve patient care, a way to deliver a client service, a way to help a group of new students, and so forth. Often, someone studied a similar topic somewhere at some time, but you want to learn about it in a specific place. The goal of descriptive research is to present a picture of the specific

details of a situation, social setting, or relationship. It focuses on questions, such as: How did it happen? Who is involved? When did it start and stop? Where does it occur? A great deal of social research is descriptive. Much of the social research found in scholarly journals or used for making policy decisions is descriptive.

Example Study Descriptive Research You have probably heard that people who complete a fouryear college degree tend to have greater long-term career success (i.e., more opportunities, higher status jobs, better earnings) than those without a college degree. Vocational and technical schools tend to offer training for very specific jobs. At four-year colleges, the majors vary greatly; some prepare students to immediately enter the job market (e.g., accounting, engineering, education, nursing, and social work), while others do not prepare students for a specific job (e.g., art, communication, math, biological science, English, history, or sociology). Many factors influence choice of major, including concerns about a career. Roksa and Levey (2010) conducted descriptive research to examine the impact of different types of college majors on career success, both in the short-run (first job after college) and in the long-term (12 years after college). They analyzed data from a national survey of over 12,600 randomly selected young people (14–22 years old). The youth were first interviewed in 1979, and they were re-interviewed every one to two years since then. The study found that students with highly job-specific majors had an easier time getting a job right out of college and got higher starting salaries than students who had less jobspecific majors. However, over the long term, students with nonspecific majors had greater career success and higher incomes than those who had narrow job-specific majors. This descriptive study documented differences by major and ruled out alternative possibilities, but it did not explain why the divergent short-term and long-term career outcomes occurred for different types of majors.

Explaining Why an Event or Situation Happens or Occurs in Specific Ways When you have a

well-recognized issue and have an accurate description of it, you might wonder why things are the way they are (Why do some parents abuse their children? Why do younger males like a certain type of music that older females dislike? Why do certain college majors result in greater long-term success?). Explanatory research builds on exploratory and descriptive research to identify the sources of social behaviors, beliefs, conditions, and events. It emphasizes the “why question” and documents causes, tests theories, and provides reasons. For example, an exploratory study discovers a new type of child abuse by parents; a descriptive study

Why Do Research? 13

documents that 8 percent of parents abuse their children in a certain way and describes the kinds of parents and conditions of its occurrence; an explanatory study could focus on why certain parents abuse their children in this manner. It may test two competing theories about why some people become abusive parents.

Example Study Explanatory Research

In related explanatory research, Scheitle and Han (2011) asked whether a high concentration of people in conservative Protestant denominations in a state explained variation across the states in laws and policies regarding sexual orientation, including same-sex marriage. They analyzed data from a public source that scored each state’s policies regarding sexual orientation, state public opinion, membership in religious denominations by state, and the strength of organized anti-LGBT advocacy groups in each state. To their surprise, the presence of many conservative Protestant denominations did not explain hostility toward sexual orientation policies. The key explanatory factor was the strength of anti-LGBT advocacy groups and the influence of such groups in state-level politics. In short, the mere presence in a state of many people with religious beliefs or membership in denominations opposed to LGBT rights is not the critical factor. Instead, the ability of political advocacy groups to mobilize people around the issue explains why some states have more LGBT hostile policies than other states.

Evaluating Whether a Program or Policy Works Researchers design an evaluation research

African Americans oppose same-sex marriage more than other racial categories. In the 2008 California Proposition 8 to ban same-sex marriage, 70 percent of African Americans voted in favor of the ban versus about half of people in other racial categories. National survey results showed no racial gap in 1988. At that time, a majority (68 percent) of whites, blacks, and those of other racial categories all opposed same-sex marriage. Twenty years later, opposition by whites and others had dropped to 45 percent, while black opposition increased and was 13 percent higher. Skerkat, Vries, and Creek (2010) conducted explanatory research to learn why African Americans oppose same-sex marriage more than other racial categories. They considered differences in education levels, age structure, geographic concentration, and religion. The authors noted that African Americans more than whites affiliate with conservative Protestant denominations that strongly oppose hom*osexuality. Fewer than 30 percent of whites but over 63 percent of blacks affiliate with a conservative Protestant denomination. By contrast, over 30 percent of whites but less than 10 percent of African Americans are Catholic. Compared to conservative Protestant sects, members of mainline Protestant denominations and Catholics tend to support lesbian, gay, bisexual, and transgender (LGBT) rights. The authors conducted statistical analyses of survey data between 1988 and 2008, and could explain most of the racial gap in same-sex marriage opposition by the religious attendance and membership in conservative religious organizations by African Americans. In short, the reason why African Americans show more opposition has little to do with race per se. The racial gap in same-sex marriage opposition is due to differences in religious participation and denominational ties that differ by racial category.

study to find out whether a program, a new way of doing something, a marketing campaign, or a policy is effective— in other words, “Does it work?” Large bureaucratic organizations (e.g., businesses, schools, hospitals, government, large nonprofit agencies) often conduct evaluation studies to demonstrate the effectiveness of what they are doing. The specific research techniques in an evaluation study are no different from other kinds of research. The difference lies in the purpose of the research. An evaluation study might ask questions, such as: Does a Socratic teaching technique improve learning over the lecture method? Does a law-enforcement program of mandatory arrest reduce spouse abuse? Does a flextime program raise employee productivity? Evaluation researchers usually measure the effectiveness of a program, policy, or way of doing something and often by using multiple research techniques (e.g., a survey and field research). If they can, many evaluation researchers like to use the experiment technique. Practitioners involved with a policy or program may conduct evaluation research on their own or do so at the request of outside decision makers. Outside decision makers sometimes place limits on the research, fix boundaries on what a study can look at, or pressure for a particular finding. This may create ethical dilemmas for a researcher. Even if an evaluation study yields clear evidence about a program’s effectiveness, people may not use the results. People might ignore solid empirical evidence and make decisions based on other factors—such as moral, political, and personal reasons. Despite clear evidence, they will continue programs found to be ineffective or end highly effective ones.

14 Chapter 1

Example Study Evaluation Study

Established in 1983, the Drug Abuse Resistance Education or D.A.R.E. program operates in about 80 percent of all U.S. school districts. It focuses on elementary schools, with middle and high school curricula reinforcing the early lessons. The D.A.R.E. elementary school curriculum, usually in the fifth or sixth grade, con-

sists of 17 lessons taught by trained uniformed police officers. The lessons teach students about various drugs and about decision-making and peer pressure resistance skills. Many evaluation research studies5 have followed students who participated in D.A.R.E. programs and compared them with students who did not; the studies have looked at the five to seven years after program participation. Findings from the studies suggest that drug use differs little between the two sets of students. In short, program participation has not achieved its primary goal, to reduce illegal drug use among teens. Despite repeated evidence that it does not work, the program continues to be popular among parents, school officials, local businesses, and police. After 30 years and many billions of dollars, the ineffective program continues due to social and political reasons, a strong desire to reduce drug use, and few alternatives that have won such wide acceptance. 5

For some of the many studies showing the ineffectiveness of D.A.R.E., see Brown and Kreft (1998), Tobler and Stratton (1997), Wysong et. al. (1994), and Zagumny and Thompson (1997).

Making It Practical: General Accounting Office Letter on Evaluation of the D.A.R.E. Program

Figure 1.3 Selection from General Accounting Office Letter on Evaluation of the D.A.R.E. Program

United States General Accounting Office Washington, DC 20548 January 15, 2003 The Honorable Richard J. Durbin United States Senate Subject: Youth Illicit Drug Use Prevention: DARE Long-Term Evaluations and Federal Efforts to Identify Effective Programs Dear Senator Durbin: The use of illicit drugs, particularly marijuana, is a problem among our nation’s youth. The adverse effects of illicit drug use play a role in school failure, violence, and antisocial and self-destructive behavior. A recent national survey* showed that for 1996 through 2002, more than 30 percent of tenth and twelfth grade students reported using marijuana in the past year. Further, about 20 percent of high school seniors reported using marijuana within the past 30 days. In fiscal year 2000, the federal government spent over $2.1 billion on illicit drug use prevention activities for youth, according to the Office of National Drug Control Policy (ONDCP). Many programs are designed to help prevent and reduce illicit drug use among youth. Often, these programs also address the use of other substances, such as alcohol and tobacco. Youth drug abuse prevention programs are implemented in school, family, and community settings. School-based prevention programs are the most prevalent because schools provide easy access to children and adolescents. The most widely used school-based substance abuse prevention program in the United States is the Drug Abuse Resistance Education (DARE) program,† which is funded by a variety of sources, including private, federal, and other public entities. DARE’s primary mission is to provide children with the information and skills they need to live drug- and violence-free lives through programs at the elementary school, middle school, and high school levels. The DARE program is usually introduced to children in the fifth or sixth grade. According to research literature, concerns have been raised about the effectiveness of the DARE fifth and sixth grade curriculum in preventing illicit drug use among youth. As agreed with your staff, this report contains information you requested on (1) the results of evaluations on the long-term effectiveness of the DARE elementary school curriculum in preventing illicit drug use among children and (2) federal efforts to identify programs that are effective in preventing illicit drug use among children. *Lloyd D. Johnston, Patrick M. O’Malley, and Jerald G. Bachman, Monitoring the Future National Results on Adolescent Drug Use: Overview of Key Findings, 2001, NIH Publication No. 02-5105 (Bethesda, Md.: National Institute on Drug Abuse, 2002). †

The DARE program is administered by DARE America—a nonprofit foundation.

Why Do Research? 15

Figure 1.3 Continued To identify evaluations on the effectiveness of DARE at preventing illicit drug use among children, we searched social science, business, and education databases, which included the Department of Health and Human Services’ (HHS) National Institutes of Health’s (NIH) National Library of Medicine, for evaluations of DARE published in professional journals. We identified articles published in the 1990s on six evaluations of the DARE elementary school curriculum that included illicit drug use as an outcome measure and that also met key methodological criteria for our review, such as a long-term evaluation design and the use of intervention and control groups for comparisons. The six long-term evaluations that we discuss in this report were conducted at different times up to 10years after student participants were initially surveyed. The six evaluations are based on three separate studies in three states. We reviewed each of the six evaluations and summarized the results of our review. We also held discussions with the researchers who conducted the evaluations. We did not independently validate the research designs or verify the results of evaluations on the effectiveness of the DARE program. (Enclosure I contains citations for the articles on evaluations of the DARE elementary school curriculum that we reviewed and enclosure II describes the methodology we used to select the evaluations). To determine federal efforts to identify programs that are effective in preventing youth illicit drug use, we interviewed federal officials and reviewed documentation on efforts by HHS and the Department of Education (Education) to recognize programs that demonstrate success in reducing illicit drug use among children and adolescents. We did not independently verify the results of prevention programs recognized by the federal agencies. We conducted our work from January through December 2002 in accordance with generally accepted government auditing standards. In brief, the six long-term evaluations of the DARE elementary school curriculum that we reviewed found no significant differences in illicit drug use between students who received DARE in the fifth or sixth grade (the intervention group) and students who did not (the control group). Three of the evaluations reported that the control groups of students were provided other drug use prevention education. All of the evaluations suggested that DARE had no statistically significant long-term effect on preventing youth illicit drug use. Of the six evaluations we reviewed, five also reported on students’ attitudes toward illicit drug use and resistance to peer pressure and found no significant differences between the intervention and control groups over the long term. Two of these evaluations found that the DARE students showed stronger negative attitudes about illicit drug use and improved social skills about illicit drug use about 1 year after receiving the program. These positive effects diminished over time. We are sending copies of this report to the Secretary of HHS, the Secretary of Education, the Director of the Office of National Drug Control Policy, and others who are interested. We will also make copies available to others upon request. In addition, the report is available at no charge on GAO’s Web site at http:// www.gao.gov. If you or your staff have questions about this report, please contact me at (202) 512-7119 or James O. McClyde at (202) 512-7152. Darryl W. Joyce and David W. Bieritz made key contributions to this report. Sincerely yours,

Marjorie Kanof Marjorie E. Kanof Director, Health Care—Clinical and Military Health Care Issues

A research topic and interest often determine the purpose of a study, and the purpose of a study tends to go hand in hand with specific research techniques. Experiments are most effective for explanatory purposes and popular for evaluation research. Researchers use survey techniques in descriptive or explanatory purposes. They can use content analysis for exploratory and explanatory purposes but primarily use it for descriptive purposes. Researchers use existing statistics research for all three purposes, but they most frequently use it for a descriptive purpose. Researchers use field research most often for exploratory and descriptive purposes. Historical-comparative research can be used for all three purposes separately or together. In sum, the purpose of a study influences many aspects of research. Ethical and political conflicts tend to arise in evaluation research because people may have opposing interests in the findings. Research findings can affect who gets or keeps a job, they can build political popularity, or they may help promote an alternative program. If you conduct a serious evaluation study of a program and find it to be ineffective and a waste of time and money, people who make a living carrying out that program may become unhappy with you. They may express their displeasure by attacking you personally or criticizing your methods of research. This tells us something else about

research: Research produces knowledge, and knowledge can be powerful. New knowledge aids in making decisions; it can also upset people who are benefiting from ignorance.

1.4.2: How Will We Use theResearch? In general, we can place research into two categories: basic and applied. Some researchers aim to advance general knowledge over the long term by adopting a detached, purely scientific, and academic orientation toward research. A researcher interested in this goal usually pursues basic social research. Others are activist, pragmatic, and interventionist oriented, and they try to solve specific, immediate problems. A researcher with this goal in mind would probably pursue applied research. However, this is not a rigid separation, and researchers in both orientations cooperate. Individual researchers often move from one orientation to another over time. Basic Research Basic social research advances fundamental knowledge. In basic research, we focus on refuting or supporting theories that explain how the social world operates, what makes things happen, why social relations are a certain way, and why society changes. Its

16 Chapter 1

Summary Review

Purposes of Research Table 1.3 Purposes of Research Type of Study

Purpose

Stage in Learning Process

Question

Main Audience

Outcome

Exploratory

Learn about something entirely new and unknown

Earliest

What?

Varies; usually a researcher

General ideas and research questions

Descriptive

Provide details on something known

Middle

Who? When? How?

Varies

Factual details and descriptions

Explanatory

Build a new or test an existing explanation

Late

Why?

Professional researchers

Test a theory; compare explanations

Evaluation

Determine the effectiveness of a program or policy

Late

Does it work?

Practitioners and policy makers

Practical recommendations

main audience are others in the research community, and it tends to be objective or detached. Basic research is the source of most new scientific ideas and the source of most new and advanced research techniques. Many nonscientists criticize basic research and consider it a waste of time and money. Although basic research might lack a practical short-term application, it provides a foundation for knowledge that advances general understanding in many policy areas, problems or issues. It is also the source of most of the tools, methods, theories, and ideas about underlying causes of how people act or think used by applied researchers. Basic research can produce significant advances in knowledge with the potential to shift how we understand a wide range of issues over the next 50 to 100 years. Frequently, the usefulness of basic research emerges years or even decades later. Practical applications may be apparent only after many accumulated advances in basic knowledge build up over time. For example, in 1984 Alec Jeffreys, a geneticist at the University of Leicester in England, was engaged in basic research studying the evolution of genes. As an “accidental” side effect of a new technique he developed, he discovered how to produce what is now called human DNA “fingerprints.” This was not his intent, and he even admitted that he would have never thought of the technique if DNA fingerprints had

been his goal. Today’s standard crime investigation technique, DNA analysis, was an u nintended outcome of basic research into a different issue over two decades ago. Applied Research Applied research addresses a specific concern or offers solutions to a problem identified by an employer, club, agency, social movement, or organization. When doing applied research, we rarely worry about building, testing, or connecting findings to a larger theory; developing a long-term general understanding; or carrying out a large-scale investigation that might span years. Instead, we conduct a quick, small-scale study that can provide results for immediate, practical use. Applied research most often is descriptive or evaluative. Its main audience includes practitioners and clinicians (such as teachers, counselors, police officers, and social workers) or decision makers (such as sales or production managers, agency administrators, and public officials). The purposes are usually exploratory, descriptive, or for evaluation. Applied research affects decisions such as the following: Should an agency start a new program to reduce the wait time before a client receives benefits? Should a police force adopt a new type of response to reduce spousal abuse? Should a political candidate emphasize his or her stand on the environment instead of the economy? Should a company market a skin care product to mature adults instead of teenagers?

Example Study Applied Research Teen pregnancy rates in the United States are among the highest in the developed world. By their eighteenth birthday, 6 in 10 teen girls and 5 in 10 teen boys have had sexual intercourse, usually

without birth control. Nearly 80 percent of the fathers of babies born to teen mothers do not marry their babies’ mothers. Sexual activity among young teens is a concern of parents, health officials,

Why Do Research? 17 pushed abstinence-only programs after 1996, Gusrang and Cheng (2010) found that others (i.e., parents, educators, physicians, local religious leaders) who disagreed with it had been shoved aside and lost decision-making influence. Today, over $128 million of federal money goes to abstinence-only programs each year. By 2002, one in four teens had received abstinenceonly education. What has been its effect? From 1995 to 2002, as more teens received abstinence-only education, fewer of them knew about birth control (see Guttmacher Institute, 2006). After years of abstinence-only programs, many of them faith based, some people began to ask: Do these programs work? Several studies provided consistent answers: Abstinence-only programs showed no lasting impact on reducing teen sexual activity, and teens in them know less about sexual diseases than teens in other programs (see Figure 1.4).

Read More After these applied research studies, most health organizations (such as the American Medical Association, the American Academy of Pediatrics, the National Institutes of Health, the Institute of Medicine, and the Office of National AIDS Policy) backed comprehensive sex education. They are not alone; 82 percent of the U.S. public support comprehensive sex education (Archives of Pediatric and Adolescent Medicine, November 2006). In April 2007, an independent research company, Mathematica Policy Research, Inc. (2007), released the most rigorous evaluation study on abstinence-only education. It showed that the abstinence-only programs were ineffective. Some people called for ending the programs, while others reaffirmed their support for abstinence-only programs. Proponents of abstinence-only programs say that study results showing no effect only mean that the programs have not been intense or long enough. Rejecting the evidence, they favor expanding abstinence-only programs. What do you think? Has the U.S. government wasted nearly $1 billion on a failed program, or is abstinence-only the best way to address issues of teen sexual behavior and pregnancy?

religious leaders, educators, and politicians. There is a broad consensus that a problem exists, but people differ on solutions. Almost everyone favors some type of sex education. They differ over its content. Most health care professionals and educators favor comprehensive sex education (also called abstinence-plus). It teaches teens about the social and biological aspects of adult sexuality, sexual diseases, and forms of birth control. Conservative politicians and some religious groups favor abstinence-only education. It promotes chastity until marriage and does not teach about birth control. Perhaps you have participated in a sex education program that promoted abstinence. Even if you have not participated, you may have heard of such programs. In 1996, the U.S. Congress authorized $50 million per year for abstinence-only programs. States added matching money. As the federal government

Figure 1.4 Teen Sex Education (Facts on Sex Education in the United States, Guttmacher Institute, December 2006)

Full Birth Control Information

Abstinence Information Only

Females

Males

100

1995

Because applied research often has immediate implications or involves contentious issues, it generates conflict and ignites social controversy more than basic research does.

100

2002

Decision Makers and Applied Research Results Decision makers who use applied research

results may or may not use them wisely. Although the primary goal for doing research is to find out what is really

18 Chapter 1 happening, many practitioners or decision-makers use findings to support other interests or priorities. Example A famous social researcher, William Whyte (1984), conducted applied research in a factory in Oklahoma and in restaurants in Chicago. In both cases, he found that the decision-makers were uninterested in his findings or wanted to suppress them. In the first case, the factory management was interested only in defeating a union, not in learning about employment relations. In the second case, restaurant owners only wanted to make their industry look good and did not want any findings about the unfavorable aspects of actual restaurant operations to appear in public. Decision-makers in private and public organizations, including the U.S. Supreme Court, sometimes make decisions based on questionable applied studies that advocates have conducted solely to advance their own position on an issue. According to one article, legal briefs presented before the court “would not pass muster in a high school research paper” and so-called “facts” cited in Supreme Court decisions came from “blog posts, e-mails, or nothing at all” or from studies that advocates had created solely to influence the Court and that had only been “published” on the Internet.6 Sometimes, the problem is an absence of quality applied studies on a relevant issue. Other times it is that decisionmakers fail to recognize quality differences in research. The problems will lessen with an increased availability of wellconducted and relevant studies, more people relying on research findings for guidance, and decision makers better educated regarding the hallmarks of quality in research.

WRITING PROMPT Questionable Applied Studies by Advocates Some advocates conduct and promote many weak applied studies to advance their point of view. What questions should you be asking when you first hear about a study? Write three questions you would ask to help identify warning signs to you or others that the study should not be fully trusted. The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

and then a bandon it to move to the next step. Rather, the steps blend into each other in an ongoing, interactive process. What we do in a later step may stimulate us to reconsider and adjust thinking about an earlier one. Thus, instead of simple forward-only flow we make back-and-forth adjustments before reaching an end. The seven steps are followed for each research project. For example, new details learned during the data collection step may cause us to adjust the earlier steps of study design and focusing the study.

Figure 1.5 Steps in the Research Process Source: Neuman, W. Lawrence, 2007. Basics of Social Research, 2nd ed. Allyn & Bacon.

1. Select Topic 7. Inform Others

6. Interpret Data

5. Analyze Data

2. Focus Question

THEORY 3. Design Study

4. Collect Data

Research is an ongoing enterprise. It builds on past research and contributes a larger, collectively created body of knowledge and understanding. One specific study is only a small part of the larger whole. Except for a very narrow applied question, we rarely end at just one study. The research process requires a constant addition of new studies and findings. A single researcher might work on multiple research projects at once, or several researchers may collaborate on one project. Likewise, one research project may result in one research report or several, and sometimes a single report describes several smaller projects.

WRITING PROMPT The Seven-Step Process

1.5: What Are the Steps inthe Research Process? 1.5 List the seven steps in the research process The seven-step process shown in Figure 1.5 is oversimplified. In practice, we rarely complete one step entirely

The full research process has seven steps. List the two steps of the research process that you think people skip or fail to take seriously. For each, explain the likely consequences. The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

Summary: What You Learned about Doing Research Although not 100 percent foolproof, research is a powerful way to improve decision making. It reduces misjudgment, bias, and distorted thinking. Research is rigorous and timeconsuming, and it uses critical thinking. By using critical thinking skills, you become more careful about what you accept as true. Critical reasoning will help you to be a better consumer of research: It rejects putting “blind faith” in research. It encourages you to understand how research works and to develop an ability to evaluate the quality of research. Many students, after just one basic class in doing research, can use the techniques, insights, and informationgathering skills to improve their decision making. There are many kinds of research. Empirical social research is an ongoing process of accumulating information based on empirical evidence. Its findings are stated in terms of probabilities, not as fixed absolutes. It is based on carefully gathered evidence that meets certain standards. The standards have been developed over many years of studies by many researchers who critiqued one another’s research, so there is improvement over time. Research is about answering questions. One way to fit research with a question is to consider the purpose of the research. There are four main purposes in doing research: exploring a previously unknown, brand-new issue; describing in depth some issue, situation, or relationship; explaining why an event or situation happens or occurs in specific ways; and evaluating whether a program/policy works. The evidence or data for research can be in a quantitative (numbers) or qualitative (words, images) form. The care and detail of gathering data, not its form (numbers versus word), determine whether the data are solid and trustworthy. Research is used both to build new knowledge (basic) and to address practical issues (applied). In either case, a research study generally follows a similar sequence of steps: You begin with a topic that is narrowed to a focused question. Next, you decide on details of a study design and how data collection will proceed. You analyze data for patterns, which are interpreted to address the original question. The final step is to communicate what you learned in the study and how you conducted the study.

Quick Review On What Basis Do We MakeDecisions? 1. Research is one of several ways to make decisions about personal, family, community, and professional issues, and to advance general understanding.

2. The word “research” is used in multiple ways. 3. People often make decisions based on alternatives to using research; although the alternatives may work, they tend to be less accurate and can be misleading. 4. Doing research is demanding but possible. 5. Research is not 100 percent correct 100 percent of the time, but it is superior to the alternatives and has internal checks. 6. Three features of research cause frustration: it is slow, changing, and stated as likelihood not as absolute. 7. Two types of argument are used to discuss issues: scientific argument and everyday argument.

What Is Empirical Social Research, and Why Is it Respected? 1. Research requires critical thinking skills that include having an open mind and looking at issues from multiple perspectives. 2. Critical thinking helps us avoid common fallacies and reveals hidden assumptions. 3. To do research, we collect and examine empirical evidence. 4. The quality of empirical evidence depends on standards developed by the scientific community. 5. Research is a process that produces a “product,” namely new knowledge.

What Are the Types of Social Research? 1. A good researcher is aware of all data collection techniques as well as their strengths and limitations. 2. Some techniques are more effective for specific types of topics, and it takes skill and practice to match a topic to an appropriate data collection technique. 3. Four quantitative data collection techniques are experiments, surveys, content analysis, and existing statistical analysis. 4. Two qualitative data collection techniques are ethnographic field research and historical-comparative research.

How Will We Use the Research? 1. There are four main purposes for doing research. One purpose of research is to explore what is happening, where the “what” refers to something new, unexpected, or relatively unknown.

20 Chapter 1 2. The second purpose is to describe in a systematic way the questions of who, how, and when for an event, situation, or activity in society. 3. The third purpose is to explain why something occurs and give reasons for an event, situation, or activity in society. 4. The last purpose is to evaluate the effectiveness of a program, designed activity, or policy. 5. We primarily use basic research studies to advance fundamental knowledge in the long term. 6. We primarily use applied research studies to address short-term practical questions and policies.

Shared Writing: How Useful is Research? Many people favor either applied or basic research. List the one you prefer most, and explain why you prefer it over the other. Review and comment on at least two classmates’ responses, whose selection differs from yours. Examine their rationale and identify at least one limitation that could affect the success of their research. A minimum number of characters is required to post and earn points. After posting, your response can be viewed by your class and instructor, and you can participate in the class discussion.

Post

0 characters | 140 minimum

Chapter 2

Planning a Study

Learning Objectives 2.1 Identify the various sources for topics of a

social science study 2.2 Analyze sources of research literature 2.3 Identify the six-steps of the literature review

process Do you have a tattoo? Have you ever wondered why people get tattoos? Your curiosity about tattoos can be the start of a research study. You might begin by looking at the several books and the nearly 25 social research articles on tattooing published in the last five years. This may help you turn the broad topic of tattoos into a research question for a study. You might ask, why tattoos are popular for people in certain cultures or times? This question looks at the cultural

2.4 Evaluate what makes a good research

question using the inductive and the deductive approaches 2.5 Analyze the various processes involved in

designing a study for a research proposal and historical development of tattoos from the origin of the word from the Tahitian tatau to their use in religious rituals. Tattoos were used thousands of years ago in what is today Japan, Siberia, India, Peru, and Egypt. Certain people, such as the Maori in New Zealand, some Amazon tribes, and certain subcultures, such as Japanese crime gangs or Neo-Nazi skinheads, regularly tattoo to communicate symbolic messages to others or as forms of ornamentation.

22 Chapter 2 Alternatively, you might ask, how many people in the United States today have tattoos and what types of people get them? To answer these questions, you might conduct survey research. An online survey by Harris Interactive of 2,000 U.S. adults in 2012 found that 21 percent of adult Americans have a tattoo (Harris, 2012). This is up from 16percent when the question was first asked in 2003 and varies by age. While 38 percent of people aged 30–39 have one, only 5 percent of people over 65 do. There is little difference by gender, but more self-identified Hispanics (30percent) than whites (20 percent) or blacks (21 percent) have them. A survey of high school students (grades 9–12) in one city by Dukes and Stein (2011) found about 18 percent of both genders had tattoos, and 16 percent of boys and 42 percent of girls had piercings. Those with tattoos or body piercings (excluding pierced ears) had less positive school attitudes, lower educational aspirations, more weapons possession, more substance use, more delinquency, and lower self-esteem than adolescents without tattoos or piercings. You might wonder, how others think about people with tattoos. To answer this question, you could conduct a survey of college students using a Martin “Stigma against Tattoo Survey” to assess stigma toward those with tattoos (Martin and Dula, 2010). Or, you might conduct an experiment similar those by Hawkes, Senn, and Thorn (2004) or Resenhoeft, Villa, and Wiseman (2008). Both experiments examined people’s reactions to college females with a tattoo. In the first study, participants in the study read about women. The researchers varied details about each woman’s characteristics and her tattoo (its size and location). In the second study, people saw photos of women with and without tattoos, and with dragon versus dolphin tattoos. If your question is about tattoos on people in music videos, you could conduct a content analysis study of music videos to see what tattoos are shown and who has them. Maybe you are curious about the business of tattooing. You could examine existing statistics and records to find the number of tattooing businesses, suppliers, and tattoo artists. If you are curious about subjective beliefs of people who get tattoos, you could conduct a qualitative field research study like that by Atkinson (2004). He spent a lot of time with tattooed people and tattoo artists and got to know them very well. Alternatively, you might focus on field research with a specific subgroup, such as gang members or Neo-Nazis, to see whether they see their tattoos differently than others. Many young people in North America today who get a tattoo say that it signals a rejection of authority, is a statement about control over their body, indicates group membership, or is a form of spiritual-artistic self-expression. Of course, once top celebrities or most of your friends get a tattoo, you may get one to mimic an idol or to conform to peer pressure.

The many studies on tattoos1 illustrate how to turn a topic into the start of a research study. We will look at how to take a topic and design a research study to examine that topic in depth.

2.1: How Do We Select a Topic to Study? 2.1 Identify the various sources for topics of a social science study The topics for a social science study arise from many sources: past studies; television or film; personal experiences; discussions with friends and family; or ideas from a book, magazine, or newspaper. It may be something that arouses your curiosity, something about which you hold deep commitments, or something you believe is wrong and want to change. In short, there is a broad range of what to study. Nonetheless, there are limits. To be appropriate for a research study, the topic needs to generalize about empirically observable patterns that operate in aggregates. Let’s look at these four features briefly: • Generalize A good research topic goes beyond one isolated unique instance. The topic should be something likely to reoccur and that applies to a number of people, places, times, or events. • Patterns A good topic examines something in which we see a regularity, structure, or form that can be summarized, or that describes a collection of complex processes, events, situations, or relationships in a simpler or more condensed way. • Aggregates The topic applies to a collection of people or other units (e.g., families, businesses, schools, hospitals, or neighborhoods). The people/units do not have to be connected to one another or even be aware of the others. There could be as few as 10 or as many as hundreds of millions. • Empirically observable The topic appears in the observable world in a way that we can detect and observe it using our senses (sight, sound, touch, smell) directly or indirectly.

1 For more on the topic of tattoos, see Atkinson (2003, 2004), Caplan (2000), DeMello (2000), Fischer (2002), Dukes and Stein (2011), Heywood et al. (2012), Horne et al. (2007), Kang and Jones (2007). Martin and Dula (2010), and Silver et al. (2011). http://webcache.googleusercontent.com/ search?q=cache: a90tIB8uP9wJ:www.harrisinteractive.com/NewsRoom/ HarrisPolls/tabid/447/mid/1508/articleId/970/ctl/ ReadCustom%2520Default/Default.aspx+&cd=1&hl=en&ct=clnk&gl=us (accessed October 20, 2014)

Planning a Study 23

2.2: The What, Why, and How of a Literature Review 2.2 Analyze sources of research literature

You are probably familiar with the English idiom: “can’t see the forest for the trees.” Being unable to “see the forest” is a way to say that the person cannot recognize general patterns, or see larger context, and usually we need to see the “big picture” for a full understanding.

These features open up many topics but rule out some others. The idea of looking for patterns in aggregates that have some generalizability rules out topics solely focused on particularistic situations (e.g., why your boy/girlfriend dumped you yesterday, why your friend’s little sister hates her third-grade teacher) and a single case (e.g., your own parent’s divorce). Nonetheless, with a little creative thinking we can convert most particular situations into a generalizable topic about patterns (e.g., boyfriends of this type tend to act in certain ways, or children often dislike thirdgrade teachers for one of four reasons). If a person has become so overly involved and focused on the immediate specific details of an event, situation, or relationship, then the bigger picture is invisible or unseen. To continue with the analogy discussed above, a good research topic enables us see the forest, or look for patterns. It does not mean we entirely ignore a particular instance or “tree;” rather, we situate the particular within a larger picture. Another type of topic ruled out for research are things that are impossible to observe, even indirectly such as unicorns, aliens from outer space, or ghosts. We can, however, do research on people’s beliefs about imaginary objects (e.g., what types of people tend to believe in unicorns and why). These and other topics about what is “invisible” (e.g., beliefs, attitudes, opinions, and feelings) can be measured indirectly using questionnaires. In summary, social research enables us to study a great many issues or questions, but it is not appropriate to study everything or anything. Appropriate topics are beyond isolated individualistic situations. They involve seeking systematic patterns across aggregates. Topics need to be about aspects of the visible, real world, i.e., things we can empirically observe or measure, directly or indirectly.

An early step when doing a study is to read what others have already learned about a topic, or to conduct a literature review. Before you head out to search for reports on studies, it is essential to be organized. To prepare a well-written, complete, literature review, you need a search plan and schedule. The ideal literature review is a carefully crafted summary of the recent studies on a topic. It discusses both study findings and how researchers reached the findings with all sources carefully documented. Reading the literature serves the following six functions: 1. It helps you narrow down a broad topic. You can use other studies as a model of how narrowly focused your research question should be. 2. It provides you with examples of possible research designs, concepts, measures, and techniques that you might use. 3. It informs you about what is known about a topic. Past studies teach you the key ideas, factors, terms, and issues surrounding a topic. You may wish to replicate, test, or extend what others have already found. 4. It presents you with examples of what final research reports look like, their parts, form, and writing style. 5. It can help you improve writing skills and learn subtle elements of conducting a good research study. 6. It is often fun and may stimulate your creativity and curiosity. Conducting a literature review builds on the communication system of social science and the assumption that knowledge accumulates. Social research is the collective effort of many people who share their results with one another; we pursue knowledge as members of a community. A core principle of social research is to share all study findings as well as all the details of how we conducted the study. This is why researchers constantly compare, replicate, or criticize other studies. Certain studies may be especially important and a few individual researchers may become famous, but each study is just one small part of the larger, collective process of expanding our common knowledge on a topic (see Figure 2.1). When you conduct a study, you build on what others have done or past studies. In the future, other people can build on the studies you or others have conducted. As Sir Isaac Newton put it, “If I have seen further it is by standing on

24 Chapter 2

Figure 2.1 Advancing Knowledge The advance of knowledge often proceeds in a manner other than a straight line.

at using specific tools. Librarians can help or may offer workshops that teach about the tools. If you have never used a search tool, expect to spend an hour or more to develop the skills to use it adeptly. Plan to locate and scan read articles: The search tools yield a list of articles with your keywords, but they cannot determine the true relevance of the articles for a research question. To decide their relevance, you must scan the articles’ titles, abstracts (to be discussed later), or first few paragraphs. Based on a quick scan-read, you decide what is relevant. If the search tool locates 50 articles, it may take two hours to scan-read all of them and decide their relevance. You may end with 10 relevant, useful studies to read in depth. Allow time to extract the major findings: Reading a scholarly research report is a skill that improves with practice. Most use a sophisticated vocabulary and contain technical information. It takes time to know what to look for. As you read, ask four questions: 1. What was this study really about?

the shoulders of giants.”2 In the larger research process, each new research achievement builds on those that came before it. Making It Practical: A Literature Review Search Plan As with most complex tasks, you can be

most effective with a literature review if you proceed systematically and first devise a plan. A literature review proceeds through six stages that you can plan out. Evaluate resources: How much time can you devote to the search? Do you have access to a college or university library? Do you know what computerized literature search tools are available at the library and how to use them? Do you want to locate a minimum number of studies? Can you easily distinguish an empirical research study from other articles? After answering these questions, you may wish to start preparing a time schedule with benchmarks or self-created deadlines for each step. The more practice you have in published studies, the faster it will go. A firsttime search by a novice can take three or more times longer than one by an experienced person. Select and narrow the topic: You search a specific question, not a general topic. The faster you can focus on a specific research question, the quicker you can proceed. Some people devote days or weeks to focusing on a research question; this is not always necessary. The question you begin with is preliminary because you can adjust and refine it as you learn more from reading past studies. Learn to use literature search tools: Use computerized search tools to search the literature. The tools require you to convert a research question’s central ideas and terms into keywords. It takes time and practice to become skilled 2

Sir Isaac Newton in letter to Robert Horne, Feb. 5, 1676. http://en.wikiquote. org/wiki/Isaac_Newton (accessed November 9, 2014).

2. How did researchers conduct the study (i.e., gather data)? 3. What is the study’s main finding? 4. What are the study’s limitations? You want to extract the essential elements from a research report and write them as notes. Plan how to take notes and record all key source details (discussed later in this chapter). You might spend over an hour reading, re-reading, and taking notes on each relevant article. Final stage—synthesize: Once you have enough articles (because there are no more, you are learning nothing new with additional ones, or you ran out of time), you must pull together and integrate what they said. You might use a few quotes, but you mostly want to paraphrase (put in your own words). Integrating different studies and synthesizing what they really say in combination is a serious thinking and writing task. Often you must return to the full report and re-read for clarification and verification.

WRITING PROMPT Literature Review Plans Consider how timeframes relate to previous papers or reports you prepared and the response or grade you received. Based on this analysis, list the six steps used to create a plan for a literature review. For each step, identify how much time (in days) you would allocate and how this will help to improve your results. The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

Planning a Study 25

2.2.1: Where Do You Find the Research Literature? You can learn about social research in newspapers, in popular magazines, on television or radio broadcasts, and in Internet news reports. They can stimulate thinking about possible study topics or research questions. Yet, they are insufficient for preparing a literature review, because they are not the full, complete study reports. Media stories of research studies are selected and highly condensed summaries that journalists prepare for a general audience. Similarly, textbooks and encyclopedias contain condensed summaries of studies to introduce readers to a topic. These sources are inadequate for conducting a literature review, because they lack essential details required to evaluate a study. To conduct a literature review, you must locate full, complete study reports. Most full reports first appear in specialized periodicals. You will need a plan to find the information you want in them. A periodical or “serial” is any publication (print or electronic) that appears regularly over time (daily, weekly, monthly, quarterly, or annually). It is easy to be confused about the types of periodicals since there are thousands of them, and they come in a vast array of types. With practice, you can learn to distinguish between the following five types: 1. Popularized social science magazines written for an educated general audience 2. Practitioner advice/opinion/news-technical publications, newsletters, and magazines 3. Opinion magazines in which scholars and experts debate and express their views 4. “Mass market” or “trade” newspapers and magazines written to inform and entertain the general public 5. Peer-reviewed scholarly journals in which researchers present the reports of studies

Articles in the first four types of periodicals may discuss study findings, but they lack many essential details about the study. Popularized social science magazines offer the informed public a simplified version of study findings butwithout all the details. Most professions have news/ communication newsletters for working professionals. They may include discussions of a research study or its implications, but they lack full study details. Experts and scholars write articles for serious opinion/public issue magazines about topics on which they may also conduct research (e.g., welfare reform, prison expansion, voter turnout, new marketing techniques). Such publications differ in purpose, look, and scope from scholarly journals. They are an arena in which people debate issues and are not where researchers provide full study reports. Mass-market publications offer the public news, opinion, and entertainment. You can find them at large newsstands, in public libraries, at bookstores, or on websites. They are source for many current events, but they do not contain full reports of research studies. You will be focused on locating scholarly journals because that is where full reports of empirical research appear.

2.2.2: Scholarly Journals The primary place where researchers disseminate information about studies is in scholarly journals. These publications are essential to a literature review because they have the complete reports of research. You rarely find them outside of college and university libraries (or an online service connected with a college library). Many have the term “journal” or “review” in their title but not all do. Most if not all of the articles are reports about original research studies, have been peer reviewed, include a reference or bibliography section that lists sources in detail, and finally are included in an indexing location system that people can access using article search tools. After completing a study and writing a report, researchers present their findings in several forums, the most accessible and respected being a scholarly journal. Almost all articles in scholarly journals are peer reviewed as part of a quality assurance system for research publications. The peer review process works as follows: 1. A researcher prepares the detailed report of a study in a specific format and sends it (in pre-publication form, it is called a manuscript) to the editor of a scholarly journal for consideration. 2. The editor (a respected, experienced researcher with a deep knowledge of the field) examines the manuscript and ensures that it meets minimum standards and is relevant for the journal. 3. The editor contacts two to six respected peer researchers and asks them to be volunteer reviewers. Each reviewer independently reads and evaluates the manuscript. He orshe evaluates a study’s contribution to

26 Chapter 2 advancing knowledge, its originality, the quality of research design and execution, organization, writing quality, use of sources, and the technical correctness of research procedures. 4. Each peer reviewer returns criticisms, comments, and suggestions to the editor. 5. The editor examines evaluations from the reviewers and then decides to do one of three things: a. to accept the manuscript as is and publish it; b. to ask the researcher to revise and resubmit the report for a second round of evaluation; or c. to reject the manuscript. Most scholarly journals use a “blind review” version of the peer review process. It is “blind” because the researcher does not know the identity of the peer reviewers who evaluate the manuscript, and reviewers do not know who conducted the study. A blind review ensures that reviewers judge the manuscript solely on its own merits. Personal relationships with the author and his or her reputation do not influence decisions. Scholarly journals publish less than one-half of the manuscripts they receive for consideration. Some highly prestigious journals publish under 10 percent of the submitted manuscripts. When you read articles in the high-prestige journals, you are seeing the top 10 percent of current research. When a high-prestige, widely read scholarly journal with a low acceptance rate rejects a research report, the researcher usually revises the paper and sends it to a less competitive journal. To a large degree, researchers consider journals with low acceptance rates as the most prestigious. They are most likely to read studies in prestigious journals because they represent the “cream of the crop” and the research community considers them to be top quality, cutting-edge studies. Scholarly journals include more than reports of research studies. They also contain letters to the editor, theoretical essays, book reviews, legal case analysis, and comments on other published studies. Some specialized journals only have book reviews; others only have literature review essays (e.g., Annual Review of Psychology, Annual Review of Nursing Research, Annual Review of Law and Social Science, Annual Review of Public Health) in which a researcher gives a “current state of the field” essay.

WRITING PROMPT Peer Review Describe the pros and cons of using peer review as a quality control tool for research studies. The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

Locating and Evaluating Studies in a Scholarly Journal Except for peer review, there

is no simple “seal of approval” to help you distinguish scholarly journals from other periodicals. Once you find a peer-reviewed scholarly journal, you need to distinguish an empirical research study from the other types of articles. This takes judgment skills or the advice of experienced researchers or professional librarians. The best way to learn to distinguish among types of publications and articles is to read many articles in scholarly journals. You can get copies of most scholarly journal articles through the Internet, but usually there is a fee to access full versions. Your college library may allow you free access (because the library paid the fee) through article databases and search tools. Databases have articles for a limited number of years and from only some scholarly journals. While there are a few online-only journals, over 95 percent are available in print form, so the online databases and search tools are based on printed material. Once you locate a scholarly journal article at a library with an online database and search tool, check that it is an article with study results and not some other type (e.g., opinion essay, book review). It is easier to identify quantitative studies because most have a methods or data section and charts, statistical formulas, and tables of numbers. Qualitative research articles are easy to confuse with theoretical essays, literature review articles, idea-discussion essays, policy recommendations, book reviews, and legal case analyses. Making It Practical: Locating Scholarly Journals Your college library has a section for

scholarly journals and magazines, or, in some cases, it mixes them with books. Look at a map of library facilities or ask a librarian to find this section. Many libraries place the most recent issues, which look like thin paperbacks or thick magazines, in a “current periodicals” section. The library stores them temporarily until it receives all the issues of a volume. Once the library has an entire

Planning a Study 27

Summary Review Table 2.1 Different Types of Periodicals Periodical Type

Example

Authors

Purpose

Strength

Weakness

Peer-reviewed scholarly journal

Social Science Quarterly, American Educational Research Journal, Journal of Applied Psychology, Social Forces

Professors and professional researchers

Report on empirical research studies to professionals and build scientific knowledge

Highest quality, most accurate, and most objective with complete details

Technical, difficult to read, requires background knowledge or training, not always about current issues

Semi-scholarly professional publication

American Prospect, Society, Psychology Today, American Demographics

Professors, professional policy makers, politicians

Disseminate and discuss new findings and their implications for professionals and the educated public

Generally accurate, somewhat easy to read

Lacks full detail and explanation, often includes opinion mixed in with discussion

Practitioner magazine or newsletter

Coach & Athletic Director, Military Police, Retail Merchandiser, Mental Health Weekly

Working professionals and some professors or “experts”

Provide a communication forum for working professionals

Current news and debates on relevant issues

Narrow focus and rarely builds general knowledge

Opinion magazine

Nation, Human Events, Public Interest, Commentary

Professors, professional policy makers, politicians

Present value-based ideas and opinions for professionals and educated public

Carefully written and reasoned

One-sided view and highly value based

Mass market magazines for the public

Time, Esquire, Ebony, Redbook, Forbes, Fortune

Professional journalists and other writers

Entertain, present, and discuss current events for lay public

Easy to read, easy to locate

Often inaccurate and incomplete

volume, it bounds all issues of a volume together and places them in the library’s collection. Libraries shelve scholarly journals from different fields together and post a list of the periodicals to which they subscribe. Most college libraries subscribe to one or more database services that provide access to thousands of journals online. Scholarly journals are published as rarely as once a year or as frequently as weekly. Most appear four to six times a year. For example, Social Science Quarterly appears four times a year, whereas the Annual Review of Nursing Research appears once a year. Librarians created a system for tracking articles in scholarly journals. Every scholarly journal has a year, volume number, and issue number. A journal begins with volume 1, issue 1, and the numbers increase thereafter. The volume usually indicates a year’s worth of articles, and an issue is section of the volume with several articles. Each issue has a table of contents with the title, author(s), and article pages. Most journals number pages by volume, not by issue or article. Page 1 is the first issue of a volume. Not all journals begin their publishing cycle in January. Issue 1 might begin in July or September. Page numbering continues throughout the entire volume and articles on consecutive pages. A single issue can have from 1 to 50 articles, but 6 to 18 articles is typical. Articles vary from 4 to over 50 pages in length. Because one volume is 1

year, a journal issue with volume 52 usually means that it has been operating for 52 years. To locate an article, we use journal name, volume, year, issue, author, article title, and page numbers. These details are the “citation” of a source in the reference section of an article or literature review. Most journal articles have a summary or abstract. A good abstract tells us the topic, research question, method, and findings. There are hundreds of scholarly journals in most fields. Each journal charges an annual subscription fee ($150 to $5,000). For this reason, only the largest research libraries subscribe to most of them. If an article is not available at your local library or in an Internet database, you can obtain a copy from a distant library through an interlibrary loan service, a system by which libraries lend materials to other libraries.

WRITING PROMPT Diverse Publications Of the multiple publication outlets, why are peer-reviewed scholarly journals so important? What are their weaknesses or limitations? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

28 Chapter 2

2.2.3: Sources Other Than ScholarlyJournals There are five other outlets that are sources of research reports: books, dissertations, government reports, policy reports, and presented papers. We will briefly consider each. Books In general, books communicate information, provoke thought, and entertain. Some books report on original research or are collections of research articles. You can find information on these types of books (e.g., title, author, and publisher) in the library’s catalog system. Only college or university libraries have books that report on research. Some publishers, such as university presses, specialize in publishing them. Qualitative types of research are more likely to appear in a book format, as are the results of long, complex studies that may also be published in scholarly journal articles. Because they are not in the article search tool system, it is difficult to locate studies in books. Three types of books can contain research reports:

• Monographs. Contain the details of a long complex study or a set of interconnected studies. • Readers. Contain articles on a topic, original or gathered from journals. Often the editor of the book has modified the research (i.e., shortened and simplified it) to make it easier for nonexperts to read. • Edited collections. A collection of new research reports, articles reprinted from scholarly journals, or a mixture of both on a common topic. Dissertations All graduate students who receive the Ph.D. degree are required to do original research and write the study as a dissertation thesis. Dissertations are in the library of the university that granted the Ph.D. About onethird of dissertation results are published later as books or

articles. Because dissertations report on original research, they can be valuable sources of information. Specialized indexes like the Dissertation Abstracts International (online and print version) list dissertations with their authors, titles, and universities. To get a copy of the dissertation, you must either borrow it via interlibrary loan from the degree-granting university, if that university permits this, or purchase a photocopy of it (see Figure 2.2). Government Reports The U.S. federal government, the governments of other nations, state- or provincial-level governments, the United Nations, and international agencies such as the World Bank all sponsor research studies and publish research reports. Many college and university libraries have some of these documents in their holdings, often in a special government documents section. Most libraries hold only the most frequently requested documents and reports. You can use specialized lists of publications and indexes to search for them, but usually you will need the help of a librarian. Some are also available online. Presented Papers Each year, the professional asso-

ciations in various fields (e.g., criminal justice, education, marketing, nursing, political science, psychology, recreation, sociology) hold annual meetings. At them hundreds of researchers gather to deliver, listen to, or discuss oral reports of recent research, with many also in written form. People who attend can pick up a copy. If you do not attend, you can obtain a meeting program with a list of each paper with its title, author, and author’s place of employment. You can also write directly to the author to request a copy. Policy Reports Research institutes and policy centers

(e.g., Brookings Institute, Rand Corporation, Organization for Economic Cooperation and Development) publish papers and reports. An organization might list its reports

Figure 2.2 Example of Dissertation Abstract Although you may find it difficult to obtain full dissertations, you can learn a great deal about the research study conducted in a dissertation from its abstract alone. Title: An ethnographic case study of two tattoo shops in Milwaukee Author(s): Mertens, Diane K. Degree: Ph.D. Year: 2014 Pages: 244 Institution: University of Wisconsin Advisor: Neuman, W. Lawrence Abstract: This ethnographic case study explored the emic knowledge of tattoo artists and tattooed individuals in Milwaukee, Wisconsin. Five tattoo artists and 12 tattooed individuals who had two or more tattoos were recruited. They were observed in two tattoo shops over the course of six months and completed a 20-item structured interview and unstructured interviews. The goal was to determine the nature of the experience and its meaning to the participants. The participants were also interviewed about their knowledge of the social consequences involved with getting tattooed, and whether or not tattooing has become socially acceptable. Two of the five tattoo artists did not complete the structured interviews, but all participated in an unstructured interview. Lastly, all 17 participants completed a task in which they created a free-list of their recreational activities and rated each on several attributes of leisure. They also sorted a list of leisure activities. The data affirmed that the idea of leisure is strongly peer-centered for the participants. Characteristics of the tattoo experience as described by the participants suggested that for them, tattooing is both a leisure activity and a form of self-expression. The participants discussed the growing acceptance of tattooing in Milwaukee, Wisconsin, as well as variables that make some tattoos socially acceptable or tolerated. They demonstrated that this morally controversial leisure activity has challenged certain social norms but remains associated with other leisure activities often perceived as deviant or morally questionable.

Planning a Study 29

on its website and make copies available electronically. To find all of them, you need to contact the organization and request a list of reports. Sometimes organizations charge a fee for their reports.

In news reports on PISA results, many Americans noticed that U.S. high school students regularly rank in the bottom onethird internationally. While many recognize problems with K-12 schools, they believe that U.S. colleges and universities are world class. Piaac was the first systematic international comparison on adult literacy and quantitative skills. It showed that U.S. college-educated adults fare no better than the U.S. highschool students when compared similarly educated people to other countries. For example, among 16- to 29-year-olds with a 4-year college (bachelor’s) degree, the United States ranked below the average of other nations on quantitative reasoning skills. Adults in 15 countries scored higher than the Americans did. Americans were eighth from the bottom in literacy. The study examined people at different ages, and found that recent college graduates are the same as older ones. Among people aged 16 to 29 with a bachelor’s degree or more, the United States ranked 16th out of 24 in numeracy. The evidence shows that, on average, American colleges are not the best in the world. As summarized in, “Time for the U.S. to Reskill? What The Survey of Adult Skills Says,” OECD 2013 (p. 20 ), “Of the three skills domains, and comparing the U.S. with other countries, the U.S. performance is weak on literacy, very poor on numeracy, but only a little worse than average on problem solving in technology rich environments.” As with all research studies, the report provides details of how the study was conducted, including example questions, how the people were sampled, and how the tests were given. For example, in the United States a national random sample of 5,010 people was selected using census data and people were individually interviewed.3

Example Study Are U.S. Colleges the World’s Best? While scholarly journal articles are the heart of a literature review, some major research organizations produce excellent policy reports. The Organization for Economic Co-operation and Development (OECD) is a major international economic organization headquartered in Paris. Thirty-four countries are members (Europe along with the United States, Canada, Australia, Japan, and South Korea). The OECD was founded in 1961 to promote social-economic progress and world trade. It grew out of the Organization for European Economic Co-operation founded in 1948 as part of the U.S. Marshal Plan. The OECD collects a vast range of statistical data on social, economic, environmental, and other indicators from member countries, and conducts policyrelated research. The OECD publishes a magazine, books, reports, statistics, working papers, and reference materials. In the area of education research, the OECD is well-known for the Program for International Student Assessment (PISA). It is a worldwide study of 15-year-old school pupils’ performance on mathematics, science, and reading that began in 2000 and has been repeated every 3 years since. The goal of PISA is to improve education policies. It is huge: 510,000 students in 65 countries between the ages of 15 years 3 months and 16 years 2 months participated in the 2012 PISA, completing the same 2-hour test. PISA results enable national leaders to see how their students compare relative to similarly educated young people in other countries on the same skills and knowledge. Another more recent OECD study was the Program for the International Assessment of Adult Competencies (known as Piaac). It tested 166,000 adults ages 16 to 65 in the OECD countries in 2011 and 2012. The study examined three areas: 1. literacy: decode written words and sentences and comprehend, interpret, and evaluate complex written text 2. numeracy: manage a situation or solve a problem in a real context by responding to mathematical content/information/ ideas presented in multiple ways 3. problem solving in technology-rich environments: solve problems for personal, work, and civic purposes by setting up appropriate goals, and accessing and making use of information through computers. Results in each area were scored on a 0- to 500-point scale. Because the Piaac test takers were adults, the test asked them to apply skills in real-world contexts. For example, they read a news article and an e-mail, each describing an innovative method to improve drinking water quality in Africa. Next, the test asked them to identify the sentence in each document that described a criticism common to both inventions.

2.3: Six-Step Literature Review Process 2.3 Identify the six-steps of the literature review process

2.3.1: Refine the Topic Your search begins with a research question, not a topic. It is impossible to examine a broad topic with any depth or seriousness. A topic such as “divorce” or “crime” or “patient care” is too broad. Narrow the topic to something, such as “the stability of families with stepchildren” or “economic inequality and crime rates across 50 nations” or “long-term care of elderly patients with a heart condition.” You further narrow this into a research question by adding conditions and limiting the range of cases/situations to which it applies. 3

See the following three links (all accessed November 1, 2015) for details on the OECD adults skills survey. http://skills.oecd.org/documents/Survey_of_ Adult_Skills_US.pdf; http://skills.oecd.org/skillsoutlook.html; http://www. oecd.org/site/piaac/

30 Chapter 2

Figure 2.3 The Six-Step Process The process of conducting a literature review for a study 1. Refine the Topic 2. Design Your Search 3. Locate the Research Reports 4. Read and Take Notes on the Reports Found 5. Organize Notes, Synthesize, and Write the Review 6. Create a Reference List

2.3.2: Design Your Search To prepare a good literature review, you will need to make a series of decisions. If you make them in a thoughtful and intentional manner, you will avoid frustration and be able to prepare a successful review. 1. Decide on the review’s extensiveness by fixing parameters for your search: how much time you can devote to it, how many years back you will look, how many reports you will examine, how many libraries you can visit, whether you will look at both articles and books or only articles, and so forth. Expect to make multiple visits to a library (online or physically). If you have 15 hours to do a literature review, do not expect to locate and include more than 10 to 12 research reports. 2. Decide which article search tools to use. 3. Decide how you will record the bibliographic information for each source and how to take notes (e.g., in a notebook, on 3 × 5 cards, in a MS Word document).

2.3.3: Locate the Research Reports Searching varies by type of research report (article, book, or dissertation). Scholarly journal articles are usually the most valuable and least time-consuming to find. Locating a copy of the reports can take time. After you find a complete copy, you must read and take notes on it. Using Article Search Tools to Locate Studies in Scholarly Journals Most studies appear in

scholarly journals. However, there are hundreds of journals, most go back many decades, and each may have over a hundred articles per year. Luckily, article databases and search tools (sometimes called indexes or research literature services) make the task easier. Until about 15 years ago, a search for scholarly journal articles required many hours of reading through specialized periodicals available only in college libraries. Today, you can use such article search tools via online services connected to article databases. There are about 50 of

them; some are general and some specialized. Most have articles from scholarly journals, although a few include papers presented at professional conferences, dissertations, and policy reports. Libraries pay fees to access the services. To locate an article, you use keywords or search by author name. Many article search tools are titled abstracts or indexes (e.g., Psychological Abstracts, Social Sciences Index, and Gerontological Abstracts). For education-related topics, the Educational Resources Information Center (ERIC) system is valuable; for medicine, MEDLINE is widely used. The name abstract is from an article’s “abstract” or a short summary, usually at the beginning of an article. It does not contain all the findings or details of a study, but you can use abstracts to screen articles for relevance. Some are highly organized and contain specific details, whereas others are less structured (see Figure 2.4). Develop a List of Keywords for Searching

You can search by author, subject, or keyword. Search tool subjects are limited to a few popular ones. Unless you know that a specific researcher did a study or you are at a late stage in searching, you probably will not use the author name for a search. For most searches, you will use keywords. You must develop key terms for your research question. Are illegal drugs more common in urban than rural high schools?

This search might include keywords, such as drug abuse, substance abuse, drug laws, and illegal drugs. You can combine these with keywords as high schools, high school students, rural schools, urban schools, and secondary schools. You should consider several synonyms for keywords. For example, a search with the keyword homicide may not find articles that use the word murder, so you want to try both. Most search tools look for a keyword in a title or the abstract. Often you can use multiple keywords using the connectors or or and. If you choose very broad keywords or many connected by or, you will get a huge number of irrelevant articles. If you use narrow keywords or several keywords connected with and, you may find zero articles. To learn what works best, try experimenting with alternatives. While a few free services are available to the public (i.e., Google Scholar), the best search tools are restricted to the students or staff through a college library. Many article search tools only provide author, titles— and abstract, not the full text of an article. Once you get search results, you can scan the titles and abstracts for relevant articles. Articles usually contain reference sections with leads to additional sources. Reading the reference section can point to relevant sources that you might not find using an article search tool.

Planning a Study 31

Figure 2.4 Example of Two Abstracts from Scholarly Articles Article abstracts provide important summary information and help you determine the relevance of an article before reading it. Example of Highly Structured Abstract Title: Demographic Correlates of a Representative Sample of Men and Women Who Have Ever Been Tattooed Abstract Objective: Despite an increase in the popularity of persons getting tattooed, we know little about the characteristics of adult men and women who have ever been tattooed. We investigated demographic and behavioral correlates of adults who have been tattooed in a representative sample. Methods: Computerassisted telephone interviews (CATIs) were completed by a representative random sample of 9,337 men and women aged 16–65 years in Canada. Results: A total of 15.8% of respondents had ever been tattooed, and 3.1% of respondents within in the year immediately prior to the interview. More men than women reported getting a tattoo; nonetheless, the highest tattooing rates were among women in their 20s (31.1%). The 20–39 age group was most likely to have been tattooed. Men without a college degree and women who were cohabitating also showed higher tattoo rates. Tattooing was correlated with several risktaking behaviors: binge drinking, smoking tobacco and marijuana, and having a large number of sexual partners. It was also correlated with having being diagnosed for anxiety or depression. Conclusions: As tattooing has grown in popularity, especially among young adults, it remains a marker for people who engage in other high-risk behaviors.

Title: College Student Attitudes About Tattoos

Example of an Unstructured Article Abstract Abstract

We test models of negative attitudes toward tattooed persons and attitudes toward getting a tattoo in the future among 205 tattooed and 307 nontattooed college students. People who have friends and family members with a tattoo reduce negative attitudes toward tattooing. In addition to having known a person with a tattoo as well as having gotten one tattoo in the past predicts a person’s attitude toward future tattooing. Looking at data for 194 tattooed respondents, we find that those who have multiple tattoos are most committed to their tattoos and least interested in their removal. Having multiple tattoos is also associated with a perception that tattooed persons suffer from negative prejudice from the nontattooed population. Persons with visible and large tattoos do not differ from persons with covered and small tattoos regarding degree of tattoo commitment, but persons with large and visible tattoos most likely to say that tattooed persons suffer from prejudice.

Example Study Sexual Harassment Literature Search Here is a search I conducted using article search tools. My general topic was “sexual harassment.” I narrowed the topic to “sexual harassment of female college students.” I looked for peer-reviewed articles in English published from 2005 to 2012. I started with the article search tool called “EBSCO-Host Academic Search Complete.” This tool examines 1,500 peerreviewed journals for all academic fields back to the late 1990s for most journals. Other article search tools contain information from different journals or cover different time spans. My first keywords were sexual harassment and university. The search located 614 articles. I found that many were not about college students. The search tool had picked up the word university from where the author worked. A search with sexual harassment and student yielded 191 articles. Some were about high school students, and some were about males being harassed by females, but many were relevant. I narrowed my search further to sexual harassment and college female and got 64 articles. Not all were on my topic of interest. I noticed that “gender-related harassment” and “unwanted sexual encounters” appeared in some articles. This gave me the idea to use them as alternative keywords for sexual harassment. I next used three other article search tools with the same restrictions and used the same keywords as before. I found that the four article search tools located many of the same articles, but often one search tool found articles not located by other search tools. This shows that it is always best to use more than one search tool (see Table 2.2).

Table 2.2 Number of Articles Found with Article Search Tools (peer-reviewed articles, 2005–2012) Keywords Used

EBSCOHost

JSTOR

Sociological Abstracts

Citation Index

Sexual harassment & University

614

192

240

114

Sexual harassment & student

191

273

Sexual harassment & college & female

Use Books and Other Outlets to Find Studies It is very difficult to find studies in books. The

subject lists in library catalog systems are broad and not very useful. Moreover, they list only books in a particular library system. Professional librarians can help you locate books from other libraries. There is no sure-fire way to locate relevant books. Use multiple search methods, including a look at journals that contain book reviews and the bibliographies of articles. Locating studies in other outlets like government documents, Ph.D. dissertations, policy reports, and presented papers is far more difficult and time-consuming. A specific study might be highly relevant to your question, but few beginning researchers have the time or skills to search other outlets systematically.

32 Chapter 2

2.3.4: Read and Take Notes on the Reports Found It is easy to feel overwhelmed as you gather studies. To help, develop a system for taking notes that works for you. The old-fashioned approach of writing individual notes on index cards is one way you can organize your information. Once you have completed your literature review/note taking, it is easy to shift and sort the note cards, place them in piles, and then look for connections among them. Many people also like to gather photocopies or printed versions of many articles and then use word-processing software to organize notes in electronic documents. Photocopying all the relevant articles can save time in recording notes, and you will have the entire report handy if you need to re-read sections. It is important to note all citation details (title, page[s], volume, etc.) and to be aware of copyright laws (most national copyright laws permit photocopying for personal research use). Unfortunately, sorting a large pile of photocopied articles is not easy, especially sorting several specific sections of the same article into different piles for analysis. Making It Practical: How to Read a Scholarly Journal Article While you can learn a lot from sim-

ply picking up an article and reading it, here are some strategies that can help you become efficient at the task and get the most from scholarly articles. 1. Start with a clear purpose in mind. Are you reading to gain background knowledge on a broad topic or to find information for a very specific research question? 2. First read the title and abstract to check an article’s relevance and general content. Next, quickly scan subheadings and the introduction and conclusion sections. 3. Form a quick mental image of the article’s topic, major findings, method, and general conclusion. 4. Consider your own opinion about or bias toward the topic, the method, the publication source. How might your own opinion color how you read and evaluate the study? Check any biases so that you can evaluate the article objectively. 5. Marshal external knowledge. What do you know about the topic and the research methods used? 6. As you read the entire article, evaluate. What errors might be present? Does the discussion of findings follow the data? Is the article’s conclusion consistent with its approach? 7. Summarize what you found: Prepare your own abstract, including the topic, the research methods

used, and the main findings. Next, take detailed notes, including quotations from the article and page numbers for use as quotes or main ideas. 8. Always review the reference section or bibliography for new sources and ideas.

WRITING PROMPT Reading a Scholarly Journal Article What benefits would you gain by skimming the introduction and conclusion of an article before reading the center? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

Create Source and Content Files Researchers

use several strategies in literature searches and article reading. My strategy is to create two kinds of files: a source file and a content file. I recommend that you adopt a similar strategy in order to keep track of all details on sources and not have content notes without references. In the source file, I record all the bibliographic information for each source, even if I may not use some of the articles. This includes journal name, full article title, date, volume and issue number, starting and ending page numbers, and the full names of all authors. It is easier to erase an unused source than to try to locate bibliographic information later when needed. The source file allows me to create a list of complete references very quickly. I put substantive details in the content file. This file contains major findings, details of methodology (such as whether it was a survey or experiment, the number of participants), definitions of major concepts, how concepts were measured, and interesting quotes. When quoting, I always record the specific page number(s) on which the quote appears. On each content note, I put the author’s last name and the publication year. This allows me to link multiple cards or computer notes in the content file to a specific source in the source file (see Figure2.5). What to Record in Notes It is best to be consis-

tent when writing notes—use all computer files or all note cards of the same size. As you decide what to record about an article, book, or other source, it is better to err by writing a little too much rather than not enough (see Figure 2.6).

Planning a Study 33

Figure 2.5 Web Page from EBSCO Host Academic Elite Advanced Search Although each search tool is a little different, the types of information you will need when searching a database of scholarly articles are similar. In addition, your speed and efficiency at finding relevant articles can improve with practice.

Tips for Recording Notes Reading research

reports with a critical eye is a skill that takes time and practice to develop. Despite a peer review procedure, errors and sloppy logic can slip into research reports. Sometimes titles, abstracts, and the introduction are misleading; they may not fully explain the study’s method and results. A

good article follows a logical progression, and all its parts fit together. Weak articles make huge leaps in logic or omit transitional steps. As you read for details and take notes, try to develop a mental image of how the researchers conducted the study. Reading many studies can expand your research design

34 Chapter 2

Figure 2.6 Example of Notes on an Article Your notes should answer the following questions ENTRY IN SOURCE FILE Guéguen, Nicolas. 2012. “Tattoos, piercings, and sexual activity.” Social Behavior and Personality, Vol. 40, Issue 9, pp. 1543–1548. CONTENT FILE Guéguen, Nicolas. 2012 Background: Many studies have examined the characteristics of people with tattoos or body piercings in the United States. U.S. studies found college students with piercings engage in risky behaviors, use illegal drugs recreationally, and engage in binge drinking, more than students who do not have piercings. Women with piercings were more likely to have premarital sex, but no difference was found between men with or without piercings found regarding sexual behavior. Similar patterns hold for tattoos. U.S. adolescents with tattoos report higher levels of substance use, violent behavior, sexual intercourse, and school problems than were adolescents without tattoos. Questions/expectations: A large majority of studies on tattooing and piercing have been conducted in the United States, with only a few studies in other countries (Canada, Switzerland). Tattooing or piercing among young adults has not been studied in France, and are new behaviors in France, beginning later than in the United States. This study focuses on sexual behavior, and the effect of a combination of tattooing and piercing on the age of the first sexual experience. The author hypothesized that respondents with tattoos and/or piercings would report earlier sexual activity than would respondents without piercings or tattoos. Definitions and measures: In a face-to-face survey, respondents were asked three questions. They were asked about age of first sexual activity, “How old were you when you had sexual intercourse for the first time?” Tattoos and body piercings were measured by asking how many piercings they had (ranging from 0 to 5+) and how many tattoos they had (ranging from 0 to 5+). Research design: The data were gathered by 104 (46 males and 58 females) research assistants who were undergraduate students studying business management. Each assistant was neatly dressed in a traditional manner for young people (jeans/sneakers/T-shirt) and on sunny days on a college campus approached 20 passerby students of the same gender as himself/herself. Data or research participants: The participants were 2,080 students (1,160 females and 920 males) enrolled in four public universities in the west of France (Brittany). All were French and Caucasian, and their mean age was 20.84 years. Findings: Among all respondents, 24 percent had either a tattoo or a piercing. Both males and females without piercings or tattoos began sexual activity later than males and females who had only tattoos, only piercings, or both tattoos and piercings. Persons with both tattoos and piercings began sexual activity younger than who had only piercings or only tattoos. No significant gender differences were found. Among male and female students without tattoos or piercings, the onset of sexual activity began one to two years later (average age 17.5) than students who had tattoos (age 16.6), piercings (age 15.8), or both (age 15.4).

skills. If you read a study in which the authors were disorganized or did not clearly provide all the details, you will recognize the importance of good organization and specifying all details. You may encounter unfamiliar terms, new theoretical ideas, advanced technical vocabulary or sophisticated statistical charts, graphs, and results beyond your background. This is because professional researchers are the primary audience for research reports. The technical terms and results communicate important information to this audience. Do not be overly concerned if you cannot follow everything. As a novice researcher and consumer of studies, you should not expect to have the sophisticated knowledge of an expert researcher. Be prepared to read an article more than once. A lack of knowledge might prevent you from fully evaluating all aspects of a study, but you can still learn from and build on sophisticated studies and improve and expand your understanding over time.

2.3.5: Organize Notes, Synthesize, and Write the Review Synthesizing and discussing findings with clear writing is the most difficult step in preparing a literature review. After gathering information, organize specific findings to create a map of how they fit together. Your organizing method depends on the purpose of the review. Usually, it is

best to organize findings around your research question or around a few core, shared findings. Most professionals try several organizing schemes before they settle on a final one. Organizing your notes is a skill that improves with practice. Some people place notes into several piles, each representing a common theme. Others draw charts or diagrams to show the connections among different findings. Others create lists of how the many study findings agree and disagree. Organizing notes is a process. Often you will find that some references and notes are no longer relevant, and you will discard them. You may discover gaps or new areas that you did not consider previously. This may require return visits to the library to refine your search. Next you will want to synthesize or blend the findings, methods, or statements from separate studies and end up with a coherent whole. Instead of just listing summaries of articles, one after the other, your literature review will show how the studies fit together as one integrated picture. Your goal is to produce a well-written, compact document that clearly summarizes what many studies say about a research question. A good literature review is a neutral summary description (do not include your personal opinion or conjecture) that communicates its purpose to the reader through its organization. You want to organize common findings or arguments together, address the most important ideas first, logically link findings,

Planning a Study 35

andnote discrepancies or weaknesses. The rules of good writing (e.g., clear organizational structure, an introduction and conclusion, transitions between sections, etc.) apply. What a Weak Literature Review Looks Like Compare Your Thoughts Sexual harassment has many consequences. Adams, Kottke, and Padgitt (1983) found that some women students said they avoided taking a class or working with certain professors because of the risk of harassment. They also found that men and women students reacted differently. Their research was a survey of 1,000 men and women graduate and undergraduate students. Benson and Thomson’s study in Social Problems (1982) lists many problems created by sexual harassment. In their excellent book The Lecherous Professor, Dziech and Weiner (1990) give a long list of difficulties that victims have suffered. Researchers study the topic in different ways. Hunter and McClelland (1991) conducted a study of undergraduates at a small liberal arts college. They had a sample of 300 students, and students were given multiple vignettes that varied by the reaction of the victim and the situation. Jaschik and Fretz (1991) showed 90 women students at a mideastern university a videotape with a classic example of sexual harassment by a teaching assistant. Before it was labeled as sexual harassment, few women called it that. When asked whether it was sexual harassment, 98 percent agreed. Weber-Burdin and Rossi (1982) replicated a previous study on sexual harassment, but they used students at the University of Massachusetts. They had 59 students rate 40 hypothetical situations. Reilley, Carpenter, Dull, and Bartlett (1982) conducted a study of 250 female and 150male undergraduates at the University of California at Santa Barbara. They also had a sample of 52 faculty. Both samples completed a questionnaire in which respondents were presented vignettes of sexual-harassing situations that they were to rate. Popovich et al. (1986) created a nine-item scale of sexual harassment. They studied 209undergraduates at a medium-sized university in groups of 15 to 25. They found disagreement and confusion among students. What a Better Literature Review Looks Like Compare Your Thoughts The victims of sexual harassment suffer a range of consequences, from lowered self-esteem and loss of self-confidence to withdrawal from social interaction, changed career goals, and depression (Adams, Kottke, and Padgitt, 1983; Benson and Thomson, 1982; Dziech and Weiner, 1990). For example, Adams, Kottke, and Padgitt (1983) noted that 13 percent of women students said they avoided taking a class or working with certain professors because of the risk of harassment.

Research into campus sexual harassment has taken several approaches. In addition to survey research, many have experimented with vignettes or presented hypothetical scenarios (Hunter and McClelland, 1991; Jaschik and Fretz, 1991; Popovich et al., 1987; Reilley, Carpenter, Dull, and Barlett, 1982; Rossi and Anderson, 1982; Valentine-French and Radtke, 1989; Weber-Burdin and Rossi, 1982). Victim verbal responses and situational factors appear to affect whether observers label a behavior as harassment. There is confusion over the application of a sexual harassment label for inappropriate behavior. For example, Jaschik and Fretz (1991) found that only 3 percent of the women students shown a videotape with a classic example of sexual harassment by a teaching assistant initially labeled it as sexual harassment. Instead, they called it “sexist,” “rude,” “unprofessional,” or “demeaning.” When asked whether it was sexual harassment, 98 percent agreed. Roscoe et al. (1987) reported similar labeling difficulties.

2.3.6: Create the Reference List The last step is to create a works cited list or bibliography. A reference list, also called a works cited list, is a list of sources to which you referred. They differ from a bibliography, which is a list of all the materials that you consulted, whether or not they are cited. For a literature review, use a reference list of the sources you discussed in the review. How you indicate sources in the text of your review and in the reference list is very important. There are several format styles, each with separate rules. Different fields (e.g., psychology, history) use specific formats. In the text of a review itself, an in-text or parenthetical citation format is most common. It has the author’s or authors’ last name and year of publication for a general statement, with page numbers for specific details or quotes. The order and format of source citation information can vary greatly. You need to learn which format style an instructor or publication requires. The citation format style precisely specifies how to organize details of source information in a reference list. Two reference books of typical styles used in social science publications are the Chicago Manual of Style and the American Psychological Association (APA) Publication Manual (see Figure 2.7). The Role of the Internet in Research The Internet has revolutionized research. Only 18 years ago, fewer people used it. Today, researchers and others use it regularly to review the literature, to communicate with others, and to search for information. Yet, it has been a mixed blessing and not proved to be the panacea that some people first thought it might be. While the Internet is an important way to find information, it remains just one tool among others. It is a powerful supplement rather than a total replacement for traditional library

36 Chapter 2

Figure 2.7 Different Reference Citations for a Book and Journal Article Reference Citations Style

Book with One Author in Reference List

MLA

DeMello, Margo. Bodies of Inscription: A Cultural History of the Modern Tattoo Community. Durham NC: Duke University Press, 2000.

ASA

DeMello, Margo. 2000. Bodies of Inscription: A Cultural History of the Modern Tattoo Community. Durham NC: Duke University Press.

APA

DeMello, M. (2000). Bodies of Inscription: A Cultural History of the Modern Tattoo Community. Durham NC: Duke University Press.

Chicago

Same as MLA for the arts, literature, or history. Same as ASA for science fields.

Style

Journal Article with Two Authors in Reference List (Journal Pagination by Volume)

MLA

Dukes, Richard A. and Judith A. Stein. “Ink and Holes: Correlates and Predictive Associations of Body Modifications Among Adolescents.” Youth and Society 43 (2011) 1547–1569.

ASA

Dukes, Richard A. and Judith A. Stein. 2011. “Ink and holes: Correlates and predictive associations of body modifications among adolescents.” Youth and Society 43:1547–69.

APA

Dukes, R.A. & Stein, J.A. (2011). Ink and holes: Correlates and predictive associations of body modifications among adolescents. Youth and Society 43, 1547–69.

Chicago

Same as APA for science fields. Same as MLA for arts, literature and history.

Others

Dukes, Richard A. and Judith A. Stein. 2011. “Ink and holes: Correlates and predictive associations of body modifications among adolescents.” Youth & Soc. 43:1547-69. Dukes, Richard A. and Stein, Judith A. (2011). Ink and holes: Correlates and predictive associations of body modifications among adolescents. Youth & Society 43 (4):1547–1569. Dukes, Richard A. and Judith A. Stein. 2011. “Ink and holes: Correlates and predictive associations of body modifications among adolescents.” Youth & Society 43 (4):1547–1569. Dukes, R.A. & Stein, J.A. (2011). Ink and holes: Correlates and predictive associations of body modifications among adolescents. Youth & Society 43:1547–1569. Richard A. Dukes and Judith A. Stein. Ink and holes: Correlates and predictive associations of body modifications among adolescents. Youth & Society. Vol. 43, no. 4 (2011):1547–1569.

Format styles for sources in a reference list.

research. On the positive side, it is easy, fast, and cheap. On the negative side, it lacks quality control and a lot of top quality information is not free. Unlike standard academic publications, there is no peer review process, or any review at all of what appears on it. Anyone can put almost anything on a website, even if it is of poor quality, undocumented, highly biased, totally made up, or fraudulent. We can quickly find useful information using a search engine (e.g., Google, Bing, Yahoo), yet the speed and ease of getting answers distracts from the fact that results we get are sometime misleading or inaccurate. Anyone who believes he or she can just “google” a question on the Internet and instantly get all the high-quality information available is very mistaken. Many excellent resource materials for social research are not available on the Internet. Most of the best information is available only through library subscription services. Contrary to early expectations, the Internet has not made all information free and accessible to everyone. One problem is that Internet sources can be “unstable” and difficult to document. After you conduct a search on the Internet and locate websites, note the specific URL (uniform record locater) or “address” (usually it starts http://) where it resides and the date you saw it. This address refers to an electronic file sitting in a computer somewhere. If the computer file moves, it may not be at the same address two days later. Unlike a journal article stored on a shelf in

undreds of libraries for many decades and available for h anyone to read, websites can quickly vanish. This means it may not be possible to check web references easily, verify a quote, or go back to original materials. Because it is easy to copy, modify, or distort and then reproduce copies of a web source, you may find multiple variations on the same material—some are fraudulent, while others are designed to be intentionally misleading. Many college libraries have guides to help students navigate the Internet. One simple helpful site to evaluate websites is EDUSCAPES.COM

Here are a few basic rules to help you locate the best sites on the Internet—ones that have useful and truthful information are the following: (1) Note the source: Sources that originate at universities, research institutes, or government agencies usually are most trustworthy. Look at the web address (URL) and its domain (.edu, .gov, .org, .net, .com). (2) Check verification information. Many websites fail to provide complete information to make verification and citation easy. The best sources provide complete information about the author, date, location, and so on. (3) Is the information current? What is the copyright date, and when was the page originally posted? When was it last updated? Be cautious of very outdated web pages. (4) Crosscheck: If you are suspicious, find three independent resources confirming each piece of questionable data.

Planning a Study 37

WRITING PROMPT Good versus Bad Web Sites When using Google or a similar search engine, how do you distinguish between trustworthy and fraudulent websites? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

2.4: How Do We Focus the Research Question? 2.4 Evaluate what makes a good research question using the inductive and the deductive approaches By now, you know that before you conduct a literature review or develop a proposal to do a study, you need to focus on a research question to make it much narrower than a topic. How you do this varies depending on whether your study follows one of two general approaches to research— an inductive or deductive approach (see Figure 2.8).

Figure 2.8 Inductive and Deductive Approach Different starting points to connect ideas with data D e d u c t i v e

IDEAS

Observed data

I n d u c t i v e

If you adopt an inductive approach, you begin with the evidence and then slowly build toward generalizations, patterns, or summary ideas. You move from examining many specific details toward developing a few general ideas about them. If you adopt a deductive approach, you start with a summary idea, a general pattern, or an “educated guess” of what you think might occur. You then move toward locating specific, observable evidence that allows you to test or verify the idea, pattern, or educated guess. Many studies are not strictly inductive or deductive, but most emphasize one approach over the other. There is no rigid rule; however, the type of data and purpose of a study are a guide. Most often, an inductive approach goes with qualitative data and the deductive approach with quantitative data. Most exploratory studies use the inductive

approach, explanatory studies use the deductive approach, and descriptive studies use both. If you follow a deductive approach with quantitative data, you will need to devote significant time early in a research study to specifying the research question precisely and planning most study details. Once you design the study, the other steps (i.e., collecting and analyzing data) can proceed in a straightforward way. By contrast, if you follow an inductive approach with qualitative data, you can devote less time to developing a research question and planning details in advance. However, you must spend far more time and effort during the subsequent stages of a study (i.e., collecting and analyzing data). In all types of research, you need to narrow a topic into a research question. It takes time to develop the judgment skills to determine whether a deductive-quantitative or an inductivequalitative study is best for a research question. Three things will help you identify the most effective type: • Reading many past studies • Learning to appreciate the specific features of qualitative and quantitative data • Understanding how to use various research techniques and recognizing their strengths and limitations You need a research question to design a study, although you can adjust the research question as the study progresses. If you adopt a deductive-quantitative approach, you must be very specific before you can proceed. If you adopt the inductive-qualitative approach, you can begin with a general research question that you narrow further during the data collection process.

2.4.1: Narrowing a Topic into a Research Question There are several strategies to focus a topic into a question, and you can combine more than one of them. 1. Examine the research literature. From a literature review, you may decide to replicate a past study exactly or with slight variations. You might explore unexpected findings discovered in past research. In many reports, authors suggest future research issues. You can also extend an existing explanation to a new topic or setting. For example, a study of workplace relations in a hospital found that nurses and other staff cooperate and are more productive under certain arrangements. You might conduct a study to see whether the same arrangements have the same outcome in a nonmedical setting (e.g., a large legal office). You can examine the intervening process. For example, a study found that increased police foot patrols produced morecalls to police when trouble occurred. You might

38 Chapter 2 examine exactly how this occurred—did the foot patrols increase familiarity, feelings of trust, and belief in the honesty in police, and in turn, these factors increased calls to police? 2. Talk over ideas with others. Ask people who are knowledgeable about the topic for questions. It is useful to seek out those who hold opinions that differ from yours on the topic and discuss possible research questions with them. A research question might help resolve different positions on an issue. 3. Specify the context. Apply a finding or topic a specific time, society, geographic unit, or category of people. Perhaps you want to study divorce. Your research question might examine divorce in a particular era (divorce in the 1960s vs. the early 2010s), location (Southern vs. Northern Mexico), or category of people (among indigenous rural people vs. urban majority Mexicans). 4. Specify the purpose of your study. Do you want the study to be an exploratory, descriptive, explanatory, or evaluation study? Tailor your research question to one or another purpose.

2.4.2: Comparing Good and Not-So-Good Research Questions Not-So-Good Research Questions: • Not empirically testable, nonscientific questions. Should abortion be legal? Is it right to have capital punishment? • General topics, not research questions. Treatment of alcohol and drug abuse. Sexuality and aging. • A set of variables, not questions. Capital punishment and racial discrimination. Urban decay and gangs. • Too vague, ambiguous. Do police affect delinquency? What can be done to prevent child abuse? How does poverty affect children? Good Research Questions • Exploratory questions. Has the actual incidence of child abuse changed in California in the past 10 years? Is a new type of abuse occurring? • Descriptive questions. Is child abuse, violent or sexual, more common in families that have experienced a divorce than in intact, never-divorced families? Are children raised in impoverished households more likely to have medical, learning, and social-emotional adjustment difficulties than children raised in nonimpoverished households? • Explanatory Questions. Does the emotional instability created by experiencing a divorce increase the chances that divorced parents will physically abuse their children? Is a lack of sufficient funds for preventive treatment a major cause of more serious

medical problems among children raised in impoverished families? • Evaluative Questions. Has the new patient tracking system produced higher satisfaction ratings? Does the automatic arrest of abusive males in domestic violence calls to police reduce later violent domestic abuse incidents? Will third-grade children’s readings scores show larger improvements under the new program than the existing reading program? Does calling customers to remind them of their appointment a day in advance reduce the percentage of customer noshows at the service center?

2.5: How to Design a Study for a Research Proposal 2.5 Analyze the various processes involved in designing a study for a research proposal The research proposal is a written document in which you review the literature and provide a detailed plan for how to carry out a study. The proposal will vary depending on whether the approach and data are deductive-quantitative or inductive-qualitative. A mixed approach using both types of data is also possible and has advantages. In all empirical research studies, you systematically collect and analyze data. If data are qualitative in the form of words, sentences, photos, and symbols, the techniques to collect and analyze data will differ from a situation when the data are in the form of numbers. Techniques appropriate for qualitative data may be inappropriate for quantitative data and vice versa. One data form is not always superior, and each has specific strengths. Ideally, a research question and a form of data will combine in a way that utilizes that data form’s strengths.

2.5.1: When and How Do You Focus the Research Question? The research question directs you to the particular data you will need to gather. For example, your question asks you to collect data about the attendance of students in specific grades, and to measure their learning (using test scores, course grades, teacher notes) in specific subject areas. If you intend to gather qualitative data, you proceed slowly and focus on a research question after you gather data. You will need a topic, such as how do high school students actually learn course material, but do not have to focus on a specific question at first. You may spend many hours gathering data by talking with, observing, and interacting with students, teachers, and parents. After examining the data, you develop a specific question

Planning a Study 39

to direct later stages of data collection. The research question emerges slowly, in an ongoing, interactive process of data gathering. For any type of research, you need to be realistic and recognize limitations as you develop the research question. Making It Practical: Practical Limitations on Study Design Designing a perfect research project

is an interesting academic exercise, but to carry out a study, practical limitations shape its design. Ask yourself the following questions: • How much time can I devote to the study? • What is the cost of conducting the study, and can I get the required funds? • Can I gain access to needed resources, people, and locations? • Do I have required approval of authorities or officials? • Have I addressed all ethical concerns? • Do I have the needed skills, expertise, and knowledge? If you can devote 10 hours a week for 5 weeks to a study but answering a research question requires a 5-year study, you need to narrow the research question. Estimating the time required for a study is difficult because the research question, research techniques, and the amount and types of data collected all have an impact. Consulting with experienced researchers is the best way to get a good estimate. As in most areas of life, the highest-quality research usually takes the most time. Access to resources is a common limitation. Beyond money and time, required resources include the expertise of others, special equipment, and information. For example, you have a research question about burglary rates and family income in the 20 largest nations. This is almost impossible to answer because data on burglary and income are not available for most countries. Some questions require the approval of authorities (e.g., to see medical records) or involve violating basic ethical principles. Your expertise or lack of it can be a limitation. Answering some research questions may require knowledge of research techniques, a statistical ability, or foreign language skills that you do not yet have.

WRITING PROMPT Preparing a Proposal Imagine that you are tasked with planning a fundraising event. Explain similarities between planning this event and planning for a study in a research proposal. The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

2.5.2: To What Universe Can You Generalize from a Study’s Findings? As you focus on a research question, you also must specify the universe to which you can generalize an answer. Only rarely do you want to restrict findings to the specific units or cases you happened to study. Instead, you want to extend them to a broader category of people, organizations, and other units. For example, your research question is: Does a new attendance policy help high school students learn more? You plan to study three high schools in one U.S. city in 2016. The universe, in this case, is all high school students. You want to generalize what you learn beyond the specific students in three high schools of one U.S. city in 2016 to all high school students, or at least all U.S. high school students in the early twenty-first century.

2.5.3: Will You Follow a Linear or Nonlinear Path When Doing Research? A path is a metaphor for a sequence of activities; it is a way of thinking and looking at issues. In general, with quantitative data you follow a relatively fixed sequence of steps in one direction or a linear path. It’s like a staircase, moving straight along a pathway upward, without deviation, and takes you to a single location. When gathering qualitative data, the pathway is less a straight line or fixed sequence; it is flexible, multidirectional, and nonlinear. A nonlinear path makes successive passes through steps and moves sideways before going forward. You advance slowly but not directly. It is more of a spiral. At each cycle or repetition, you collect data and gain new insights, then move ahead. If you are accustomed to the direct, linear approach with fixed steps, the nonlinear path may look inefficient and sloppy. A nonlinear approach does not have to be disorganized, and it’s never an excuse for doing poor-quality research. It has its own discipline and rigor. It can be highly effective when adjusting to a fast-changing, fluid situation. It can create a feeling for the whole, allow subtle shades of meaning to emerge, pull together divergent information, and permit switching perspectives. In contrast, if you are used to a nonlinear approach, the linear approach may appear to be rigid and artificial. It may appear too set, fixed, and standardized. You may believe it misses what is most interesting in dynamic human relations. The linear path can offer a highly efficient, disciplined, and simple-to-follow sequence that makes it easy to spot a mistake and to repeat a past study.

40 Chapter 2

2.5.4: Do You Examine Variables and Hypotheses or Cases and Contexts? The decision to collect quantitative or qualitative data influences the research process in several ways. Studies that depend on quantitative data “speak a different language” than studies that use qualitative data and the corresponding research approach. In general, studies that discuss variables and hypotheses rely on quantitative data and the corresponding research logic. By contrast studies that focus on examining cases and how the context affects them use qualitative data and a qualitative approach to research. Quantitative Data Variables The variable is a central idea in quantitative research. Simply defined, a variable is a concept that varies. Research with quantitative data uses a language of variables and emphasizes relationships among variables. Once you begin to look for them, you will see variables everywhere. For example, gender is a variable; it can take on at least two values: male or female. Marital status is a variable; it has the values of never married single, married, divorced, or widowed. Type of crime is a variable; it can take on values of robbery, burglary, theft, murder, and so forth. A person’s attitude toward abortion is a variable; it ranges from strongly supporting the right to a legal abortion to strongly opposing abortion. It is easy to confuse a variable with the categories or values of variables. A category in one variable can itself become a separate variable with a slight change in definition. For example, “male” is not a variable but it describes a category of the variable gender. A related idea, “degree of masculinity,” is a variable because it describes the intensity or strength of attachment to attitudes, beliefs, and behaviors associated with being masculine as part of the broader idea of gender. If you wish to gather quantitative data, you must convert most of your ideas into the language of variables. We can classify variables into three basic types:

• independent, • dependent, and • intervening. How you classify a variable depends upon its location in a cause-effect statement. The cause variable is the independent variable. The result or effect variable is the dependent variable. The independent variable is “independent of” prior causes that act on it. Two questions help identify the independent variable: • Does it come earlier in time? Independent variables always come before any other type.

• Does it have an impact on another variable? Independent variables have an impact on other variables.

The dependent variable “depends on” the cause. You can reword most research questions in terms of the dependent variables because a dependent variable is what you will explain. If your research question is about why the crime rate increased in Dallas, Texas, then your dependent variable is the Dallas crime rate. A simple cause-effect relationship requires only an independent and a dependent variable. A third type of variable, the intervening variable, appears in complex relationships and shows the link or mechanism between the independent and dependent variable. To advance knowledge, we both document simple cause-and-effect relationships and try to specify the mechanisms within a causal relation. In a sense, the intervening variable acts as a dependent variable with respect to the independent variable and as an independent variable toward the dependent variable. Three-Variable Example The famous French sociologist Emile Durkheim developed a theory about a causal relationship between marital status and suicide rates. He found that married people are less likely to commit suicide than single people and believed it was because married people were more socially integrated (i.e., had feelings of belonging to a group or family). We can restate his theory as the following: Being married (independent variable) increases social integration (intervening variable), which in turn reduces likelihood of suicide (dependent variable). Specifying the chain of variables clarifies linkages in a causal explanation. Complex theories have multiple independent, intervening, and dependent variables. For example, you notice that people from disruptive family settings have lower incomes as adults. Why? Family disruption causes lower self-esteem among children, which causes greater psychological depression, which causes poor grades in school, which causes reduced prospects for getting a good job, which causes a lower adult income (see Figure 2.9). Family disruption is the independent variable. Adult income level is the dependent variable. All the rest are intervening variables. Two explanations of the same dependent variable may use different independent variables, or agree about the independent and dependent variables but differ on the intervening variable. Both may say that family disruption causes lower adult income. One holds that disruption encourages children to join deviant peer groups that are not socialized to norms of work and thrift. Another emphasizes the impact of the disruption on childhood depression and poor academic performance.

Planning a Study 41

Figure 2.9 Chain of Variables A complex causal explanation often contains a chain of several connected variables Family Disruption

Low SelfEsteem By Child

Psychological Depression

Poor Grades

In a single study, you usually test only one part of a larger causal explanation. Even though you might test one small part, you want to link it to the larger explanation. In a study, you connect independent and dependent variables with the hypothesis. A hypothesis is a tentative statement of a relationship between two variables. It is a guess about how the world works. We can restate the hypothesis as a prediction about what we expect. A hypothesis for a causal explanation has the following five characteristics: • It has at least two variables. • It specifies how the variables are connected, which is the cause, and which is the effect. • It includes a time order assumption (what comes first). • You can restate it as a prediction or expected finding. • You can show that it is supported or false with empirical data. Example The more a couple attends religious services together, the lower the chances that they will divorce. • Two variables: (1) Frequency of attending religious services together and (2) probability of divorce. • Connection: Lower attendance causes higher chance of divorce, and vice versa. • Time order: Attendance is earlier and divorce comes later. • Prediction: Couples who attend religious services together tend to have fewer divorces than couples who never or rarely attend religious services together. • Testable with empirical data: We can look at 1,000 couples and ask how often they attend religious services together, then see how many of them are still married to one another 10 years later.

WRITING PROMPT Causal Diagrams Create a hypothesis that has three variables (independent, intervening, and dependent). The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

Low Job Prospect

Adult Income Level

Hypotheses Knowledge cannot advance far with one

test of a single hypothesis. In fact, you may get a distorted picture of the research process if you focus too much on a single study that tests one hypothesis. Knowledge develops over time as many researchers test many hypotheses. It grows out of the shifting and winnowing of many findings about hypotheses. If data fail to support some hypotheses, researchers gradually drop them from consideration. If data support a hypothesis, they keep it in contention. Researchers constantly create new hypotheses that challenge existing ones that have received support. Over time, if a hypothesis continues to receive empirical support in test after test and repeatedly is better than alternative hypotheses, we accept it as likely to be true. Acceptance requires multiple tests with consistent and recurring empirical support. Making It Practical: From the Research Question to Testable Hypotheses Going from a well-

formulated research question to a hypothesis is a short step. A good research question contains hints about hypotheses. A hypothesis is a tentative answer to the research question. Consider the broad research question, “Is age at marriage associated with chances of divorce?” It has two variables: “age at marriage” and “chances of divorce.” Age at marriage is the independent variable because marriage logically precedes divorce, the dependent variable. Beyond stating that two variables are connected, you must specify the relationship’s direction. You have two choices: 1. The lower the age at time of marriage, the greater the chances that of divorce 2. The higher the age at time of marriage, the greater the chances of divorce. Let us say the hypothesis makes a prediction with choice #1: People who marry younger have a greater chance of divorce than those who marry older. This focuses the research question into, “Are couples who marry young more likely to divorce than those who wait to marry until they are older?” Often a broad research question contains multiple potential hypotheses. The broad question was, “Is age at marriage associated with chances of divorce?” Another hypothesis from it is, “The smaller the difference between the ages of the marriage partners at the time of marriage, the lower the chances of divorce.” Here we specified “age at marriage” differently. We can also specify conditions under which a relationship works. For example, “The

42 Chapter 2 lower the age at time of marriage, the greater the chances that the marriage will end in divorce, except for marriages between members of a tight-knit traditional religious community in which early marriage is the norm.” Besides answering a research question, a hypothesis can be an untested proposition from a theory. We can express a hypothesis at two levels: 1. an abstract, conceptual level of theory; 2. a concrete level that can be empirically measured in a study. The theory explains why the empirical prediction of the hypothesis is true. Clear explanation and logic connect the theory to specific measures. Let us continue with the same example but now state it as a theoretical statement. Adults stabilize a self-identity and develop mature coping skills as they move from their late teens to their late 20s. A stable self-identity and mature coping abilities enable people to sustain a long-term committed intimate relationship, such as marriage. If two adults enter into a marital relationship before they have acquired a stable a self-identity and developed mature coping abilities, the relationship is unlikely to persist very long.

Now we can rephrase it as empirically testable statement with specific measures: The rate of divorce within the first 10 years of a marriage is much higher when both partners are 25 years old or younger at the time of a marriage than when both marriage partners are 26 years old or older.

WRITING PROMPT Two Levels of the Hypothesis Create a two variable hypothesis on a topic that interests you and that you believe to be true. State this hypothesis two ways—at a theoretical and at an empirical level. The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

Hypothesis Our confidence in the truthfulness of a

hypothesis grows as it defeats its competitors in repeated tests. A curious aspect of hypothesis testing is that the evidence in support of a hypothesis is not equal to evidence that negates it. We give the negative evidence greater importance. Technically, researchers do not say that they have proved a hypothesis to be true; however, they do say that they reject a hypothesis. If your evidence supports a hypothesis, the hypothesis is a possibility; it is still in the running and the case is not

closed. When evidence fails to support a hypothesis, it is tarnished and starts to fall out of the running. This occurs because a hypothesis makes a prediction. Negative evidence shows that the prediction is wrong. Positive evidence is less critical. Alternative hypotheses may make the same prediction. Confirming evidence may reinforce belief in a hypothesis, but it does not automatically beat out the alternative hypotheses that make the same prediction. Serious negative evidence can eliminate a hypothesis, yet piling up more and more evidence in favor of a hypothesis is not as noteworthy. Researchers test hypotheses in a straightforward way, or they use the null hypothesis. Most of us talk about a hypothesis as a way to predict a relationship between two variables. The null hypothesis does the opposite; it predicts no relationship. Many quantitative researchers, especially experimenters, use a null hypothesis approach. They seek evidence that allows them accept or reject the null hypothesis. For example, Sarah believes that students who live on campus in dormitories get better grades than students who live off campus and commute. Her null hypothesis is that residence and grades are unrelated. She matches the null hypothesis with a corresponding alternative hypothesis. It is that a relationship exists; more specifically that a student’s on-campus residence has a positive effect on grades. Many people think that the null hypothesis looks like a “backward way” to test hypotheses. It rests on the assumption that hypothesis testing should make finding evidence for a relationship between variables very demanding. With the null hypothesis approach, we directly test the null hypothesis. If evidence supports the null hypothesis (technically—it is accepted as true), we are forced to conclude that the alternative hypothesis is false. If the evidence rejects the null hypothesis, then the alternative hypothesis remains as a possibility—it is still in contention. As we repeatedly test and reject the null hypothesis, the alterative hypothesis gains support over time. Researchers use the null hypothesis because they want to be extremely cautious. Until there is a mountain of evidence, they hesitate to say that a relationship exists. It is similar to the Anglo-American legal idea of innocent until proved guilty. We assume the null hypothesis, until enough evidence suggests otherwise.

WRITING PROMPT The Null Hypothesis In what ways does the concept of “null hypothesis” correspond to the “innocent until proven guilty” idea? Describe at least one other situation when we posit the absence of something when testing for it. The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

Planning a Study 43

Qualitative Data Cases and Contexts You have seen how studies with quantitative data emphasize variables. By contrast, studies with qualitative data examine cases and contexts. A researcher who uses qualitative data may not think in terms of variables or testing hypotheses. He or she sees many areas of social life, human relations, and social activities as being intrinsically qualitative. Rather than try to convert fluid qualitative social life into variables or precise numbers, he or she retains the loose images or ideas that people use in natural social contexts. In a qualitative research study, we frequently examine a limited number of cases in depth. The cases are usually the same as a unit of analysis. Instead of using precise numerical measures of a very large number of cases, as in quantitative data analysis, we examine in detail many aspects of a few cases. The rich detail of and astute insight into the few cases replace precise measures across numerous cases. Because we closely examine the same case over time, we can better see an issue evolve, a conflict emerge, or a social relationship develop. This puts us in a good position to detect and observe processes. In historical research, the passage of time may involve years or decades. In field research, it may be days, weeks, or months. In both, we see what unfolds and can notice when something unusual or important occurs. The social context is very important for studies with qualitative data. This is because an event, social action, or statement’s meaning depends, in an important way, on the context in which it appears. When we remove an event, social action, or conversation from its social context, or ignore the context, we can seriously distort its meaning. Without the context, its real importance or significance is often lost. This requires us to pay close attention to what surrounds an event, action, or statement. It also implies that the same events, actions, or statements can have different meanings in different situations, cultures, or historical eras. Let us say we want to study voting. Instead of simply counting votes across time or cultures, we might ask: What does voting mean in the context?

The same action (e.g., voting for a presidential candidate) may differ depending on the context, such as intense argument and competition among several parties, no difference among candidates, or a situation of total one-party dominance. Until we place the parts of social life into a larger whole, we may not grasp the part’s meaning. For example, it is hard to understand a baseball glove without knowing something about the game of baseball. If a person looks at it as a glove, like a mitten for cold weather, driving gloves, or gloves to use for working in the garden, the baseball glove makes little sense. The glove’s meaning comes from its use and placement within the flow of a baseball game. The whole of the game—innings, bats,

curve balls, hits—gives meaning to each of the parts. Each part without the whole has little meaning.

2.5.5: How Will You Analyze Patterns in the Data That You Gather? You look for patterns in the quantitative and qualitative data, but in different ways. With quantitative data, you use charts, tables, and statistics to rearrange, examine, and evaluate the numbers in ways that can reveal patterns. You then connect the patterns back to the research question. Likewise, the hypothesis simultaneously answers the research question and makes a prediction about what will appear in the charts, tables, and statistics. With qualitative data, you also look for patterns. The patterns emerge as you rearrange, examine, and evaluate textual or visual data. While doing this, you also try to retain an authentic voice and remain true to the original understandings of the people that you studied. Instead of relying on charts, statistics, and displays of numbers, you identify patterns (i.e., sequences, cycles, contrasts) in the data, (i.e., observed events, conversations, or situations) as they occurred in a specific context. You might discuss the patterns in terms of themes or as narratives. A narrative is a story that has a beginning and ending and major actors or forces that pull the reader from start to finish. Qualitative data are often more complex and filled with specific meaning than numbers. Your job is to translate, or make understandable, the data for people who lack a direct experience with the specific research setting. For example, you may describe the following 30-second social interaction in which no one spoke as follows: A middle-aged man in a business suit rushes into a coffee shop, opens his wallet, and puts a five-dollar bill on the counter. Without a word, the clerk at the shop quickly pours a cup of coffee into a take-away container and adds cream for the man. The man picks up the container, turns, and quickly walks out the door.

A “translation” of this interaction integrates observations, conservations, and the context. For example, you

44 Chapter 2 know that the man is a regular patron of the coffee shop near a train station. He has been coming each morning for five years. Today he is in a rush to catch a commuter train to his office downtown. He will drink the coffee while on the train. The clerk knows the man and knows what he wants. The man orders the same thing every day. In return, the clerk rushes to take care of the man each time. When the man is very rushed, he just puts down a five-dollar bill. The coffee only costs $2.50; the money covers the cost of the coffee, $1.50 for the newspaper that the man took from outside the front of the coffee shop before entering, and a $1.00 tip. On the days that he has time, the man sits and chats with the clerk about baseball and current events.

Summary Review Table 2.3 Quantitative versus Qualitative Research Overall Type of Study

Quantitative Research

Qualitative Research

Approach

Usually deductive

Usually inductive

Research Question

Developed and refined before gathering data

Developed and refined while gathering data

Path

Linear

Non-linear

Main goal

Test a hypothesis that you started with

Discover/capture the meaning of a social setting

Concepts and Ideas

Are expressed in the form of distinct variables

Are expressed in the form of themes and motifs

Measurement

Plan precise measurements before data collection

Create measures ad hoc as gathering data

Data

In the form of numbers

In the form of words and images

Theory

Usually causal

Can be causal or other

Data Analysis

Data analysis includes statistics, tables, or charts with relationships among numbers

Often includes narrative story with detailed descriptions of people and events in a social setting

WRITING PROMPT Two Types of Data List the pros and cons of quantitative and qualitative data and research, in your perspective. Based on this analysis, explain which approach you would prefer. The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

2.5.6: What Type of Explanation WillGive Meaning to the Patterns inthe Data? We use the word “explanation” in two ways: 1. an everyday way in which explanation means making something clear or comprehensible to another person with examples 2. answering the “why?” question and making something comprehensible by placing it within a relevant theory, set of ideas, or context. In explanatory research, the goal is to create a research explanation. The most common type of explanation is the causal explanation. In it, you explain by identifying one or more causes for an effect or outcome. The cause corresponds to your independent variable and the effect to your dependent variable. A causal explanation often connects to a theory and has the following three elements: • Time order: The cause must come earlier in time than the effect or the result it produces. • Association: The cause and effect are associated, or they go together and vary with one another. Some people call it correlation, although technically correlation is a specific measure of association. • Alternative causes ruled out: There is no better or stronger cause than the one you identified. To be a cause (an independent variable), something must happen first. Usually you can observe or logically determine time order. Two factors that occur together are associated; that is, when one factor is present or at a high level, the other one is also present or at a high level. Several statistics measure an association. The best known is the correlation coefficient. The last item in the list is the most difficult one to document or observe. If you claim that one factor causes another, there should not be any stronger, truer, or better cause present that you are not including. This is an important element because there can be multiple causal factors, some obvious and some hidden. If you say one factor causes another, yet in fact, an unacknowledged and stronger cause is present, your explanation will be misleading. To avoid this problem, you want to rule out other possible causes. In a causal explanation, you link a generalization to a specific instance of it, as follows: • A causes B generally. • This situation is A; therefore, we expect to find B. Example People who spent many years in prison have difficulty finding stable, well-paid work after their release. Joe Brown was imprisoned for many years; he is having difficulty finding stable, well-paid work after his release.

Planning a Study 45

You fit a specific observable instance within the more general rule or pattern. You can covert a causal explanation into independent and dependent variables. Independent variable: Whether a person was previously imprisoned for many years Dependent variable: Amount of difficulty in finding stable, well-paid work Sometimes researchers who use qualitative data use causal explanation. At other times, they instead develop ideas or theories during the data collection process. They build up from specific data to general ideas. Instead of a causal connection between two variables, their explanation is in the form of motifs, themes, or distinctions. Many explanations with qualitative data take the form of grounded theory. To build a grounded theory explanation, you make comparisons. For example, you observe an event (e.g., a police officer confronting a speeding motorist). You look for similarities and differences. Does the police officer always radio in the car’s license number before proceeding? After radioing the car’s location, does the officer ask the motorist to get out of the car sometimes but at other times casually walk up to the car and talk to the seated driver?

When data collection and theorizing are interspersed, theoretical questions arise that suggest future observations. You collect new data so that they can answer theoretical questions that came from thinking about previous data.

WRITING PROMPT Grounded Theory Consider something you recently observed that peaked your curiosity. Develop a mini-theory or hypothesis you could develop from this observation. The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

2.5.7: What Are the Units of Analysis in Your Study? Every study has units of analysis. They are an essential element for clearly thinking through and planning a research study. Yet, most research reports do not explicitly identify the units of analysis as such. Your research question shapes the unit of analysis, which in turn influences study design. This means an awareness of them enables you to design a better study and to avoid errors. Common units are the individual, the group (e.g., family, friendship group), the organization (e.g., corporation, university), the social category (e.g., social class, gender,

race), the social institution (e.g., religion, education, the family), and the society (e.g., a nation, a tribe). Let us say that you want to conduct a descriptive study to find out whether colleges in the North spend more on their football programs than do colleges in the South. The variables are college location and college spending for football. The unit of analysis in this situation is the college. It flows from the research question. The individual is the most commonly used unit of analysis in social research, but it is by no means the only one. Different research questions imply different units of analysis, and different research techniques work best for specific units. For example, in the Piaac study of adult skill levels described above, the nation is the unit of analysis because it compared average adult high-level skills across various nations. On the other hand, you might conduct a study to compare the level of skills of graduates from 20 different universities and colleges in one nation. Perhaps you think learning environments that provide close, personal faculty–student relations best develop and nurture the highest level skills in students. You could contrast three learning environments, the selective small, private college, a large, public university, and an open-enrollment mass online college. You could measure the skills of graduates from each type of learning environment, and examine school characteristics (e.g., entrance requirements, student body size, tuition level, and online or face-to-face teaching). Your unit of analysis would be the learning environment (i.e., organization), or specifically the college or university. This is because you are comparing characteristics of the school. Units of analysis influence how to gather data and the level of analysis. What Is the Level of Analysis of Your Study? The last study design issue is the level of analy-

sis. The social world operates on a continuum from a small scale or micro level (e.g., a few friends, a small group) to a large scale or macro level (e.g., entire civilizations or a major structure of a society). The level of analysis is the level of reality you examine. It is a mix of the number of people, the expanse of geographic space, the scope of the activity, and the length of time. A micro-level study might involve 30 minutes of interaction among five people in a small room. A macro-level study could involve a billion people on three continents across a century. The level of analysis delimits the kinds of assumptions, concepts, and theories you will use. It also influences the appropriate units of analysis. Let us look at examples at each end of the continuum. Micro Level Suppose you want to study the topic of dating among college students. A micro-level analysis uses ideas such as interpersonal contact, mutual friendships, and common interests among individual students. Suppose you believe that students tend to date someone with

46 Chapter 2 whom they have had personal contact in a class, share friends in common, and share common interests. You might gather data from 100 students on their friends, contacts, and relationships. The individual student is your unit of analysis. Macro Level Suppose you want to learn how social-economic inequality affects violent behavior in a society. You may be interested in the degree of inequality (e.g., the distribution of wealth, property, income, and other resources) throughout a society. Likewise, you may look at patterns of societal violence (e.g., aggression against other societies; level of violent crime; violent feuds between families; organized crime with gangs, bandits, and warlords; religiousracial-based conflicts). You develop a macro-level explanation because of the topic and the level of social reality. You gather data on the level of inequality in each of 50 countries for 20 years, as well as data on how many acts of violence occurred in each country. The country is your unit of analysis. Avoid Spuriousness As you design a study with a

causal explanation, you need to be aware of an issue that may totally upset your explanation. As you learned previously, a causal explanation requires time order, association, and ruling out of alternative causal factors. You can observe or test the first two, but the third element can be tricky. You must make certain that there are no alternative causes of the outcome. The alternative cause may not be obvious. If an unseen alternative cause strongly affects your dependent variable, then any claims you make about the cause (independent variable) would be wrong. In short, establishing time order and a strong association between two variables is not enough. The observed relationship could be an illusion, just like the mirage that resembles a pool of water on a road during a hot day.

Table 2.4 Quick Checklist of Study Design Issues in a Research Proposal Study Design Issues

Options

When do you focus the research question?

Very early or it emerges later with data collection

What is the universe of your study?

Select a broad set of units to which you can generalize

What is your research path?

Linear or nonlinear

What do you examine?

Variables and hypotheses or cases and contexts

How do you interpret patterns in the data?

Statistics and charts or themes and narratives

What type of explanation do you use?

Causal explanation or grounded theory

What are your units of analysis?

Select the type of cases or units you will measure

What is your level of analysis?

Micro to macro

Spuriousness is an illusionary relationship due to an unacknowledged other variable that is a cause of both the independent and the dependent variable. Finding a strong correlation between the two variables does not mean you definitively found a cause and effect. You must also check for spuriousness before claiming causality. Spuriousness may seem complicated, but it relies on common-sense logic. For example, you know that there is an association between the use of air conditioners and ice cream cone consumption. If you measured the number of air conditioners in use and the number of ice cream cones sold each day, you would find a strong correlation. Shops sell more cones on the same days when more people turn on air conditioners. However, you know that eating ice cream cones does not cause people to turn on air conditioners, or turning on an air conditions does not cause people to crave ice cream cones. Instead, a third factor causes both variables: hot days. How can you tell whether a relationship is spurious? How do you discover if there is a mysterious, unseen third factor? How can you build in a safeguard to prevent spuriousness?

You can gather data to help to control for spuriousness. In survey data and existing statistical sources, you must decide on control variables that measure possible alternative causes. You then use statistical techniques to test whether an association is spurious. In all situations, including qualitative data analysis, you need a theory, or at least a good guess, about what alternative causes might influence what you see as the cause and effect. You then take them into consideration by gathering data. One way to grasp the idea of spuriousness is with an example. Example Let us consider the following relationship: using illegal drugs causes more suicides, school dropouts, and violent acts. We see positive correlations between people who take drugs and being suicidal, dropping out of school, and engaging in violence. Many people look at the relationship and argue that drugs are the cause and ending illegal drug use will stop social problems such a suicide or violence. An alternative view says the relationship is spurious—many people turn to illegal drugs as a way to cope with emotional distress or high levels of disorder in their communities (e.g., high unemployment, unstable families, high crime, few community services, and lack of civility). Likewise, people who experience great emotional distress or who live in highly disordered communities are more likely to commit suicide, drop out of school, and engage in violence. Perhaps reducing emotional distress and community disorder can end both illegal drug use and the other social problems. If the alternative view is correct, reducing drug taking alone may have a little effect because it does not address the root causes (i.e., emotional problems and community disorder).

Planning a Study 47

This makes the apparent relationship between illegal drugs and social problems spurious because emotional distress and community disorder are the true and initially unacknowledged alternative causes of both illegal drug use and the other social problems (see Figure 2.10).

Figure 2.10 Spuriousness Example—The Relationship between Illegal Drugs and Suicide Illegal Drugs

Emotional Problems and Community Disorder

Observed association

Suicide

True cause

Learning from History Night-Lights and Spuriousness For many years, researchers observed a strong positive association between the use of a nightlight and children who were nearsighted. Medical professionals thought that the night-light somehow caused children to develop vision problems and advised parents against using a night-light for their children. Other researchers could find no good reason for night-light use causing nearsightedness. A 1999 study provided the answer. It found that nearsighted parents are more likely to use night-lights. Parents also genetically pass on their vision deficiency to their children. The study found no link between nightlight use and nearsightedness once the effect of parental vision was considered. Thus, the initial causal link was misleading or spurious once researchers considered the previously unrecognized impact of parental vision impairment and nighttime behavior (see New York Times, May 22, 2001).

Summary: What You Learned about Planning a Study In this chapter, you saw how to conduct a literature search, narrow a topic into a focused research question, and identify units and levels of analysis. The decision to use qualitative or quantitative data suggests a different sequence of decision making. Choosing a qualitative or quantitative approach (or a mix of both) depends on your topic, your purpose, and intended use of study results, as well as your own assumptions. If you decide that quantitative data are best, you take a linear path and emphasize objectivity and use explicit, standardized procedures and a causal explanation. You use the language of variables and hypotheses testing. The process is a set of discrete steps that precede data collection: Narrow the topic to a more focused question, transform concepts into variables, and develop hypotheses to test. In actual practice, you move back and forth, but the general process flows in a single, linear direction. Your explanations usually take a cause-effect form. If you decide that qualitative data are best, you follow a nonlinear path that emphasizes becoming intimate with the

details of a natural setting or a particular context. You use fewer standardized procedures or explicit steps, and you must devise on-the-spot techniques. You use a language of cases and contexts that directs you to examine particular cases or processes in detail. You do not separate planning and design decisions into a distinct pre-data collection stage, but you continue to develop the study design throughout early data collection. You slowly evolve toward a specific focus based on what you learn from the data. As you reflect on the data, you can develop a grounded theory explanation. The qualitative and quantitative distinction is overdrawn and is not a rigid dichotomy. You can mix the two types. Before you mix the data types, you need to understand each and appreciate each on its own terms. You should recognize that quantitative and qualitative data each have strengths and limitations. The ultimate goal is to better understand and explain events in the social world, and the best way to do so is to appreciate the value of each style of data collections has to offer. Studying people and doing a research study about human relations also have an ethical-moral dimension.

48 Chapter 2

Quick Review

How Do We Focus the Research Question?

How Do We Select a Topic to Study?

1. Before conducting a study, it is essential to narrow the focus from a topic into a specific research question that the findings of the study will answer.

1. The topics appropriate for a social science study can arise from a very wide range of sources.

2. How and when we narrow to a research question depends on whether we use an inductive or deductive approach. With an inductive approach, we begin looking at the specific details in the evidence, then move toward general ideas; while with a deductive approach, we start with general ideas, then move toward testing or verifying the ideas with specific evidence. While we can find a mix of both approaches in many studies, most emphasize one or the other.

2. Appropriate research topics have the following four features: (a) it is generalizable and occurs beyond one single, isolated instance; (b) it is about patterns or reoccurring situations; (c) it is about what occurs in aggregates, i.e., across many cases/units; and (d) it is about something we can empirically observe, either directly or through indirect means. 3. Social research looks at the “big picture” or “forest,” and if it considers a particular instance or “tree” it is within a larger context.

Six-Step Literature Review Process 1. The “literature” refers to past research reports on a topic, and reading the literature can serve several purposes. Two important purposes are to find out what researchers already learned about a topic and to place a new study in the context of previous studies so that we can build on prior knowledge. 2. The “literature” is part of the communication system of research that operates on the principle that researchers share all study findings and how they conducted a study. 3. Stories about research studies appear in many places (e.g., newspapers, in popular magazines, on television or radio broadcasts, and in Internet news reports), but these are insufficient for preparing a literature review, because they do not have the full, complete study reports. 4. The primary place where researchers disseminate information about studies is in scholarly journals. The journals are usually available only through college and university libraries. They contain the full study reports, and we can locate articles in the journals by using specialized article search tools. 5. Most scholarly journals only accept a small percentage of study reports they receive. They decide which ones to accept after the reports have been evaluated in the peerreviewed process. The peer-reviewed process is a kind of quality assurance system. 6. Although full study reports appear in some formats other than scholarly journals, it is much more difficult to locate the reports in such formats. 7. To conduct a literature review, it is necessary to focus on a specific topic, develop a plan for locating relevant study reports, and take careful notes on the contents of the research reports. 8. A crucial step in the literature review is to organize the notes, then synthesize and interconnect the findings from many study reports into a well-written review.

3. A deductive approach tends to go with quantitative data and explanatory studies, while the inductive approach tends to go with qualitative data and exploratory studies. These are only general tendencies.

How to Design a Study for a Research Proposal 1. We systematically collect and analyze data in all research studies, but the techniques appropriate for qualitative data may be inappropriate for quantitative data, and vice versa. Neither qualitative nor quantitative data are best; rather, each has strengths. 2. The universe is the set of units or situations to which the study applies, and we need to decide to what universe we can generalize a study’s findings. 3. When planning and conducting research, we tend to follow a linear or nonlinear path. A linear path is like a straight line in which each step in the process follows directly after the previous one. A nonlinear path moves back and forth, or is more of an upward moving spiral. Research on quantitative data tends to be linear while that with qualitative data is more likely to be nonlinear. 4. If we use a linear, deductive approach and gather quantitative data, we start by converting ideas or characteristics of the units we are examining into variables. We then link the variables together as hypotheses. There are three major types of variables: Independent (Cause), Dependent (Result or Outcome), and Intervening (Comes in between). A hypothesis is a prediction or statement about which variables affect other variables and the types of impact they have on one another. In linear-deductive approach, we start with the hypothesis and then gather data to test whether they support the hypothesis. 5. If we use a nonlinear, inductive approach and gather qualitative data, we emphasize looking at the details of specific cases and the contexts in which the cases appear. Rather than starting with a hypothesis to test, we start with looking at specific details, and then build toward generalizations. 6. In all studies, we look for patterns in data, but this varies depending on whether the data are quantitative or qualitative.

Planning a Study 49 We use charts, tables, and statistics to rearrange the numbers in quantitative data to reveal patterns. With qualitative data, we look for patterns as we rearrange, examine, and evaluate textual or visual data, and often discuss the patterns in terms of themes or as narratives. A narrative is a story with a beginning and endand major actors or forces that pull the reader from start tofinish. 7. The word “explanation” can mean making something clear to another person or answering the “why?” question by placing it within a relevant theory, set of ideas, or context. In explanatory research, we create a research explanation, and the most common type, especially for deductivequantitative data, is a causal explanation. A causal explanation has three elements: time order, association, and ruling out alternative causes. These features are in the hypothesis. In the inductive approach commonly used with qualitative data, we often use grounded theory instead. It emerges after examining the data and builds an explanation by making comparisons. 8. Every study has units of analysis. The research question shapes the unit of analysis. They are the kind of case or unit in the data about which we gather information and make comparison, e.g., individual people, schools, or entire countries. Most research reports do not explicitly identify the units of analysis as such. 9. Every study has a level of analysis. It refers to the level of the social world we explore, and operates on a continuum

from a small scale or micro level (e.g., a few friends, a small group) to a large scale or macro level. The research question shapes the level of analysis of a study, which in turn is unit of analysis and appropriate data. 10. When making a causal explanation, we always want to avoid spuriousness. It is an illusionary relationship due to an unacknowledged other variable that is a cause of both the independent and the dependent variable.

Shared Writing: Spuriousness Develop an example where there is a strong correlation of two factors, but a third factor really explains the dependent variable. Don’t identify the third factor. Comment on at least examples provided by two of your classmates and suggest a possible third factor as for what may have caused the relationship. Identify what steps would help to analyze the example to reveal the third factor as the true cause. A minimum number of characters is required to post and earn points. After posting, your response can be viewed by your class and instructor, and you can participate in the class discussion.

Post

0 characters | 140 minimum

Chapter 3

Becoming an Ethical Researcher

Learning Objectives 3.1 Recognize the importance of personal

integrity in research 3.2 Understand the origin and application of

the rules of conduct in using humans in research

3.4 Recognize the impact that politics and large

corporations have on research 3.5 Examine the complexity of “value-free” and

“objective” research

3.3 Identify some of the tactics used by research

sponsors to aid their causes

In 1932, the U.S. Public Health Service began a study on how the disease syphilis progressed. The goal was to improve treatment programs for infected African Americans. Six hundred low-income black men in Macon County, Alabama participated in the “Tuskegee Study of Untreated Syphilis in the Negro Male.” Of participants, 399 had syphilis and 201 did not have the disease. Researchers never told

the participants that the study was on syphilis or that they had syphilis. Instead, they told the participants that they were being treated for “bad blood,” a local term for many ailments, including syphilis, anemia, and fatigue. In exchange for participating in the study, the men received free medical exams, free meals, and burial insurance. Researchers instructed local physicians not to treat

Becoming an Ethical Researcher 51

participants for syphilis. Although the study was supposed to be six months long, it continued for 40 years. Researchers never treated the participants for syphilis; even after a highly effective treatment, penicillin became available in 1947. They followed the men until death. Untreated syphilis goes through several stages. Second-stage syphilis develops three months to three years after onset. It causes fever, swollen lymph glands, sore throat, patchy hair loss, headaches, weight loss, muscle aches, and fatigue. If untreated, it becomes late-stage syphilis. This stage damages the victim’s heart, eyes, brain, nervous system, bones, and joints. It can produce mental illness, blindness, deafness, memory loss, serious neurological problems, heart disease, and death. The “Tuskegee Study of Untreated Syphilis in the Negro Male” ended in 1972 only because of an exposé by a news reporter. By the time it ended, 28 participants had died of untreated syphilis. One hundred others died due to syphilis-related complications. Participants had infected 40 wives, and 19 children had contracted the disease atbirth. The reporter who wrote the story that ended the study had talked to a research interviewer working for the study. The interviewer had tried to raise ethical concerns with the U.S. Public Health Service for five years but could not get top research officials to end the study and provide proper medical care for the participants. With the media exposé, a national scandal erupted over the study in 1973. After the U.S. Congress held public hearings, the federal rules on ethics in research with humans were overhauled. Eventually, the surviving research participants sued the U.S. government and were awarded $10 million in an out-of-court settlement. Even after the 1974 settlement, it took another 23 years before the President of the United States officially apologized to the surviving research participants. The Tuskegee syphilis study is an outrageous instance of a disregard of basic ethical principles in research with humans in the United States. Two books, a play, and a dramatic movie, Miss Evers’ Boys (1997), document this incident. This unethical study and the publicity surrounding it highlight the limits on research with humans. When studying people, you must follow ethical principles. In this chapter, you will learn about ethical issues involved with doing social research. We conduct a study to create knowledge, answer questions, solve problems, or help humanity, but we must always do this in a morally responsible manner. Research ethics includes the concerns, dilemmas, and conflicts over the proper way to conduct a study. Ethics defines what is or is not morally proper. This is not an easy task. There are few clear ethical absolutes. Various ethical principles or guidelines require you to exercise judgment, some principles may conflict with others in practice, and you often must balance competing priorities.

You must balance potential benefits from research— such as advancing understanding and improving decision-making—against its potential costs—such as loss of dignity, self-esteem, privacy, or democratic freedoms.

3.1: What Is the Ethical Imperative? 3.1 Recognize the importance of personal integrity inresearch Ethics is something you should think about early in the research process, as you plan a study or prepare a research proposal. While it is difficult to appreciate ethical issues fully until you begin to conduct a study, waiting until you are in the middle of doing research is too late. By preparing ahead of time and considering ethical concerns, you can build sound ethical practices into a study’s design and be alert to potential ethical concerns that may arise. Being aware of research ethics also help you to better understand the process of social science research. Researchers have a strong moral and professional obligation to act ethically at all times and in all situations. This applies even if research participants are unaware of or are unconcerned about ethics, or if an employer or the sponsor of a study asks you to engage in unethical research. “The research participants didn’t care,” or, “My boss told me to do it,” are never acceptable reasons for engaging in unethical behavior. Although most professions (e.g., journalists, law enforcement, medicine, accounting, etc.) have ethical standards, standards for doing research with humans may be more rigorous. Being ethical is rarely easy. Moral, legal, and political philosophers over the centuries have debated the issues that a researcher may face. Ethical behavior begins and ends with you, the individual researcher. The best defense against unethical behavior is a strong personal code of moral behavior. Before, during, and after conducting a study, you will have opportunities to, and should, reflect on the ethics of research actions and consult your conscience. Ultimately, doing ethical research rests on your personal integrity. Given that most people who do research are genuinely concerned about other people, you might ask, why would a researcher ever act in an ethically irresponsible manner? The most common causes of unethical behavior are lack of awareness and pressures to take shortcuts. People feel pressures to build a career, publish new findings, advance knowledge, gain prestige, impress family and friends, satisfy job requirements, and so forth. Doing ethical research takes longer, is more costly, and is more complex to complete.

52 Chapter 3 If you act ethically all the time, few people will rush to praise you. This is because ethical behavior is expected. However, if you act unethically and are caught doing so, expect public humiliation, a ruined career, and possible legal action against you. The best way to prepare for e thical decision-making is to internalize a sensitivity to ethical concerns, adopt a serious professional role, and maintain constant contact with others in the research community.

3.1.1: Scientific Misconduct Professional researchers, research centers, and government agencies that fund research all have rules against scientific misconduct. Two major types of scientific misconduct are research fraud and plagiarism. Both are serious ethical violations for which there is never an excuse. Research fraud happens when a researcher invents data that he or she did not really collect and fails to disclose honestly and fully how he or she conducted a study. Fraud in research is rare. It includes significant, unjustified departures from the generally accepted practices for doing and reporting on research. When someone conducts a study that sharply differs from generally accepted research practice, people suspect either incompetence or possible fraud. Plagiarism is misrepresenting what someone else wrote or thought as your own. If you copy two sentences from someone’s research report or use another researcher’s questionnaire but fail to report the source, you have committed plagiarism. You must put very serious effort into keeping track of sources and properly citing them. Good documentation should allow others to track borrowed ideas back to the original source.

Example Study Scientific Misconduct and theMiracle Study In October 2001, a peer-reviewed publication, the Journal of Reproductive Medicine, published a study by three Columbia University medical school researchers. The researchers, Cha, Wirth, and Lobo (2001) claimed to have demonstrated that infertile women who had people pray for them became pregnant twice as often as women for whom no one said prayers. The researchers reported that 199 women undergoing in vitro fertilization in Korea, and who had Christian groups in Australia, Canada, and the United States pray for them, conceived at twice the rate as women who did not receive prayer. Researchers never told patients they were part of a study or that anyone was praying for them. Articles touting the findings appeared in newspapers worldwide, and television news programs announced the miracle study.

Read More Serious readers of the study became suspicious due to a lack of details and the use of an unusual and extremely complex study design. Suspicions increased when the researchers refused to share data or to answer questions. Rogerio Lobo, the lead author, first failed to respond to inquiries, then said he did not know about the study until 1 month after it was finished. He subsequently removed his name from the study and stepped down as chairman of the Obstetrics and Gynecology Department. Dr. Lobo had connections with the journal and might have influenced the peer review process. The second author, Dr. Cha, refused to answer questions and left Columbia University shortly after the study appeared. A few years later, another scholarly journal found him guilty of plagiarism. Shortly after the study appeared, the third author, Daniel Wirth (a lawyer without a medical degree who had used a series of false identities over the years) pleaded guilty to conspiracy to commit fraud in shady business dealings. It was revealed that the researchers had failed to get informed consent from participants. Informed consent means a research participant “consents,” or voluntarily agrees to be part of a study, and is “informed,” i.e., he or she knows something about the study. This triggered an investigation by the United States Department of Health and Human Services. Over time, suspicions grew that researchers faked study data, making the study a case of scientific fraud. The reputations of all the authors, of the sponsoring university, and of the scholarly journal in which it appeared have been tarnished. A different study examined the healing power of prayer several years later. An article by Mitchell Krucoff appeared in 2005 (Krucoff et al., 2005) in the scholarly journal Lancet. In this study, 700 heart patients received prayers by Buddhist, M uslim, Jewish, and Christian congregations around the world. The authors fully disclosed all study details, answered all questions, and obeyed all ethical guidelines. The study design was straightforward, and there was no evidence of fraud. However, this study found that prayer had no effect on healing. The study authors stated that until more research is conducted, we do not know whether prayer has any effect on medical recovery.

3.1.2: Unethical but Legal Do not confuse being ethical with acting legally. A research action can be fully legal (i.e., not breaking any law) but clearly unethical (i.e., violating accepted standards of ethical research). This happens in other areas of life. You might be a dishonest, deceptive, and untrustworthy person who fails to keep your word and often lies. This makes you an immoral person, but you are not violating a law (unless you engage in deceptive business practices or lie while under a sworn oath). You may not have friends and few people will trust you, but you will not be sent to jail.

Becoming an Ethical Researcher 53

Figure 3.1 Typology of Legal and Ethical Actions in Research As shown in the figure, most research actions are both moral and legal. People canquickly recognize actions that are both illegal and immoral. A few rare instances occur when a research action is illegal but ethical. More common are research actions that are legal but violate ethical standards, because ethics arebroader and less explicit than the law. Acting within the law does not guarantee that you are acting ethically in research. You may want to seek guidance about ethics in research from several resources: colleagues, ethical advisory committees, institutional review boards, professional codes of ethics, and published discussions of research ethics. Ethical Legal

YES

Moral and Legal

Legal but Immoral

Illegal but Moral

Immoral and Illegal

3.2: The Ethical Issues Involved in Using People as Research Participants 3.2 Understand the origin and application of the rules of conduct in using humans in research Have you ever been a participant in a research study? If so, how were you treated? More than any other ethical issue, most attention in research ethics focuses on possible negative effects on the research participants. This is because of past situations of abuse and because total protection for research participants with absolute rights of noninterference would make research impossible. It is important to protect participants in research while still involving them in research studies. You must balance competing values in the many grey areas of ethics; however, the community of researchers, codes of ethics, and sometimes the law recognize a few clear prohibitions: • Never cause unnecessary or irreversible harm to research participants. • Always get voluntary consent from research participants before a study begins. • Never unnecessarily humiliate or degrade research participants. • Never release harmful information about specific individuals collected for research purposes. A simple rule guides you to act ethically: always show respect for the research participant.

As the person conducting a study, you have a clear ethical responsibility to provide research participants with basic protections.

3.2.1: The Origin of Ethical Principles for Research withHumans We can trace concerns over the treatment of research participants to medical studies in the early 1900s. This concern expanded after the public learned of gross violations of basic human rights in the name of research. The most notorious violations were “medical experiments” in Nazi Germany and Japan in the 1940s. Despite the mistreatment of people in the name of research, such as the Tuskegee syphilis study, each exposure of these incidents helped to extend and advance the discussion of ethical principles. In the syphilis study and in horrific medical experiments during World War II, vulnerable, powerless people suffered in the name of scientific research and advancing knowledge. No one voluntarily agreed to participate in a study or knew what was going to happen. This situation gave rise to the principle of voluntary participation. It would be wonderful to report that all abuse ended in the 1940s. However, incidents of unethical research have reappeared. Until the 1970s, U.S. medical researchers did not always provide full ethical protection to participants. For example, in 1940 U.S. researchers injected 400 prisoners with malaria to study the disease. They did not tell inmates about the nature of the experiment. Similar studies on malaria continued through 1946. As the U.S. military developed atomic weapons, researchers conducted studies on the effects of radioactive substances on people from the 1940s and the 1960s. In addition to prison inmates, researchers often used U.S. soldiers, hospital patients, or children with mental disabilities as research participants. In one case, researchers put radioactive material in the breakfast cereal of children.

54 Chapter 3 In other studies, the U.S. military gave unsuspecting people hallucinogenic drugs, such as LSD, to study their effects. During the 1960s, medical researchers injected patients at the Jewish Chronic Disease Hospital with live cancer cells and injected the hepatitis virus into children with developmental disabilities institutionalized at the illowbrook School. Through the 1970s, New York W researchers tested over 90 percent of new pharmaceuticals on prison inmates, despite increased questions about ethical protections.

Learning from History Nazi Doctors

alive. Vivisected prisoners included men, women, children, and infants. Prisoners had limbs amputated in order to study blood loss. Researchers froze the limbs (hands, arms, legs) and amputated them for study, or froze intact limbs and then thawed them to study the effects of untreated gangrene and rotting. Researchers also used humans as targets to test grenades, flamethrowers, germ-releasing bombs, chemical weapons, and explosive bombs. Other researchers looked at how long it took a person hung upside down to choke to death. When World War II ended, the Allies (mostly British, French, and American) brought Nazi doctors before a war crimes tribunal in Nuremburg, Germany. They found many of the researchers to be guilty of “crimes against humanity.” The trials resulted in the creation of an international set of ethical standards for conducting research with humans. Postwar trials charged few Japanese researchers with war crimes, and the events of Unit 731 received less publicity. This was because the Pacific War ended later, European involvement was less, and most of the victims were Asians, and therefore seen as less important given the prevailing racial bias at the time.

WRITING PROMPT It May Be Legal But Is It Ethical? Legal standards can be country specific, yet ethical standards are worldwide. Describe an example of a case, even outside of research, where legal and ethical standards did not match.

The Hippocratic Oath that physicians take pertains to the ethical practice of medicine. It says, “Never to do deliberate harm to anyone.” During the 1940s, respected scientific and medical experts in Germany violated this Oath. They conducted horrible acts on innocent men, women, and children for the purpose of studying them. The researchers used high research standards and carefully gathered data. They even published results in scholarly journals. Nonetheless, they acted unethically and failed to protect the research participants. Using the large populations in concentration camps, German researchers gassed, poisoned, and froze research participants to death. They injected participants with typhus and malaria to study the diseases. Researchers purposely exposed people to mustard gas and incendiary bombs to study how they caused injury, and they starved people to death to study the starvation process. At the same time that German research was occurring, the Japanese conducted similar horrific studies on humans at research Unit 731 to improve bacteriological warfare. As many as ten thousand people were research participants, both civilians and prisoners of war. Physicians performed vivisections (opening the body and cutting flesh) and infected prisoners with various diseases and then removed organs to study disease. These procedures were conducted while the patients were

The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

3.2.2: Protecting Research Participants from Harm Most discussions about not harming people in research focuses on medical research, but social research can also cause harm in several ways: • Physical harm or bodily injury • Great emotional distress or psychological harm • Legal harm and damage to a person’s career, reputation, or income. Certain types of harm are more likely in certain types of research (e.g., in experiments versus field research). As a researcher, you have a responsibility to be aware of potential harm and to take precautions to minimize the risk of harm to participants. The guiding ethical principle is that no person should experience harm as the direct result of participating in a research study.

Becoming an Ethical Researcher 55

Physical Harm Physical harm is rare in social research studies. The ethical rule is simple: Never—under any circ*mstance—purposely cause physical harm to a research participant.

To be ethical, you must anticipate risks, including basic safety concerns (e.g., safe buildings, furniture, and equipment). This means screening out high-risk participants (those with heart conditions, mental breakdown, seizures, etc.) if you will subject them to great stress. A researcher accepts moral and legal responsibility for injury due to participation in research. You should terminate a project immediately if you can no longer guarantee the physical safety of those involved. Of course, if you do research in a dangerous situation, you also want to protect yourself from harm.

Example Study Zimbardo Prison Study

Philip Zimbardo (Zimbardo, 1972) designed an experiment on prison conditions. Before the experiment, he gave male volunteer students personality tests and only included those in the “normal” range. He randomly divided the volunteers into two role-playing groups: guards and prisoners. The study took place in a simulated prison in the basem*nt of a Stanford University building. Zimbardo informed participants assigned to play prisoner that they would be under surveillance and would have some civil rights suspended, but that no physical abuse was allowed. He had the prisoners de-individualized (dressed in standard uniforms and called only by their numbers), and had the guards militarized (given uniforms, nightsticks, and reflective sunglasses), to make the simulation feel real. He told guards to maintain a reasonable degree of order. Guards

served 8-hour shifts while he had prisoners locked up 24 hours per day. The study was to last two weeks. Unexpectedly, participants got very caught up in their roles. The prisoners became passive, depressed, and disorganized. The guards became aggressive, arbitrary, and dehumanizing. By the sixth day, Zimbardo called off the experiment for ethical reasons. The risk of permanent psychological harm, and even physical harm, was too great.

Psychological Abuse, Stress, or Loss of Selfesteem Some social research studies place participants

in stressful, embarrassing, anxiety-producing, or unpleasant situations. By placing participants in realistic situations with psychological distress, we can learn about people’s responses in real-life, high-anxiety situations. However, is it unethical to cause discomfort? Researchers still debate the ethics of the famous Milgram obedience study. Some say that the precautions Milgram undertook and the great knowledge gained outweighed the potential psychological harm to participants. Others believe that the extreme stress and the risk of permanent harm were too great. Today, no one would conduct such an experiment due to a heightened sensitivity to the ethical issues involved. Social psychologists who study helping behavior often place participants in stressful, emergency situations to see whether they will lend assistance. For example, in a classic study, Piliavin and associates (1969) studied helping behavior in subways by faking a person’s collapse onto the floor. In the field experiment, the riders in the subway car were unaware of the experiment and did not volunteer to participate in it. The study’s findings were valuable, but the lack of informed consent and anxiety created were ethically controversial. Only experienced researchers should consider conducting a study in which they purposely induce great stress or anxiety. They must take all necessary precautions before inducing anxiety or discomfort. This includes consulting with others who have conducted similar studies as well as with mental health professionals. They should screen out high-risk populations, arrange for emergency interventions, and be prepared to end the study immediately if a dangerous situation arises. They must always obtain written informed consent and always debrief participants. Even with these safeguards, they can never create unnecessary stress. Unnecessary means beyond the minimal amount required for the desired effect. Any discomfort they create must have a very clear, legitimate research purpose. Knowing what “minimal amount” means comes with experience. It is best to begin with too little stress, risking a finding of no effect, than to create too much. In addition, it is best to work in collaboration with other researchers. Involving several sensitive researchers reduces the chance of making an ethical misjudgment.

56 Chapter 3

Example Study Milgram Obedience Study

Stanley Milgram’s studies on obedience (Milgram, 1963, 1965, 1974) are widely discussed. He wanted to learn how ordinary people could have carried out the horrors of the Holocaust under the Nazis. The study examined the impact of social pressure on people obeying authority figures. After signing informed consent forms, he assigned a volunteer-participant, in rigged random selection, to be a “teacher” while a confederate working for him was the “pupil.”

Figure 3.2 Arrangementof Milgram’s obedience

experiment.

S = pupil,L = teacher-participant, V = experimenter

The pupil was located in a nearby room, where the research participant could hear but not see the pupil. The pupil was connected to electrical wires. Milgram told the participant to test the pupil’s memory of word lists and to increase the electric shock level if the pupil made mistakes. The shock apparatus was clearly labeled with increasing voltage that indicated its danger. As the pupil increasingly made mistakes and

the teacher-participant turned switches, the pupil would make noises as if in severe pain. The researcher was always present and made quiet comments such as, “You must go on,” to the participant. As the voltage levels got higher, Milgram reported, “Subjects were observed to sweat, tremble, stutter, bite their lips, groan, and dig their fingernails into their flesh. These were characteristic rather than exceptional responses to the experiment” (Milgram, 1963:375). At the end, he told participants what really took place. The “pupil” was actually a confederate who was acting, and no one was shocked. The percentage of “teacher” participants who applied electrical shocks to very dangerous levels was dramatically higher than Milgram expected; 65 percent gave shocks at 450 volts. This study raised ethical concerns over using deception and about the extreme emotional stress participants experienced. Despite the precautions, many people believe that the degree of distress and risk of long-term emotional problems among the participants was too great.

Legal Harm A researcher is responsible for protect-

ing research participants from an increased risk of arrest simply because of their involvement in a study. If an increased risk of arrest is associated with research participation, few people will want to participate in research. Potential legal harm is one criticism of the “tea-room trade” study by Humphreys (1973). A related ethical issue occurs if you learn of illegal activity when collecting data. You must weigh the value of protecting the researcherparticipant relationship against potential harm to innocent people if you do not report what you discover. In the end, you alone, as the researcher, are morally and legally responsible. In field research on police, Van Maanen (1982:114–115) reported seeing police beat people and witnessing illegal acts and irregular procedures but said, “On and following these troublesome incidents, the choices I made followed police custom: I kept my mouth shut.” Field researchers who study the “seamy side” of society can face difficult ethical decisions. For example, when studying a mental institution, Taylor (1987) discovered the mistreatment and abuse of inmates by the staff. He had two choices: abandon the study and call for an immediate investigation, or keep quiet and continue with the study for several months, publicize the findings afterward, and then become an advocate to end the abuse. After weighing the situation, he followed the latter course and is now an activist for the rights of mental institution inmates. In some studies, observing illegal behavior may be central to the research project. A researcher who works closely with lawenforcement officials must face the question, “Are you an independent professional who ethically protects participants to advance knowledge in the long term, or are you a freelance undercover informant who is working for the police and is trying to catch criminals now?”

Becoming an Ethical Researcher 57

Example Study Tearoom Trade

Laud Humphreys studied male hom*osexual behavior (Humphreys, 1973). He focused on “tea-rooms”—places where anonymous sexual encounters occur. He observed about 100 men engaging in sexual acts in a public restroom in a park. To do this, Humphreys pretended to be a “watchqueen” (a voyeur and lookout). He also followed research participants to their cars and secretly recorded their license numbers. He next obtained the names and addresses of participants from law enforcement registers by posing as a market researcher. One year later, in disguise, Humphreys used a deceptive story about conducting a health survey to interview the men in their homes. Humphreys kept names in safety deposit boxes and took precautions. He significantly advanced knowledge about the men who frequent “tea-rooms” and overturned previous false beliefs about them. He learned that most of the men were married and had stable jobs. They were not isolated loners without stable relationships or jobs, as was previously thought. The study generated an ethical controversy. It was covert research based on deception. None of the participants voluntarily agreed to be in the study. Despite precautions, Humphreys put participants at risk of great legal and personal harm. Had he lost control of the names, someone might have used them to blackmail participants, destroy their marriages and careers, or initiate criminal prosecution against them.

3.2.3: Participation Must Be Voluntary and Informed A fundamental ethical principle is: Never coerce anyone into participating; participation must be voluntary at all times. The ethical principle of voluntary consent says getting permission is not enough. To make an informed decision, participants need to know in what activities you are asking them to participate. Researchers must make participants aware of their rights. They do this by providing participants with an informed consent statement about the study and ask participants to sign it. For survey research, we know that participants who receive a full informed consent statement

respond the same as those who do not. If anything, people who refuse to sign the statement are likely to guess or say, “No response,” to survey questions. A formal, signed informed consent statement is optional for most survey, field, existing statistics, and secondary data research, but it is mandatory for most experimental research. The general rule is, the greater the risk of causing possible harm to a research participant, the greater the need to obtain a written informed consent statement. Government agencies require informed consent in most studies involving people, with a few exceptions. You should get informed consent unless there are good reasons for not obtaining it (e.g., covert field research, use of secondary data, etc.). Making It Practical: Obtaining Informed Consent Informed consent upholds the principle of

voluntary participation and is mandated in various laws, regulations, and codes of ethics. There are a few exceptions to the use of consent, such as field research observation in a large public setting. In situations such as a survey questionnaire that do not ask for highly personal information, consent can be a short oral statement. Consent must always be in written form with a signature for research that involves physical discomfort, induced stress, or the collection of highly personal information. Informed consent statements contain the following eight elements: 1. A brief description of the purpose and research procedure, including how long the study will last 2. A statement of any risks or discomfort associated with participation 3. A guarantee of anonymity and the confidentiality of data records 4. Identification of who the researcher is and contact information for more information about the study 5. A statement that participation is voluntary and participants can withdraw at any time without penalty 6. A statement of any alternative procedures that may be used 7. A statement of any benefits or compensation that research participants may receive 8. An offer to provide a summary of the findings when the study is completed. Informed consent that discloses research details along with the researcher’s identification also protects research participants against fraudulent research. It lessens the chances that a con artist can use a bogus identity to defraud or abuse research participants, market products, or obtain personal information for unethical purposes. Secretly collecting information on people’s behavior without informed consent both violates their privacy and risks violating ethical standards.

58 Chapter 3

Learning from History Unethical Research byFacebook A research study on the social network site, Facebook, reported in a peer-reviewed scholarly journal, The Proceedings of the National Academy of Science (PNAS), in March 2014 (Kramera, Gillory, and Hanco*ck, 2014) caused a great deal of public discussion over ethics among academics, business and government officials in the U.S. and Europe.1 Professor Robert Klitzman, a leading expert on ethics at Columbia University said the study by researchers from Cornell University and the University of California at San F rancisco working with the “Core Science Team” at Facebook is “scandalous and violates accepted research ethics” (Klitzman, 2014). In the study, nearly 700,000 Facebook users were part of a twenty-day experiment without their knowledge. The researchers manipulated the users’ news feeds, reducing positive or negative content, and examined the emotional contents of the users’ subsequent posts. The goal was to manipulate people’s moods. The basic finding was that users who saw positive news feeds were more likely to say positive things in subsequent posts, and those who see negative ones do not. The Facebook study was unethical in the view of many researchers. The 1974 National Research Act in the U.S. states all researchers must respect the rights of research participants by explaining to them the purposes of the study and the procedures and foreseeable risks or discomfort. Researchers must have the study design pre-reviewed by the local Institutional Review Board (IRB) and get informed consent from participants before proceeding. The Cornell University IRB did review the study and did not believe informed consent or “opt out” provision was required, but others, including the journal editors, called it “a matter of concern” in a special note on the study. Facebook claims that its data use policy, which essentially says, “If you use Facebook everything you do on Facebook can be analyzed (i.e., we do not have to inform when we manipulate your feeds), exempted it from standard research ethics.” Facebook also said that such practices are common throughout the online industry. Nonetheless, university researchers are required to obey research ethics and obtain approval from the Institutional Review Board. In the end, as Meyer (2014) concluded, the Facebook study might have been technically legal but it likely violated established research ethics standards. It was unethical to involve people, many of whom were legal minors, in an experiment that manipulated their emotions without the participants having given prior informed consent to participate and without any opportunity to “opt out” if they wished.

The ethics of the controversial Facebook study is discussed in Albergotti andDwoskin (2014), Goel and Vindu (2014), Gorski (2014), and Klitzman and Appelbaum (2014).

WRITING PROMPT Does Informed Consent Work? In what ways does informed consent implement the principle of voluntary participation and help to stop abuse? Can you think of a situation where researchers would be overdoing it by getting informed consent? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

3.2.4: Limits to Using Deception inResearch Has anyone ever told you a half-truth or lie to get you to do something? How did you feel about it? Deception is a mild type of harm to participants; it reduces trust and honesty in human relations. When a study uses deception, voluntary participation and a person’s right not to participate can be a critical issue. Researchers debate whether using deception is ethically acceptable. Some say it is never acceptable. Others say it is ethical, but only for specific purposes and with strict conditions. Sometimes an experimental researcher has legitimate methodological reasons for deceiving participants or using misrepresentation. Experimenters often use deception when they expect that participants would modify their behavior or statements if they knew the study’s true purpose. For example, you want to study body posture. If you tell participants you are studying their body posture, they might adjust how they stand or sit because they know this is something you are observing. This could make it impossible for you to learn about their real body posture. Another use of deception is in covert field research. For example, you could not gain access to a research site if you told the truth, so you lie or hide your identity as a researcher. Say you want to conduct a field study on teens who use illegal drugs and commit minor crimes. If you told the teens when you first approached them you were going to carefully observe and study them, they might not cooperate or reveal all their actions. In any type of study, deception is never preferable if you can accomplish the same thing without using it. It is a last resort. Deception is acceptable only within strict limits, if you do the following: • Show that it has a clear, specific methodological purpose. • Use it only to the minimal degree necessary and for the shortest time. • Obtain informed consent and do not misrepresent any risks. • Always debrief (i.e., explain the actual conditions to participants afterward).

Becoming an Ethical Researcher 59

It is possible to obtain prior informed consent and use deception if you describe the basic procedures involved but conceal limited information about specific details. You could inform participants that they will sit alone in a room for 10 minutes, chat with one other participant for 5 minutes, and then complete a 15-item questionnaire, but not tell them your hypotheses or that the other participant in the room is secretly working for you. Experimental researchers often invent stories about the purpose of a study to distract participants from the true purpose of a study. For example, a researcher studying male-female eye contact might tell participants the study is about college student opinions on current affairs. Some field researchers use covert observation to gain entry to a field research setting. In studies of cults, small extremist political sects, and illegal or deviant behavior, it may be impossible to conduct research if you disclose your purpose. If a covert stance is not essential, the ethical rule is very clear: do not use it. If you are unsure of whether covert access is necessary, then use a strategy of gradual disclosure. Begin with limited disclosure, then reveal more as you learn more about a setting and participants. When in doubt, it is best to err in the direction of disclosing your identity and purpose. Covert field research is controversial among researchers. Many feel that all covert research is unethical. Even researchers who accept covert research in certain situations place limits on its use. It is only acceptable when overt observation is impossible. If you use covert research, you must inform participants of the covert observation immediately afterward and give them an opportunity to express concerns. Using deception and covert practices can increase mistrust and cynicism and reduce public respect for social research. Secretly spying on people without their permission and lying to get information carries real dangers. It is analogous to being an undercover government informer in a nondemocratic society.

believes that it will help the criminal. The criminal is being coerced to participate; the alternative is two years in prison. In such a case, the researcher and others must honestly decide whether the benefits to the criminal and to society and far outweigh the ethical prohibition on coercion. Moreover, the coercion must be limited. Even so, such decisions are risky. History shows many cases in which researchers forced powerless participants to be in a study allegedly to help them, but later it turned out that the research participants received few benefits but experienced great harm. Perhaps you have been in a social science class in which a teacher required you to participate in a research project. This is a special case of coercion. Usually, it is ethical. The legitimate justification for it is that students learn more about research when they experience it directly in a realistic situation. Such minor, limited coercion is acceptable as long as the teacher meets three conditions: • The participation in research is attached to a clear educational objective of the specific course. • Students have a choice of the research experience or an alternative activity of equal difficulty. • The teacher follows all other ethical principles of conducting research.

WRITING PROMPT Is Deception Acceptable? Deception in research and covert research are controversial ethical issues among researchers. Some justify them by saying they are the only way to obtain valuable information about human behavior. Others say they are always unethical, never justified and should be banned. Do you think deception and covert research are justifiable in certain situations or do you think should they be banned? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

Avoid Coercion Coercion can be physical, social,

legal, professional, financial, or other pressure put on people so that they will agree to participate in a study. The general rule is clear: Never coerce people to participate. This includes offering people a special benefit that they cannot attain other than through participation in a study. For example, it is unethical for a commanding officer to order a soldier to participate in a study, for a professor to require a student to be a research participant to pass a course, or for an employer to demand that an employee complete a survey as a condition of continued e mployment. The rule against coercing people to participate in a study can become tricky in specific situations. For example, a convicted criminal faces the alternative of two years’ imprisonment or participation in an experimental rehabilitation program. The convicted criminal does not believe in the benefits of the program, but the researcher

3.2.5: Privacy, Anonymity, andConfidentiality How would you feel if someone learned private details about your personal life and shared them with the public without your permission or knowledge? Because researchers often learn intimate details about participants, they take special precautions to protect research participants’ privacy. Privacy When you study people, you learn details

about them. As survey research probes into beliefs, backgrounds, and behaviors, it often reveals intimate, private details. Experimental researchers sometimes use two-way mirrors or hidden microphones to “spy” on a person’s behavior. Even if research participants are aware they are in a study, they may not know what a researcher is

60 Chapter 3 looking for. Field researchers observe private aspects of behavior or eavesdrop on personal conversations. They have studied people in public places (e.g., in waiting rooms, walking down the street, in classrooms, etc.), but some “public” places are more private than others are (consider, for example, the use of periscopes to observe people who thought they were alone in a public toilet stall). Eavesdropping on conversations and observing people in quasi-private areas raises ethical concerns. When you conduct research, you need to protect the privacy of participants. To be ethical, you can only violate a participant’s privacy to the minimum degree necessary and collect private information only for a legitimate research purpose. In addition, you must take several steps to protect the information you have learned about participants from public disclosure. This takes two forms that many people confuse: anonymity and confidentiality. Anonymity Anonymity means to remain anonymous

or nameless. No one can trace information back to a specific individual. Methods of protecting anonymity vary by research technique. In survey and experimental research, do not collect names or addresses of participants and only refer to them by a code number. In survey research by mail, include a code on the questionnaire to determine which respondents failed to respond. Respondents are not anonymous during that phase of the study. After completed questionnaires are returned, researchers will not keep respondent names. To protect anonymity, you should not able to narrow down which person submitted a particular questionnaire based on details about them and a list of names. This is why a promise of anonymity could be breached unintentionally in small samples. For example, you survey students at a small college of 250 students and ask questions including age, sex, religion, hobbies, and hometown. You notice answers from a 22-year-old Jewish male born in Stratford, Ontario, whose hobby is to be on the football team. Such information among a small group allows you to learn who the specific individual is, even if you did not ask his name. This violates a promise of anonymity. Protecting anonymity in field research is difficult. You learn details about research participants and their names. You can disguise some details and give false names, but even this does not always work. In one community study, researchers invented a false town name, “Springdale,” and altered facts to protect the anonymity of the people studied, but they did not do enough. As a result, readers could identify the town and particular individuals when the study appeared as a book, Small Town in Mass Society (Vidich and Bensman, 1968). Town residents became very upset about how the researchers portrayed them. They even staged a parade mocking the researchers. An additional issue is that when you use fictitious information, the gap between what you studied and what you report may raise questions about what you found and what you made up.

Confidentiality Even if you cannot protect anonymity, you should always protect participant confidentiality. Anonymity means that no one can learn the identity of specific individuals. Confidentiality allows you to attach the information to particular individuals, but keep it secret from public disclosure. You only release data in ways that do not permit anyone else to link specific individuals to information. You do this by presenting data publicly only in an aggregate form (e.g., as percentages, statistical means, etc.). As you can see in Table 3.1, it is possible for you to provide anonymity without confidentiality, or vice versa; however, they usually go together.

Table 3.1 Anonymity and Confidentiality Anonymity Confidentiality

YES

Gather data so it is impossible for anyone to link details to any name and release findings in aggregate form.

Privately link details about a specific participant to a name, but only publicly release findings in aggregate form.

Release details about a specific participant to the public, but withhold the name and details that might allow someone to traceback to the person.

Unethical Reveal publicly details about a person with his/her name.

The table below includes an example of each situation listed.

Table 3.2 Examples of Anonymity and Confidentiality Example

Situation

You conduct a survey of 100 people but do not know names of any of the participants and only release data as percentages of the total.

Anonymity with confidentiality

You conduct a survey of 100 people but do not know names. You report all the details about the specific answers given by a person publicly but block out a few details to make it impossible for anyone to track down the person.

Anonymity without confidentiality

You conduct a survey of 100 people and have each person’s name listed on his or her questionnaire but only publicly release the data aspercentages of the total.

Confidentiality without anonymity

You conduct a survey of 100 people and have each person’s name on his or her questionnaire. You publicly release a person’s answers with thename, or with enough details to allow easydiscovery of the person’s name.

Neither anonymity nor confidentiality (this is unethical)

In a few situations, other principles overrule the principle of protecting research participant privacy. One exception is a clear, immediate danger to a person’s safety, such as learning that a participant is considering suicide or plans to injure or kill another person. For example, you study parents, and during an interview with a father, you learn that he is abusing his child physically or sexually. In

Becoming an Ethical Researcher 61

this case, you must weigh protecting participant privacy against the harm that the child would experience. The ethical course of action is to protect the child from imminent harm and notify the authorities. Ethics often requires weighing completing demands.

Learning from History Not Breaking the Confidence Guarantee

research means. If you wish to have “incompetent” people (e.g., children, mentally disabled, etc.) participate in a research study, you must meet two minimal conditions: • The person’s legal guardian/parent grants written informed consent permission. • You closely follow all standard ethical rules to protect participants from any form of harm. Anonymity and confidentiality become more complicated if you study special populations. In many large bureaucracies, people in positions of authority may restrict research access unless you violate confidentiality and give them information about participants. Examples 1. You want to study drug use and sexual activity among high school students. School authorities only agree if you give them the names of all drug users and sexually active students. They might say they want to assist the students with counseling and to inform the students’ parents. An ethical researcher must refuse to continue. If the school officials really wanted to assist the students and not use researchers as spies, they could develop their own outreach program.

Social researchers can pay a high personal cost for being ethical. Washington State University Professor Rik Scarce went to jail. Professor Scarce had studied extremist social-political movements using accepted field research techniques. He introduced himself, slowly gained access and won trust, explained his research interests, and guaranteed confidentiality. He attended a local group’s meetings, talked with activists and leaders, and spent time observing. In a second study of extremist groups, Dr. Scarce was studying a radical animal liberation group when police suspected the group leader of breaking into a nearby animal facility, releasing the research animals, and causing $150,000 in vandalism damage. Although some past court rulings appeared to offer social researchers protection, other rulings have not upheld confidentiality protections for social research data. When police asked for all of Dr. Scarce’s research notes, he followed professional ethical rules and principles. He refused to break the research confidentiality guarantee and hand over his notes or testify to a grand jury about his observations. As a result, he spent 159 days in a Spokane, Washington, jail for contempt of court (see Scarce, 1994,1999).

3.2.6: Extra Protections for SpecialPopulations Some research participants may not be able to give true voluntary informed consent. Special populations such as students, prison inmates, employees, military personnel, the homeless, welfare recipients, children, or the developmentally disabled may not be fully capable of giving consent freely. Some may agree because it is a way to a desired good—such as higher grades, early parole, promotions, or additional services; others may not understand what

2. You conduct a survey of workplace conditions. A manager asks to see all complaints about work conditions in a survey, along with the names of the employees making a complaint. The supposed reason for the survey is to address the complaining individual’s concerns, not to single out complainers. An ethical researcher protects participant privacy and only releases results without names to protect the employees from potential retaliation.

Privacy rules operate to protect special population research participants from legal and physical harm. Example Draus and associates (2005) protected the participants in a study of illegal drug users (a special population) in rural Ohio. To do this, they conducted interviews in large multiuse buildings, avoided any reference to illegal drugs in their written documents, did not include the names of drug dealers and locations, and did not affiliate with drug rehabilitation services. This is because the rehabilitation services had ties to law enforcement. They (2005:169) noted, “We intentionally avoided contact with local police, prosecutors, or parole officers,” and, “surveillance of the project by local law enforcement was a source of concern.”

WRITING PROMPT Is Privacy Protection Outdated? Do you think the attempts at strict privacy protection in social research are ethically necessary, or do you believe they should be abandoned as outdated in today’s world of on-going surveillance? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

62 Chapter 3

3.2.7: Formal Protections for Research Participants Different countries have slightly different protections. In the United States, the U.S. Department of Health and Human Services Office for the Protection from Research Risks issues regulations to protect research participants. Federal regulations follow a biomedical model and protect participants from physical harm. It is one federal government agency, and technically its rules apply only when federal money is involved, but in practice, all other government agencies and researchers follow its guidance. Local governments, hospitals, universities, and private companies model their internal policies on the federal rules. Other U.S. government rules require the creation of institutional review boards (IRB) at all research institutes, medical facilities, colleges, and universities where research with humans occurs. An institutional review board has a mix of researchers and nonresearchers. Its members review research procedures at a proposal or preliminary stage, making certain that ethical principles are upheld. Some forms of research are exempt from a formal, full review by the IRB. These include educational tests, normal educational practice, most nonsensitive survey questionnaires, observation of public behavior, and studies of existing public data in which individuals cannot be identified. Submitting a proposal to an IRB for review requires a little extra time and planning. IRB members are an “extra set of eyes” looking at a research design to ensure that research participants will be fully protected. Most professionals (e.g., physicians, attorneys, family counselors, social workers, and others) have organizations that developed a written code of ethics, peer review boards, or licensing regulations. A code of ethics is a written statement of ethical rules that identify proper and improper behavior. Most professional social science associations have codes of ethics that represent a consensus of professionals on ethics. Although not all researchers may agree on every ethical issue, they uphold ethical

standards as members of a profession. These include the following: • Ensure that participants have voluntarily consented to be in the study. • Avoid unnecessary physical and mental suffering. • Avoid any research where death or disabling injury to participants is likely. • End a research study immediately if its continuation is likely to cause injury, disability, or death. • Highly qualified people using the highest levels of skill and care should conduct research studies. • Study results should be for the good of society and unattainable by any other method. Making It Practical: Codes of Ethics We can trace today’s codes of ethics for human research to the Nuremburg Code created after war crime trials in Germany at the end of World War II. The codes of most professional organizations, such as nursing, social work, public opinion research, psychology, or sociology are not identical but they do overlap a great deal. Professional associations create codes of ethics and hear about possible violations, but there is no formal policing of the codes. The penalty for a minor ethical violation rarely goes past public embarrassment and a letter of complaint. Those who commit a serious ethical violation, even if they violated no law, will face loss of reputation, loss of employment, a ban on the research findings being published, or restrictions from future jobs. Besides making explicit the beliefs of the research community and providing researchers with guidance, codes of ethics help universities and other institutions defend legitimate, ethical research against political or other pressures. If researchers receive unjustified demands to stop legitimate research or to reveal protected details about research participants, written codes of ethics that are widely endorsed within the research community provide an important line of defense.

Ethics Code of theAmerican Association for Public Opinion Research We, the members of the American Association for Public Opinion Research, subscribe to the principles expressed in the following code. Our goals are to support sound and ethical practice in the conduct of public opinion research and in the use of such research for policy- and decision-making in the public and private sectors, as well as to improve public understanding of public opinion and survey research methods and the proper use of public opinion and survey research results.

We pledge ourselves to maintain high standards of scientific competence and integrity in conducting, analyzing, and reporting our work; in our relations with survey respondents; with our clients; with those who eventually use the research fordecision-making purposes; and with the general public. We further pledge ourselves to reject all tasks or assignments that would require activities inconsistent with the principles of thiscode.

Becoming an Ethical Researcher 63

The Code I.

Principles of Professional Practice in the Conduct of Our Work A. We shall exercise due care in developing research designs and survey instruments, and in collecting, processing, and analyzing data, taking all reasonable steps to assure the reliability and validity of results. 1. We shall recommend and employ only those tools and methods of analysis that, in our professional judgment, are well suited to the research problem at hand. 2. We shall not knowingly select research tools and methods of analysis that yield misleading conclusions. 3. We shall not knowingly make interpretations of research results that are inconsistent with the data available, nor shall we tacitly permit such interpretations. 4. We shall not knowingly imply that interpretations should be accorded greater confidence than the data actually warrant. B. We shall describe our methods and findings accurately and in appropriate detail in all research reports, adhering to the standards for minimal disclosure specified in Section III. C. If any of our work becomes the subject of a formal investigation of an alleged violation of this Code, undertaken with the approval of the AAPOR Executive Council, we shall provide additional information on the survey in such detail that a fellow survey practitioner would be able to conduct a professional evaluation of the survey.

II. Principles of Professional Responsibility in Our Dealings with People A. The Public: 1. When preparing a report for public release we shall ensure that the findings are a balanced and accurate portrayal of the survey results. 2. If we become aware of the appearance in public of serious inaccuracies or distortions regarding our research, we shall publicly disclose what is required to correct these inaccuracies or distortions, including, as appropriate, a statement to the public media, legislative body, regulatory agency, or other appropriate group, to which the inaccuracies or distortions were presented. 3. We shall inform those for whom we conduct publicly released surveys that AAPOR standards require members to release minimal information about such surveys, and we shall make all reasonable efforts to encourage clients to subscribe to our standards for minimal disclosure in their releases. B. Clients or Sponsors: 1. When undertaking work for a private client, we shall hold confidential all proprietary information obtained about the client and about the conduct and findings of the research undertaken for the client, except when the dissemination of the information is expressly authorized by the client, or when disclosure becomes necessary under the terms of Section I-C or II-A of this Code. 2. We shall be mindful of the limitations of our techniques and capabilities and shall accept only those research assignments that we can reasonably expect to accomplish within these limitations. C. The Profession: 1. We recognize our responsibility to the science of survey research to disseminate as freely as possible the ideas and findings that emerge from our research. 2. We shall not cite our membership in the Association as evidence of professional competence, since the Association does not so certify any persons or organizations. D. The Respondent: 1. We shall avoid practices or methods that may harm, humiliate, or seriously mislead survey respondents. 2. We shall respect respondents’ concerns about their privacy. 3. Aside from the decennial census and a few other surveys, participation in surveys is voluntary. We shall provide all persons selected for inclusion with a description of the survey sufficient to permit them to make an informed and free decision about their participation. 4. We shall not misrepresent our research or conduct other activities (such as sales, fund raising, or political campaigning) under the guise of conducting research. 5. Unless the respondent waives confidentiality for specified uses, we shall hold as privileged and confidential all information that might identify a respondent with his or her responses. We also shall not disclose or use the names of respondents for nonresearch purposes unless the respondents grant us permission to do so. 6. We understand that the use of our survey results in a legal proceeding does not relieve us of our ethical obligation to keep confi nonymity. dential all respondent identifiable information or lessen the importance of respondent a III. Standards for Minimal Disclosure Good professional practice imposes the obligation upon all public opinion researchers to include, in any report of research results, or to make available when that report is released, certain essential information about how the research was conducted. At a minimum, the following items should be disclosed. 1. Who sponsored the survey, and who conducted it 2. The exact wording of questions asked, including the text of any preceding instruction or explanation to the interviewer or respondents that might reasonably be expected to affect the response 3. A definition of the population under study, and a description of the sampling frame used to identify this population 4. A description of the sample design, giving a clear indication of the method by which the respondents were selected by the researcher, or whether the respondents were entirely self-selected

64 Chapter 3 5. Sample sizes and, where appropriate, eligibility criteria, screening procedures, and response rates computed according to AAPOR Standard Definitions. At a minimum, a summary of disposition of sample cases should be provided so that response rates could be computed. 6. A discussion of the precision of the findings, including estimates of sampling error, and a description of any weighting or estimating procedures used 7. Which results are based on parts of the sample, rather than on the total sample, and the size of such parts 8. Method, location, and dates of data collection.

From time to time, AAPOR Council may issue guidelines and recommendations on best practices with regard to the release, design, and conduct of surveys.

Summary Review

Basic Principles of Ethical Research • Accept responsibility for all ethical decisions and the protection of research participants. • Use the research techniques that are most appropriate for a topic or situation. • Follow accepted methodological standards and strive for high accuracy. • Detect and remove any threats of harm to research participants. • Never exploit research participants for personal gain. • Get informed consent from the research participants before beginning. • Treat the research participants with dignity and respect at all times. • Only use deception if absolutely needed, and always debrief participants afterward.

sponsors believe that because they are paying, they can ask researchers to compromise ethical research standards as a condition of employment or part of a contract for research. While few would ask a medical doctor to prescribe medications unethically or a lawyer to violate professional ethics in a court of law, some sponsors are not aware of or are unconcerned about professional ethics in social research. If a sponsor makes an illegitimate demand, you have three basic choices: be loyal to an organization and cave in to the sponsor, exit from the situation by quitting, or voice opposition and become a whistle-blower. You need to set ethical boundaries beyond which you refuse a sponsor’s demands and choose your own course of action. Whatever the case, it is best to consider ethical issues early in a relationship with a sponsor and to express concerns up front. While sponsors often provide essential funding for research, and most allow great autonomy to professional researchers, some interfere with the research process in unethical ways. This occurs in three main ways: to require certain research findings, to restrict how research is done, and to suppress unwanted results.

• Honor all guarantees of privacy, confidentiality, and anonymity you make to participants. • Be candid and honest when interpreting and reporting study results.

Tips for the Wise Consumer

• Identify the sponsors of funded research to participants and to the public.

Who Paid for a Study?

• Release all details of the study procedures with the results.

It is unethical to hide the identity of a research sponsor. You should tell study participants who the sponsor is and inform the readers of research reports. Participants in a study have a right to know the sponsor. Telling participants is rarely controversial, but it becomes tricky in a few instances. For example, a pro-choice organization sponsors a study to look at the attitudes of members of religious groups opposed to abortion. The organization asks that you not reveal the sponsor to participants. You must balance the ethical rule to reveal a sponsor’s identity against the sponsor’s desire for confidentiality and possible bias or reduced cooperation by study participants. In general, unless you have a very clear, strong methodological reason for not doing so (such as reduced cooperation and strong bias), tell participants of the sponsor of a study. If telling participants of the sponsor will create a bias or noncooperation, then wait until after you have gathered the data. When reporting study results, the ethical mandate is unambiguous: You must always reveal sponsors who fund a study.

• Act with integrity and adhere to the behaviors outlined in professional codes of ethics.

3.3: How Do Sponsors Affect Research Ethics? 3.3 Identify some of the tactics used by research sponsors to aid their causes You might find a job in which you are assigned to conduct research for a sponsor—an employer, a government agency, or a private firm. Special ethical issues can arise when a sponsor pays for research, especially applied research. Some

Becoming an Ethical Researcher 65 One study can have multiple sponsors, especially if it is a large one; you should list all the sponsors. Government agencies, foundations, or nonprofit organizations fund most research studies. Here is an example of sponsor information from an article by David Kirt and Andrew Papachristos (2011) on Neighborhood Violence: “This research was supported by grant 2008-IJ-CX-0012 from the National Institution of Justice. The Project on Human Development to the Chicago Neighborhoods was conducted with support of the John D. and Catherine T. MacArthur Foundation, the National Institution of Justice, and the National Institute of Mental Health.”2 If you see no funding source listed, the study was probably conducted as part of the researcher’s professional duties or supported by an employer. As you read more research reports, you may notice that certain organizations regularly sponsor studies on a topic.

3.3.1: Arriving at Particular Findings A sponsor might tell you, directly or indirectly, what results you should come up with before you conduct a study. The ethical choice is to refuse to continue if told you must reach specific findings as a precondition for doing research. Legitimate research does not have restrictions on the possible findings. In an ethical, legitimate study, you will not know the findings for certain until after you have gathered the data and completed the study.

3.3.2: Limits on How toConductStudies Sponsors can legitimately set some conditions on research techniques used (e.g., survey versus experiment) and limit costs for research. However, as a researcher, you must follow generally accepted research standards. Often there is a trade-off between research quality and cost. You should give a realistic appraisal of what you can accomplish for a given level of funding. If you cannot uphold generally accepted standards of research, refuse to do the research. Unfortunately, a few sponsors care little for the actual results or truth and have little respect for research or its ethical principles. They only see research as “a cover” to legitimate a predetermined decision or as a way to deflect criticism. They are abusing the reputation of research and its integrity to advance their own narrow goals. If a sponsor requests the use of illegitimate research techniques (e.g., a biased sample or leading survey questions), the ethical researcher refuses to cooperate. Ethical violations harm the sponsor, researchers, the scientific community, and society in general. You must make a moral decision: Will you prostitute your reputation and skills to give sponsors whatever

This information is included in a footnote on the first page (p. 1190) of the article.

they want, even if it is unethical? Or, are you an honorable professional who teaches, guides, or even opposes sponsors based on widely accepted research principles?

3.3.3: Suppressing Findings Perhaps you conduct a study, and the findings make the sponsor look bad. The sponsor decides to suppress thestudy’s results. This kind of situation happens often inapplied research in both the private and public sector. In one case, a state government created a lottery commission to examine starting a government-sponsored lottery. Some politicians and members of the public asked for a study on the likely effects of a state lottery, so the commission hired a sociologist with expertise in that area. After she completed the study, but before releasing the report to the public, the commission asked her to remove sections of the report that discussed the negative social effects of gambling. They tried to eliminate sections of the report that predicted that the lottery would cause a large increase in compulsive gamblers and the recommendation that the state create social services to help them. The commission ordered and paid for the study, but the researcher felt a professional ethical obligation to show the public the full, uncensored report. Unless she went beyond the commission and released the complete report, the public would see a distorted, biased picture of study findings. You will read more about political influence later in this chapter, but even the U.S. federal government has a record of suppressing research findings that contradict the political goals of high officials. In 2004, many leading scientists, including Nobel laureates, medical experts, former federal agency directors, and university presidents, voiced their concern over the dramatic increase in the government’s misuse of research. One complaint was that government officials had suppressed research findings that disagreed with their political goals. In addition, they suppressed findings on the poor safety or pollution records of industries that were major political campaign contributors. Researchers employed by government agencies reported that their politically appointed managers had suppressed findings or omitted important technical information to advance nonscientific, political goals that the findings had contradicted.

Learning from History Why So Little Research on Gun Violence or Medical Marijuana? For a few high-profile topics in the U.S., including gun violence and medical marijuana, there is great public interest but very little solid scientific research to guide policy decisions. The reason

66 Chapter 3 has little to do with a lack of interest by researchers; rather, politicians responding to advocacy groups have blocked funding or permission to do research.3 In 1970, the United States Congress “failed to follow its usual review process dictated by the Controlled Substances Act that requires scientific evaluation and testimony before legislative action. It declared cannabis illegal in the absence of such evidence” (Bostwick, 2012:181). Since then, researchers have been prohibited from studying it. As Bostwick (2012: 182) noted, “the political climate at the federal level has essentially quashed the type of research that is routine before commercial introduction of new drugs.” As a result, the gradual introduction of “medical marijuana” and cannabis legalization by several states has proceeded based on opposing political pressure, not based on neutral evidence from scientific study. Likewise, in 1996 the U.S. Congress enacted strict restrictions and cut Center for Disease Control research on the issue of firearms and violence, and funding dropped by 96 percent. This effectively stopped government-funded research on the issue until a Presidential order in 2014 in the wake of the “Sandy Hook” mass shooting of elementary school children. As a result, the amount of research-based knowledge has been miniscule. In both situations, politicians and special interest groups feared that objective, neutral research might uncover findings that could contradict their ideological position. They preferred a situation of ignorance, where their ideological position would prevail in the absence of evidence, to a climate of open inquiry and research-based truth seeking. Pay attention to who paid for research and when doing sponsored research, negotiate conditions for releasing findings prior to beginning the study or signing a contract. It is best to begin with an explicit guarantee that you will only conduct ethical research. It is legitimate to delay the release of findings to protect the identity of informants, to maintain access to a research site, or to protect your personal safety. It is not legitimate to censor findings because a sponsor does not want to look bad or wants to protect its reputation. The researcher directly involved and knowledgeable about a study shoulders a responsibility for both conducting the research and making its findings public.

Blowing the Whistle Whistle-blowing occurs

when a researcher informs an external audience of a serious ethical problem that is being ignored. It is never a first step; rather it occurs after the researcher has repeatedly attempted to inform superiors and fix the problem internally. The whistle-blowing researcher must believe that the situation is a serious breach of ethics and the organization will not end it without public pressure. Whistle-blowing is risky in several ways: • Outsiders may not be interested in the ethical abuse and simply ignore it.

• Outsiders might not care about ending unethical behavior, but instead use the exposure of unethical behavior to advance their own goals. • Managers will try to protect the organization and discredit the whistle-blower. • The whistle-blower often experiences emotional distress and strained relations, and even lawsuits. • Future employers may not trust the whistle-blower and may avoid hiring him or her. A whistle-blower needs to be prepared to make sacrifices—loss of a job or no promotions, lowered pay, an undesirable transfer, abandonment by friends at work, or legal costs. There is no guarantee that doing the ethicalmoral thing will end the unethical behavior or protect an honest researcher from retaliation.4

3.4: Politics and Social Research 3.4 Recognize the impact that politics and large corporations have on research The ideals of a free, open, and democratic society include advancing and sharing knowledge. People have a right to study and inquire into any question and to share their findings publicly. Ethical issues largely address moral concerns and standards of professional conduct that are usually under the researcher’s control. Political concerns can also influence and interfere with the research process. Organized advocacy groups, powerful interests, government officials, or politicians may try to restrict or control the direction of research. In the past, powerful political interests and groups have tried to stop research or the spread of legitimate research findings to advance their own political goals. They have used their political power to threaten researchers or their employers, to cut off research funds, to harass individual researchers and ruin their careers, and to censor publication of findings that they disliked. Politically powerful groups have directed research funds away from studying questions that researchers saw as important and toward studies that supported the policy positions of their political views. Members of the U.S. Congress have targeted and removed funds for social research projects that panels of independent scientists evaluated as being well designed and critical to advanced knowledge. Often the reason has been that the politicians personally disliked the study topics (e.g., sexual behavior of teens, illegal drug users, voting behavior). Politicians are not the only

See American Psychological Association (Crawford [2014], Gutting [2012], Jamieson [2013], and Underwood [2013] on gun violence). See Bostwick (2012), Crawford (2014), and Harris (2010) on medical marijuana.

Summary Review Table 6.2 Survey Question Writing Pitfalls Things to Avoid

Not Good

A Possible Improvement

1. Jargon, slang, abbreviations

Did you drown in brew until you were totally blasted last night?

Last night, about how much beer did you drink?

2. Vagueness

Do you eat out often?

In a typical week, about how many meals do you eat away from home, at a restaurant, cafeteria, or other eating establishment where you pay for food?

3. Emotional language 4. Prestige bias (also Double-barreled)

The respected Jasper Commission documents that a staggering $500 Billion of your tax dollars are being wasted on public schools through poorly trained or incompetent teachers, outdated curriculum, weak management unable to deal with unions, “and other wasteful practices. Is cutting excessive public school spending and providing vouchers that will allow parents to choose effective private schools a top priority for you?”

How important is it to you to reduce wasteful public school spending? Very Important Somewhat Important Neither Important or Unimportant Somewhat Unimportant Not Important at All

5. Double-barreled questions

Do you support or oppose raising social security benefits and increased spending for the military?

Do you support or oppose raising social security benefits? Do you support or oppose increasing spending on the military?

6. Confuse beliefs about a hypothesis with testing the hypothesis

Do you think more educated people smoke less?

What is your education level? Do you smoke cigarettes?

7. Leading questions

Did you do your patriotic duty and vote in the last election for mayor?

Did you vote in last month’s mayoral election?

8. Issues beyond respondent capabilities

Two years ago, how many hours did you watch TV every month?

In the past two weeks, about how many hours do you think you watched TV on a typical day?

9. False premises

When did you stop beating your girl/boyfriend?

Have you ever slapped, punched, or hit your girl/boyfriend?

10. Distant future intentions

After you graduate from college, get a job, and are settled, will you invest a lot of money in the stock market?

Do you have definite plans to put some money into the stock market within the coming two months?

11. Double negatives

Do you disagree with those who do not want to build a new city swimming pool?

There is a proposal to build a new city swimming pool. Do you agree or disagree with the proposal?

12. Unbalanced responses

Did you find the service at our hotel to be: Outstanding, Excellent, Superior, or Very Good?

Please rate the service at our hotel: Outstanding, Very Good, Adequate, or Poor.

124 Chapter 6 Making It Practical: Improving Unclear Questions Here are three survey questions written by

experienced professional researchers. They revised the original wording after a pilot test revealed that 15 percent of respondents asked for clarification or gave inadequate answers (e.g., don’t know). As you can see, question wording is an art that may improve with practice, patience, and pilot testing (see Table 6.3).

Table 6.3 Responses to Question and Percentage Askingfor Clarification Original Question

Problem

Revised Question

Do you exercise or play sports regularly?

What counts as exercise?

Do you do any sports or hobbies, physical activities, or exercise, including walking, on a regular basis?

What is the average number of days each week you have butter?

Does margarine count as butter?

The next question is just about butter—not including margarine. How many days a week do you have butter?

[Following question on eggs] What is the number of servings in a typical day?

How many eggs is a serving? What is a typical day?

Responses to Question

On days when you eat eggs, how many eggs do you usually have?

Percentage Asking for Clarification

Original

Revision

Original

Revision

Exercise question (% saying “yes”)

48%

60%

Butter question (% saying “none”)

33%

55%

18%

13%

Egg question (% saying “one”)

80%

33%

Source: Adapted from Fowler (1992).

Avoiding leading questions requires some thought. The question, “Stable gay and lesbian couples have same rights as any other law-abiding citizens. Do you agree that your fellow citizens who are gay or lesbian should have equal rights to the benefits and responsibilities of a legal marriage as anyone else?” may lead respondents to state that they favor same-sex marriage even if they are uncertain. The phrases about the “same rights as other law-abiding citizens” and “your fellow citizens” make it leading. You can state leading questions to get a positive or a negative answer. For example, “Should we change laws to give sexually deviant people the privileges and benefits of a marriage, or should we continue to uphold the natural basis of a marriage as only between one man and one woman as it has been for centuries worldwide in custom, law, and religious teachings?” leads a respondent to disagree. Leading questions are just one of several potential pitfalls to avoid when writing survey questions.

WRITING PROMPT Leading Questions Many survey questions asked in public opinion polls about public issues are “leading” to some degree (especially those by political or advocacy groups, or by companies checking customer satisfaction). Write a question that you believe to be “leading.” Describe what makes it a leading question. The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

6.3.2: Getting Answers to Survey Questions Should you use an open- or closed-ended response format?

In an open-ended response format (also called unstructured or free response), respondents can give any answer, whereas in the closed-ended response format (also called structured or fixed response), they must choose among fixed answer choices. Each format has advantages and disadvantages. You must decide which format is most appropriate for the conditions or purpose of your particular survey. This depends on your study’s goal and the practical limitations of your research project. An open-ended response format takes more time for respondents to answer than a closed-ended format. It also requires a respondent to have minimal writing or verbal skills. It is easy for respondents to get off track and give unrelated information in the open-ended format. Interviewers or recorders must get down extensive verbatim answers. Researchers must develop a coding system and place open-ended responses into codes for data analysis, which is more complex and time-consuming than for data in the closed-ended format. This makes the open-ended format impractical for all but small-scale samples (under 200 people) with a few questions (under 20). Nonetheless, it is very useful for exploratory research, when we know little about an issue. It is also more effective when the goal is to capture a respondent’s thinking process. Writing survey questions with the closed-ended response format requires making some decisions. • How many response choices should you offer? • Should you offer a middle or neutral choice? • How should you order answer choices? • How should you measure the direction of answers? Making such decisions is not easy. For example, two response choices are too few, but more than five choices can create confusion. You want to measure meaningful

The Survey 125

istinctions and not collapse them. Very specific answer d choices yield more information, but being too specific may create confusion for the respondents. You can offer more choices and get more information without confusion if you rephrase a yes/no question, “Are you satisfied with your dentist?” by using a Likert scale format such as, “How satisfied are you with your dentist: very satisfied, somewhat satisfied, somewhat dissatisfied, or not satisfied at all?” A closed-ended format is much faster and easier for both respondents and the researcher. It permits large-scale samples and long questionnaires. Yet, you can lose important insights by forcing an individual’s complex beliefs and feelings into a few fixed categories. One way to reduce the disadvantages of a closed-ended format is to mix open-ended and closed-ended formats in a questionnaire or to use a partially

open format. Mixing formats also changes the pace and helps interviewers establish rapport. Periodic probes with closedended answers can reveal a respondent’s reasoning. For large-scale surveys, you can use open-ended format in small, preliminary pilot tests, then develop closed-ended response categories based on responses the open questions. The partially open format uses a closed-ended format with an openended “other” option. For example, the following is a survey question used in CBS/Newsweek/New York Times polls that I modified to make partially open: Which of the following comes closest to your view? _____ Gay couples should be allowed to marry, OR _____ form civil unions, OR _____ there should be no legal recognition of gay couple’s relationship, OR some other arrangement, please specify ________________

Summary Review

Open- and Closed-Ended Response Formats Table 6.4 Closed-Ended Response Formats Advantages of Closed-Ended Format

Disadvantages of Closed-Ended Format

• • • •

• Questions can suggest ideas that the respondent would not otherwise have. • Respondents without an opinion or knowledge on an issue may answer anyway, creating meaningless data. • Respondents can be frustrated because their desired answer is not a choice. • Respondents get confused if many (e.g., 15) response choices are offered. • A respondent’s misinterpretation of a question can go unnoticed. • Distinctions between respondent answers to a question can become blurred. • Respondent mistakes in marking the wrong response are possible. • Respondents are forced to give simplistic responses to complex issues. • Respondents are forced to make choices they would not make in the real world.

• • • •

Easier and quicker for respondents to answer. Easier to compare answers across respondents. Easier to code and statistically analyze answers. The response choices can help clarify question meaning for respondents. Respondents are more likely to answer about sensitive topics. There are fewer irrelevant or confused answers to questions. Less articulate or less literate respondents are not at a disadvantage. Replication of the survey is easier.

Table 6.5 Open-Ended Response Formats Advantages of Open-Ended Format

Disadvantages of Open-Ended Format

• They permit an unlimited number of possible answers. • Respondents can answer in detail and can qualify and clarify their responses. • Researchers can discover unanticipated findings. • Respondents can give adequate answers to complex issues. • Respondents can be creative, express themselves, and provide rich detail. • Respondents can reveal their thinking process and frame of reference.

• • • • •

Different respondents give different degrees of detail in their answers. Respondent answers may be irrelevant or buried in useless detail. Data coding, comparisons, and statistical analysis become very difficult. Articulate and highly literate respondents have an advantage. Questions may be too general for respondents who lose direction or the questions may intimidate unsophisticated respondents. • A greater amount of respondent time, thought, and effort is necessary. • Answers require a lot of space and time to record.

How Do You Write Good Closed-Format Responses? Most surveys offer preset answers from

which a respondent chooses. Writing good answer choices is just as important as writing a good question. The answer choices should have three features: • Mutually exclusive. Response categories do not overlap. You can easily correct numerical ranges (e.g., 5–10, 10–20, 20–30) that overlap (e.g., 5–9, 10–19,

20–29). The ambiguous verbal choice is another type of overlapping response category—e.g., “Are you satisfied with your job, or are there things you don’t like about it?” It is not clear how a person who is generally satisfied but has a few minor complaints would answer this. • Exhaustive. This means that each respondent has a choice—a place to go. For example, asking respondents

126 Chapter 6 “Are you working or unemployed?” leaves out respondents who are not working but do not consider themselves unemployed (e.g., full-time homemakers, people on vacation, students, people with disabilities, or retired people). When writing a question, first think about what you want to know and then consider the circ*mstances of all possible respondents. For example, if you ask about a respondent’s employment, do you want information on the primary job or on all jobs? Do you want both fulland part-time work? Do you only want jobs for pay or also unpaid and volunteer jobs as well? If someone is temporarily unemployed, do you want the last job he or she held? • Balanced. This means you offer the favorable or unfavorable choices equally in a set of responses. A case of unbalanced choices is the question, “What kind of job is the mayor doing: outstanding, excellent, very good, or satisfactory?” It offers three favorable and one neutral response choice. Another type of unbalanced question omits information—e.g., “Which of the five candidates running for mayor do you favor: Eugene Oswego or one of the others?” You can balance responses by offering bipolar opposites. Unless there is a specific purpose for doing otherwise, offer respondents equal polar opposites at each end of a continuum. Asking, “How strongly you support a ban on same-sex marriage? Do you strongly support it, somewhat support it, or just barely support it?” is a set of unbalanced answers. To make it balanced, you could ask, “How do you feel about a ban on same-sex marriage; do you support it, oppose it, or neither support or oppose it?” Should You Offer a “Don’t Know” or “No Opinion” Response Choice? Professional survey

researchers debate whether to include choices for neutral, middle, and nonattitudes (e.g., “not sure,” “don’t know,” “undecided,” or “no opinion”) in closed-ended questions. They want to avoid two errors: • Getting a “no opinion” or “don’t know” response when a respondent actually holds a nonneutral opinion • Forcing a respondent to choose a position when he or she has no opinion on an issue or knows nothing about it. We have three ways to address the “don’t know” response in attitude questions: 1. standard-format 2. quasi-filter 3. full-filter questions. A standard-format question does not offer a “don’t know” choice, and respondents must volunteer their lack of knowledge or opinion. A quasi-filter question offers

respondents a “don’t know” or “not certain” answer alternative. A full-filter question is a special type of contingency question (contingency questions are discussed later). It is a two-part question. It first asks whether respondents have an opinion, and then asks for the opinion among those who say that they have an opinion. Studies of survey formats vary, but several suggest that without a “no opinion” choice, as in the standard question, some respondents will offer to answer a question, even if they are very uncertain or unaware of an issue. Many people find saying “I don’t know” or “I have no opinion” difficult to assert or may feel embarrassed in doing so. With a quasi-filter question, most such respondents choose “don’t know” because it appears as a legitimate response. A full-filtered question takes “no opinion” or “don’t know” options one step higher. You should use a full filter for issues about which many people may not be informed or have a firm opinion. An option is to ask about an opinion using a quasi-filter question, then follow up all those with an opinion with a second question about how strongly they feel. Example What is your opinion about the issue of global warming? Do you feel in the future it is going to be a major threat, a minor threat, or no real threat to how our society will be, or do you have no opinion? _____ _____ _____ _____

Major Threat Minor Threat No Real Threat No opinion {Go directly to next question}

[ASK ONLY IF FIRST THREE ANSWERS ARE GIVEN] How strongly do you hold that opinion? Do you hold the opinion _____ very strongly, _____ somewhat strongly, or _____ not very strongly at all?

Example Study Questionnaire Items from the2009 Pew Research CenterSurvey This marital status question has mutually exclusive and exhaustive answer choices. Are you currently married, living with a partner, divorced, separated, widowed, or have you never been married? (IF R SAYS “SINGLE,” PROBE TO DETERMINE WHICH CATEGORY IS APPROPRIATE) {QID:MARITAL1} 1. Married 2. Living with a partner 3. Divorced 4. Separated

The Survey 127 5. Widowed

The response entered here will appear in the performance dashboard and can be viewed by your instructor.

6. Never been married 9. Don’t know/Refused (VOL.) The question below reduces the social desirability bias (discussed later in this chapter) to say that a person registered by offering the out of “can’t find the time to register.” REGIST These days, many people are so busy they can’t find time to register to vote, or move around so often they don’t get a chance to re-register. Are you NOW registered to vote in your precinct or election district or haven’t you been able to register so far? [INSTRUCTION: IF RESPONDENT VOLUNTEERS THAT THEY ARE IN NORTH DAKOTA AND DON’T HAVE TO REGISTER, PUNCH 1 FOR REGIST AND REGICERT] {QID:REGIST} 1. Yes, registered 2. No, not registered 9. Don’t know/Refused (VOL.)

Making It Practical: Standard-Format, QuasiFilter, and Full-Filter Questions

Standard Format Here is a question about another country. Do you agree or disagree with this statement? “The Russian leaders are basically trying to get along with America.” Quasi-Filter Here is a statement about another country: “The Russian leaders are basically trying to get along with America.” Do you agree, disagree, or have no opinion on that? Full Filter Here is a statement about another country. Not everyone has an opinion on this. If you do not have an opinion, just say so. Here’s the statement: “The Russian leaders are basically trying to get along with America.” Do you have an opinion on that? If yes, do you agree or disagree? Source: Schuman and Presser. 1981:116–125.

Table 6.6 Example of Results from Different QuestionForms

Submit How Can You Help Respondents Remember?

Many survey questions ask about past events, such as when you last saw a doctor or when you last purchased a cell phone. Recalling events accurately takes more time and effort than the few seconds respondents typically use when answering. Also, accurate answering declines over time. Most respondents can recall highly significant events that occurred in the past four to six weeks, but after that, their recall accuracy erodes. The following might interfere with good recall: • A sensitive or threatening topic—people often suppress a bad memory and forget unpleasant or embarrassing events. • Events that occurred simultaneously—when several things occur at once, they can blur together in memory. • Events that occurred after that being recalled—more recent events can distort the memory of what happened earlier. • An issue or event that was not significant—people remember what was important for them but if they did not consider it important, they are likely to forget it. • A person’s need to be consistent and not appear to contradict him/herself—people tend to selectively remember what is consistent and forget what is contradictory. Such influences on recall do not mean that you cannot ask about past events, but when asking about the past you need to customize questions and interpret results cautiously. It is best to give respondents special instructions, extra thinking time, and recall aids. One recall aid is to use a fixed time frame or location. Rather than ask: “How often did you attend sporting events last year?”

Standard Form (%)

Quasi-Filter (%)

Full Filter (%)

Agree

48.2

27.7

22.9

Disagree

38.2

29.5

20.9

No opinion

13.6*

42.8

56.3

* Volunteered Source: Adapted from Schuman and Presser (1981:116–125). Standard format is from Fall 1978; quasi- and full-filter are from February 1977.

WRITING PROMPT What If You Have No Opinion? If you were presented with a survey question in which there was no “Don’t Know” or “No Opinion” choice but you did not know or had no opinion, how would it make you feel? What could you do in this situation?

Instead, ask it this way: “I want to know how many sporting events you attended last winter. Let’s go month by month. Think back to December. Did you attend any sporting events for which you paid an admission in December? Now, think back to January. Did you attend any sporting events in January?”

If you want information about activities over a long period, it is best to ask about a short time frame that a respondent is likely to know about, and then do the math yourself afterward. Instead of asking: “How many hours did you watch TV in the past year?”

128 Chapter 6 Try asking: “In a typical weekday during the past year, about how long did you watch TV? What about in a typical weekend day, how often?”

You can then multiply the respondent’s answers by the number of weekdays and weekend days to create an estimate for annual TV watching. How Can You Ask Respondents about Sensitive Issues? Most respondents want to present a posi-

tive image of themselves. They may feel ashamed, embarrassed, or afraid to give truthful answers regarding unpleasant or unflattering behaviors or events. They may find it emotionally painful to confront their own actions honestly, let alone admit them to a stranger. People may underreport or self-censor reports of behavior or attitudes they wish to hide or believe to violate social norms such as having an illness or disability (e.g., cancer, mental illness, venereal disease) or engaging in illegal or deviant behavior (e.g., evading taxes, taking illegal drugs, engaging in uncommon sexual practices). They may be hesitant to reveal their financial status (e.g., income, savings, or debts). Alternatively, they may over report positive or generally accepted behaviors. Researchers developed several techniques to increase getting truthful answers to sensitive issues. One is to alter the context and question wording to be less threatening. You should only ask about sensitive issues after a warmup, when respondents feel more trust in the survey or interviewer. You can emphasize to respondents that you want honest answers and reassure them of confidentiality. You can provide a context that makes it easier for respondents to answer and appear less unusual. For example, rather than asking “Have you stolen from a store?,” you can ask, “In past surveys, many people have reported that at some point they took items from a store without paying. Have you ever taken something from a store without paying for it?” Another technique is first to ask about more serious activities, making the sensitive question issue appear less unusual. A respondent may hesitate to admit that he or she shoplifted. However, a question about shoplifting appears after several questions about armed robbery or burglary, respondents may admit to shoplifting because it appears to be less serious than the other crimes. Social desirability bias occurs when respondents distort answers to look good or to conform to social norms. Many people over report being highly cultured (i.e., reading books, attending high-culture events), giving money to charity, having a good marriage, loving their children, and so forth. For example, one study found that one-third of people who said they gave money to a local charity in a survey really did not. Because a norm says that one should vote in elections, people often say they voted when they

did not. To reduce social desirability bias, you can phrase questions to make norm violation appear less objectionable, or present a wider range of behavior as acceptable or give respondents “face-saving” alternatives. The National Election Survey asked about voting in the following way to reduce the social desirability bias: “In talking to people about elections, we often find that a lot of people were not able to vote because they weren’t registered, they were sick, or they just didn’t have time. Which of the following best describes you?—One, I did not vote. Two, I thought about voting this time but didn’t. Three—I usually vote, but didn’t this time. Four—I am sure I voted.” What Are Contingency Questions? A contingency question (also called screen or skip question) is a two-question sequence that increases relevance. A first question selects respondents for whom the second question is relevant. It screens in/out respondents who get the second part. The following example is a contingency question. 1. Did you vote in the mayoral election last April when Guo, Smith, and Lopez were candidates? [ ] Yes (GO TO QUESTION 2) [ ] No (SKIP TO QUESTION 3) 2. Which candidate did you vote for? _____ Guo _____ Smith _____Lopez _____ Don’t remember 3. What kind of overall job is the new mayor doing in your opinion? _____ Excellent _____ Good _____ Fair _____ Poor How Can You Avoid Specific Words That Affect Answers? Wording effects occur when a par-

ticular word evokes a response. Professional survey researchers recognize that particular words in a survey questions may trigger strong feelings or have connotations that color answers. Because respondents react to one word rather than thinking about the issue in a question, you want to avoid such words in survey questions. It is easier to write survey questions if you have a large vocabulary, know the connotations and meanings of many words, and are sensitive to the vocabulary of respondents. In general, you want to use simple vocabulary and grammar to minimize confusion. Unfortunately, it is not possible to know in advance whether a word or phrase will affect responses.

Learning from History The Power of Words Survey researchers have uncovered several powerful wording effects in surveys. One well-documented effect is the difference between forbid and not allow.1 Both terms mean the

See Foddy (1993) and Presser (1990).

The Survey 129 same thing, but many more people are willing to “not allow” something than to “forbid” it. In general, less educated respondents are most influenced by minor wording differences. Certain words trigger an emotional reaction or have significant connotations that we are just beginning to learn about. Smith (1987) found large differences (e.g., twice as much support) in U.S. survey responses depending on whether a question asked about spending “to help the poor” or “for welfare.” Heated political attacks on welfare in the 1970s and 1980s changed connotations of the word welfare, and it took on negative connotations that it did not previously have. The once neutral word came to imply lazy and immoral people as well as wasteful, ineffective, and expensive government programs. Today, it is best to avoid using it. Likewise, Hurwitz and Peffley (2005) discovered that in recent years many Americans have come to associate the term inner city with negative racial stereotypes about African Americans. Racially prejudiced whites gave negative responses when the phrase “inner city” appeared in a survey question but neutral responses for the same issues when it did not appear. In a 2005 Pew Research survey, 51 percent of respondents said they favored “making it legal for doctors to give terminally ill patients the means to end their lives” but only 44 percent favored “making it legal for doctors to assist terminally ill patients in committing suicide.” Both questions asked about the same thing, but the respondent reactions differed because of the word “suicide.” Respondents also can become confused about the meaning or connotations of key words. One survey asked respondents whether they thought television news was “impartial.” Impartial is a ninth-grade vocabulary term, and researchers assumed everyone knew its meaning. They later learned that fewer than half of the respondents had interpreted the word with its proper meaning. Over one-fourth had ignored it or had no idea of its meaning. Others gave it unusual meanings, and one-tenth gave it a meaning directly opposite to its true meaning. You need to be cautious, because although a few wording effects (e.g., the difference between forbid and not allow) are known, we are still learning about the power of specific words to shape respondent answers.

6.4: How Can You Design an Effective Questionnaire? 6.4 Analyze some of the features of an effective questionnaire Once you have created a collection of survey questions, you face two decisions in designing an effective questionnaire. First, how many questions can you include in the questionnaire, or total questionnaire length. Second, how should you organize or arrange the questions that are in the questionnaire.

6.4.1: Length of Survey orQuestionnaire The length of a questionnaire depends on the format of your survey and on respondent characteristics. A 3–5 minute telephone interview is rarely a problem. You can often extend it to 10 minutes. Web surveys vary but few people spend more than 10 minutes taking them. Mail questionnaires are more variable. A short (three-page) questionnaire is appropriate for the general population. Some researchers used questionnaires as long as 10 pages (about 100 items) with the general public, but responses drop significantly for longer questionnaires. For highly educated respondents and a salient topic, using questionnaires of 15 pages may be possible. Many face-to-face interviews last a half-hour. In special situations, researchers have been able to conduct face-to-face interviews for as long as 3–5 hours.

6.4.2: Question Sequence You face three survey question sequence issues when designing a questionnaire: 1. How to organize questions on a questionnaire 2. How to reduce question order effects 3. How to control context effects

WRITING PROMPT Wording Effects in Survey Questions Sometimes a term with a strong emotional impact appears in a survey question and people answer based on their emotions about the term rather than the word’s actual meaning. Brainstorm a short list ofwords that you should avoid using in a survey question because they have a lot of emotional impact. The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

Several years ago, my students conducted a telephone survey on two topics: concern about crime and attitudes toward a new anti-drunk-driving law. A random half of the respondents heard questions about the drunk-driving law first; the other half heard about crime first. I examined the results to see whether there was a context effect—a difference by topic order. This study found that respondents asked about the drunk-driving law first expressed less fear about crime than did those asked about crime first. Likewise, those asked first about drunk driving expressed more support for a tough new drunk-driving law than those who first heard about crime. Apparently, the first topic created a context influencing answers to the second topic. After hearing questions about crime in general and about violent

130 Chapter 6 crime, respondents may have considered drunk driving to be a less serious offense. By contrast, respondents asked about drunk driving first may have thought about it as criminal behavior. When asked about crime in general, they still are thinking of drunk driving as a type of crime. How to Organize Questions on a Questionnaire Every questionnaire has opening, middle, and

ending questions. Sequence the questions in a way to minimize respondent discomfort and confusion. The early questions should help a respondent feel positive and comfortable with the survey process. After an introduction that explains the purpose of the survey, make opening questions pleasant, interesting, and easy to answer. Demographic questions (e.g., age, education level, and so forth) are easy but not interesting, and should go toward the end. In addition, place questions on the same topic together and introduce the section with a short statement (e.g., “Now I would like to ask you questions about housing”) to orient respondents. Question topics should flow smoothly and logically. Organize them to assist respondents’ memory and comfort levels. How to Reduce Question Order Effects The

order in which questions appear may influence a respondent’s answers. Order effects are strongest for people who lack strong views, for less educated respondents, and for older respondents or those with memory loss. For example, support for a single woman’s having an abortion predictably rises if it comes after a question about abortion being acceptable when a fetus has serious defects, but not when the question is by itself or it comes before a question about fetus defects. Making It Practical: The Effect of Previous Questions Previous questions in a questionnaire influ-

ence later ones in two ways: 1. Question content (i.e., the issue). This occurred in the above example of my student’s study about drunk driving and crime. In another case, researchers compared three forms to ask how much a respondent followed politics: the question alone, after asking what the respondent’s elected representative recently did, and after asking about what the representative did and about the representative’s “public relations work” in the area. The percentage of respondents’ reporting that they followed politics “now and then” or “hardly at all” was 21 percent, 39 percent, and 29 percent, respectively, for the three forms. The second form apparently made respondent’s feel they knew little, but the last form gave respondents an excuse for not knowing the first question—they could blame their representative for their ignorance. 2. Respondent’s response. A respondent having already answered one part of an issue may assume no overlap. For example, a respondent is asked, “How are you doing in the classes in your academic major?” The next

question is, “How are you doing in your classes?” Most respondents will assume that the second question only means classes outside their major because they already answered about the major. If you wanted to ask about classes overall, then you should place that question before the question about the classes in the major.

WRITING PROMPT Survey Question Order In many surveys, we ask about multiple topics. Can you think of two topics for a survey where answering one topic first might shift how people think about a second topic? What might we do to find out whether topic order makes a difference, and if so how to deal with the issue? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

How to Control Context Effects Respondents tend to answer questions based on a context of preceding questions and the interview setting. Context effects are strongest for ambiguous or unclear questions. This is because respondents draw on the context when interpreting a question. It is not always possible to control context effects. A first step is to be aware of them. You want to ask a more general question before a more specific one. It takes a bit more work, but making two versions of the questionnaire and altering topics to two random parts of a sample allows you to check whether context effects are operating.

Example Study Question Order Effects In a study with 9,000 respondents, Wilson (2010) found that the ordering of survey questions influenced reports of interracial prejudice expressed by American blacks and whites. When questions specifically deal with salient emotional cues, such as interracial hostility among blacks and whites, question context and racial group membership strengthen in-group and out-group categorization. In-group members (e.g., blacks) tend to view members of an out-group (e.g., whites) as having greater dislike toward them (e.g., whites dislike blacks) when respondents were first asked about the in-group (e.g., blacks). This produced a desire to favor one’s in-group over the outgroup—a contrast effect. When the opposite order occurred, a norm of “evenhandedness” suggests that evaluating an outgroup in a noncomparative context, the in-group member feels a need to appear balanced or evenhanded, particularly in the presence of an interviewer. When first asked if their own group dislikes the out-group, respondents felt compelled to say that more of the out-group members disliked them in a follow-up

The Survey 131 question. So, when black respondents were asked “Do whites dislike blacks?” 41% said “yes.” When white respondents were asked “Do blacks dislike whites?” 42.8% said “yes.” However, when asked about their own in-group first, perceptions shifted. When black respondents were asked the same question after they first answered whether “blacks dislike whites,” then 55.4% said “yes” instead of 41%. When white respondents were asked “Do blacks dislike whites?” after having answered whether whites dislike blacks, 51% said “yes” instead of 42.8%. In short, roughly 10 percent more people said the out-group disliked their in-group in one situation over the other. It all depended upon the context, or which question they answered first. (See Figure 6.4.)

Figure 6.4 Race of Respondent by Question Order Interaction Pattern (Predicted Probability of Perceiving More Dislike toward Out-Group Members) Answering Survey Questions Vary by Question Order and Respondent’s Race 60 55.4

Percentage

51 50 45 41

42.8

40 35

Black Respondents

White Respondents

Whites dislike Blacks first asked Blacks dislike Whites first asked

6.5: The Advantages and Disadvantages of Different Survey Formats 6.5 List the advantages and disadvantages of various survey formats

Each survey format has advantages or disadvantages: selfadministered, mail, face-to-face-interview, phone interview, and web survey.

6.5.1: Mail and Self-Administered Questionnaires Advantages Mail and self-administered question-

naires are popular because they are easy and inexpensive. You distribute or mail questionnaires directly to respondents, respondents read instructions and questions, and then record their answers. If you use mail, you can cover a wide geographical area. In addition, a mail survey allows respondents to check personal records at home if necessary, offers anonymity, and avoids interviewer bias. Disadvantages Physical location is limited for distributed, self-administrated questionnaires. Distribution by mail provides large geographic area, but people do not always complete and return questionnaires and the biggest problem with mail questionnaires is a low response rate. You might mail out 500 questionnaires but get back only 50. Increasing the number mailed out to 50,000 so that you have 500 both becomes much more expensive and can create a bias. The 10 percent of people who respond are unlikely to be representative. Perhaps only people who are highly interested in the survey topic or who have a lot of free time (e.g., unemployed, retired, traditional homemakers) respond. The opinions, education, income, age, and other characteristics of those who respond may not adequately reflect the entire sample. It might seriously distort your data. Another limitation is that you do not control the conditions under which a person completes a mail questionnaire. A questionnaire completed during a drinking party by a dozen laughing people may be returned along with one filled out by an earnest respondent. With mail questionnaires, no one is present to clarify questions or to probe for more information when respondents give incomplete answers. Someone other than the sampled respondent (e.g., spouse, new resident, etc.) may open the mail and complete the questionnaire. Respondents can complete the questionnaire weeks apart or answer questions in an order different from that intended by researchers. Incomplete questionnaires can also be a serious problem. A mail questionnaire format limits the kinds of questions that a researcher can use. Questions that require visual aids (e.g., look at this picture and tell me what you see), open-ended questions, many contingency questions, and complex questions do poorly in mail questionnaires. Likewise, mail questionnaires are ill suited for the illiterate or semi-illiterate in a country’s main language.

132 Chapter 6

6.5.2: Telephone Interviews Advantages The telephone interview is a popular

survey method because you can reach about 95 percent of the population by telephone. You call a respondent, ask questions, and record answers. The sample of respondents can come from lists, telephone directories, or random-digit dialing (RDD) and can be from a wide geographical area. A staff of interviewers can interview 1,500 respondents across a nation within a few days and, get response rates as high as 85 percent with 7 to 10 callbacks. The telephone interview is a flexible method and has most of the strengths of face-to-face interviews. Interviewers pick a specific respondent, control the sequence of questions, and can use some probes. A specific respondent answers the questions alone. Interviewers can use contingency questions effectively, especially with computer-assisted telephone interviewing (CATI) (discussed later in this chapter). Also, supervisors can monitor interview quality by listening in. Disadvantages High cost and limited interview

length are major disadvantages of telephone interviews. Respondents without telephones are impossible to reach, or the call may come at an inconvenient time. The use of an interviewer reduces anonymity and introduces potential interviewer bias. Open-ended questions are difficult to use, and questions that require visual aids are impossible. Interviewers can only note serious disruptions (e.g., background noise) and respondent tone of voice (e.g., anger or flippancy) or hesitancy. Over the past 30 years, respondent cooperation with telephone interviews has steadily declined, with more refusals and inability to reach people, even with 10 callbacks at different times.

6.5.3: Face-to-Face Interviews Advantages Face-to-face interviews have the highest response rates and permit the longest questionnaires. Interviewers also can observe the surroundings and can use nonverbal communication and visual aids. Welltrained interviewers can ask all types of questions, can ask complex questions, and can use extensive probes. Disadvantages The training, travel, supervision, and

personnel costs for interviews can be very high. Interviewer bias is also greatest in face-to-face interviews. The appearance, tone of voice, question wording, and so forth of the interviewer may affect the respondent. In addition, interviewer supervision is less than for telephone interviews.

6.5.4: Web Surveys Advantages Web surveys came into widespread use

over the past 10 years. They are now the lowest cost format and get answers the fastest. Three other advantages are that web surveys can span a large geographic area, include

visual materials as well as the survey questions, and use elaborate contingency questions. Disadvantages A drawback with web surveys is that not everyone has Internet access, especially low-income, elderly, and less educated people. Low response rates and incomplete surveys are also common problems with web surveys. As incentives for web survey participation improve, these issues may become less important. Another disadvantage is a lack of control over the conditions under which a person completes a web survey. Someone who is not serious or someone other than the selected respondent may complete the web survey. Also, as with mailed questionnaires, no one is present to clarify questions or to probe for more information when respondents give incomplete answers.

WRITING PROMPT Different Survey Formats Of the various types of survey formats, what type of survey format orpresentation encourages you to take the survey most seriously? Describe what qualities in that type of survey encourage more respondents to take a survey questionnaire seriously and give the most truthful, thoughtful answers. The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

6.6: How Do You Interview in Survey Research? 6.6 Apply best practices for preparing the interviewer to administer social surveys Interviews are a special type of social relationship in which one person asks the questions while another person provides answers. There are several types of interview. There are police interrogation interviews, job candidate selection interviews, and celebrity entertainment interviews. In survey research interviews, the primary goal is to obtain accurate information from another person. The survey interview is a short-term social interaction between two strangers with expectations, social roles, and norms. Its explicit purpose is for one person to obtain specific information from the other. The roles are those of the interviewer and the interviewee or respondent. The interviewer obtains information by asking prearranged questions and recording responses given by the respondent. You have several choices if you plan to collect survey data with interviewers. Will you interview by phone or faceto-face? Will you personally do the interviews or have others conduct interviews for you?

The Survey 133

If you use other people to conduct interview, you need to train them with the questionnaire. Interviewers must become very familiar with the wording and purpose of each question and practice following the flow of questionnaire items. They will also need to exhibit proper survey interview behavior.

6.6.1: The Interviewer’s Role The interviewer’s task is complex. He or she must control the interview and the flow of interaction, obtain cooperation and build rapport, and simultaneously remain neutral and objective. Interviewers encroach on the respondents’ time and privacy for information that may not directly benefit the respondents. They try to reduce embarrassment, fear, and suspicion so that a respondent feels comfortable revealing information. A good interviewer constantly monitors the pace and direction of the social interaction as well as the answers and the behavior of respondents. He or she helps respondents to feel that they should give complete, truthful answers. It is important to note that not all participants will be familiar with their role as a survey respondent. A survey interviewer may need to clarify the roles in the social relationship and follow them consistently or explain the nature of survey research. Some respondents will view the research interview as an intimate conversation similar to a therapy session, a boring bureaucratic exercise in completing forms, a referendum on public policy choices, or a competitive testing situation. Some suspicious people even treat it as a contest in which interviewers try to trick or entrap respondents. A misunderstanding about the respondent role can produce a misunderstanding of the meaning of survey questions. A survey interviewer is nonjudgmental and never reveals his or her opinions, verbally or nonverbally (e.g., by a look of shock). If a respondent asks for an interviewer’s opinion, the interviewer politely redirects the respondent. For example, if a respondent asks “What do you think?,” the interviewer may answer “Here we are interested in what you think; what I think doesn’t matter.” Likewise, if the respondent gives a shocking answer (e.g., “I was arrested three times for beating my infant daughter and burning her with cigarettes”), an interviewer should not show shock, surprise, or disdain but treat the answer in a matter-of-fact manner. If the survey interviewer must be neutral and objective, why not use a computer? Compare Your Thoughts Machine interviewing has many limitations. It lacks the human warmth, sense of trust, motivation, and rapport that a human interviewer creates. It cannot answer respondent questions

or clarify misinterpretations. An interviewer helps to define the situation for respondents, offers guidance and determines whether respondents have the information sought, understands what is expected, and is providing relevant and serious answers.

6.6.2: The Interview Stages A survey interview proceeds through three major stages: introduction and entry, main part of the interview, and the exit. Stage 1: Introduction and Entry The interviewer gains access, provides authorization, and reassures and secures cooperation from the respondent. He or she is prepared for reactions, such as “How did you pick me?” “What good will this do?” “I don’t know about this,” “What’s this about?” The interviewer explains why the specific respondent is being interviewed and not a substitute. In this stage, the interviewer takes control, and establishes role expectations. Informed consent and assurances of confidentiality occur in this stage. Stage 2: The Main Part of the Interview The main part consists of asking questions and recording answers. The interviewer follows the exact wording on the questionnaire—no added or omitted words and no rephrasing. In this stage, it is important to set a comfortable pace and give nondirective feedback to maintain respondent interest. Nondirective feedback might include a slight nod or smile, making eye contact, or quietly saying “OK.”

He or she asks all applicable questions in order, without returning to or skipping questions unless the directions specify this. Interviewers need to be aware that respondents often reinterpret survey questions to apply to their personal situations or to make them easy to answer. In addition to asking questions, the interviewer accurately records all answers. This is relatively easy for closedended questions, where interviewers just mark the correct box. For open-ended questions, the interviewer’s job is more difficult. An interviewer listens carefully, has legible writing or good recording machine skills, and records what a respondent says verbatim without correcting grammar or slang. The interviewer does not summarize or paraphrase because doing so leads to lost information or distorted answers. For example, the respondent says, “I’m really concerned about my daughter’s heart problem. She is only 10 years old and already she has trouble climbing stairs. I don’t know what she’ll do when she gets older. Heart surgery is too risky for her and it costs so much. I guess she’ll have to learn to live with it.” If the interviewer summarizes, “respondent is concerned about daughter’s health,” much is lost.

134 Chapter 6 Stage 3: The Exit In this last stage, the interviewer

thanks the respondent and leaves. He or she then goes to a quiet, private place to edit the questionnaire. During this time, other details like the date, time, and place of the interview; a thumbnail sketch of the respondent and interview situation; the respondent’s attitude (e.g., serious, angry, or flippant); and any unusual circ*mstances (e.g., “Telephone rang at question 27 and respondent talked for four minutes before the interview started again”) should be recorded. The interviewer will also note anything disruptive that happened during the interview (e.g., “Teenage son entered room, sat at opposite end, turned on television with the volume loud, and watched a music video”). The interviewer also records his or her personal feelings and anything that was suspected (e.g., “Respondent became nervous and fidgeted, changed answers once each on questions 14, 15, and 16 about his marriage”).

6.6.3: Training Interviewers Perhaps someday you may get a job interviewing for a professional survey organization. A large-scale survey employs many interviewers. Few people other than professional survey researchers appreciate the difficulty of the interviewer’s job. A professional-quality interview requires a carefully selected and trained interviewer. As with any employment situation, interviewers need adequate pay and good supervision to perform consistently at their peak. Good interviewers are pleasant, honest, accurate, mature, responsible, intelligent, stable, and motivated. They are patient and calm. They have experience with many types of people and possess poise and tact. Face-to-face interviewers must have a nonthreatening appearance. If the survey involves face-to-face interviewing in high-crime areas, the interviewers need to have proper “street smarts” and may require extra protection (e.g., a partner or assistant).

interviews in the office and in the field that are recorded and critiqued, many practice interviews, and role playing. Interviewers are taught about survey research and the role of the interviewer. They must become very familiar with the questionnaire and the purpose of questions.

6.6.4: Using Probes in Interviews A good interviewer knows how and when to use probes. Probes clarify a respondent’s ambiguous or irrelevant response and confirm that respondents understand the questions as intended. This means interviewers need to understand the survey and recognize an irrelevant or inaccurate answer. There are many types of probes. A three- to five-second pause is often effective, as is nonverbal communication (e.g., tilt of head, raised eyebrows, or eye contact). The interviewer can repeat the question or the reply and then pause. He or she can ask a neutral question, such as: “Any other reasons?” “Can you tell me more about that?” “How do you mean?” “Could you explain more for me?” Making It Practical: Using Probes

Interviewer Question: What is your occupation? Respondent Answer: I work at General Motors. Probe: What is your job at General Motors? What type of work do you do there? Interviewer Question: How long have you been unemployed? Respondent Answer: A long time. Probe: Could you tell me more specifically when your current period of unemployment began? Interviewer Question: Considering the country as a whole, do you think we will have good times during the next year, or bad times, or what? Respondent Answer: Maybe good, maybe bad, it depends, who knows? Probe: What do you expect to happen? Record Response to a Closed Question

Interviewer Question: On a scale of 1 to 7, how do you feel about capital punishment or the death penalty, where 1 is strongly in favor of the death penalty, and 7 is strongly opposed to it? When hiring interviewers, researchers consider the candidate’s physical appearance, age, race, sex, languages spoken, and even the sound of their voice. Most training for professional interviewers requires full-time sessions lasting one to two weeks. It usually includes lectures and reading, observation of expert interviewers, mock

(Favor) 1 ___ 2 ___ 3 ___ 4 ___ 5 ___ 6 ___ 7 ___ (Oppose) Respondent Answer: About a 4 or 5. I think that all murderers, rapists, and violent criminals should get death, but I don’t favor it for minor crimes like stealing a car. Interviewer: Would that be a 4 or 5 on this rating from 1 to 7?

The Survey 135

Probes are not substitutes for writing clear questions or creating a framework of understanding for the respondent. Unless carefully stated, probes might shape the respondent’s answers. Yet flexible or conversational interviewing, in which interviewers use many probes, can improve accuracy on questions about complex issues for which respondents do not clearly understand basic terms or about which they have difficulty expressing their thoughts. For example, to the question, “Did you do any work for money last week?,” a respondent might hesitate and then reply, “Yes.” An interviewer can probe, “Could you tell me exactly what work you did?” The respondent may reply, “On Tuesday and Wednesday, I spent each afternoon helping my buddy John move into his new apartment. For that he gave me $50, but I didn’t have any other job or get paid for doing anything else.” If the question intended only to get reports of regular paid employment, the probe revealed a misunderstanding.

WRITING PROMPT Probes in Interviews Does using many probes impede the goal of standardizations (i.e., each respondent hears the exact same survey question, so we knowall answers from respondents are to the same question)? Explain your answer. The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

6.6.5: Interviewer Bias Survey researchers prescribe interviewer behavior to reduce bias. Bias is when a particular interviewer ’s actions have influenced how a respondent answers, and responses differ from what they would be if another interviewer had asked the questions. Proper interviewer behavior and exact question reading are critical, but there is a larger issue. An interviewer’s uncontrolled visible characteristics, including race, stature, age and gender, often affect interviews and respondent answers. This means noting the characteristics of both interviewers and respondents, especially in questions about issues related to visible characteristics. For example, African American and Hispanic American respondents tend to express different policy positions on race- or ethnic-related issues depending on the apparent race or ethnicity of the interviewer. This occurs even with telephone interviews when a respondent has clues about the interviewer’s race or ethnicity.

Gender also affects interviews in terms of both obvious issues, such as sexual behavior, and support for gender-related collective action or gender equality. In general, interviewers of the same gender or ethnic-racial group as the respondent tend to get the most accurate answers.

Example Study Interviewer Race Effects AreSubtle and Pervasive Survey researchers have been long aware that an interviewer’s race can influence how respondents answer racially sensitive questions. Other research documented that women or racial minorities tend do poorly on certain tests administered by members of outside groups (such as white males) because of a stereotype that they will do poorly. They feel great pressure to disconfirm the negative stereotype, but the pressure creates test anxiety that lowers their test score. In contrast, when a member of their same group administers the test, the stereotype threat is not activated, test anxiety is reduced, and they score higher. To test whether the stereotype-triggered test anxiety also operates in survey interviews, Davis and Silver (2003) conducted a telephone survey of whites and African Americans in the Detroit area. They wanted to see whether a survey interviewer’s race affected how a respondent answered. The research question was, “Do African Americans score differently on survey knowledge questions depending on whether they think the telephone interviewer is white or African American?” Results showed that white respondents answered the same irrespective of interviewer’s race. However, African American respondents scored higher on the knowledge questions when they believed their interviewer was an African American. The authors concluded that beyond widely known social desirability and racial conformity issues in surveys, the race of an interviewer can create subtle anxiety based on negative stereotypes that influence how a respondent answers many survey questions.

6.6.6: Computer-Assisted Telephone Interviewing

Most professional survey organizations that do phone interviewing have installed computer-assisted telephone interviewing (CATI) systems. With CATI, the interviewer sits in front of a computer and makes calls. Wearing a

136 Chapter 6 headset and microphone, the interviewer reads the questions from a computer screen for the specific respondent, then enters the answer via the keyboard. Once he or she enters an answer, the computer shows the next question on the screen. CATI speeds interviewing and reduces interviewer errors. It also eliminates the separate step of entering information into a computer and speeds up data processing. The CATI system works well for contingency questions because the computer shows the questions appropriate for a specific respondent; interviewers do not have to look for the next appropriate question. In addition, the computer can check an answer immediately after the interviewer enters it. For example, the interviewer enters an answer that is impossible or clearly an error (e.g., an H instead of an M for “Male” or F for “Female”). The computer catches the error immediately.

6.7: How Can You Be Ethical in Survey Research? 6.7 Analyze some of the ethical issues in conducting survey research There are ethical and unethical ways to conduct surveys. A major ethical issue in survey research is the invasion of privacy. People have a right to privacy, but asking about intimate actions and personal beliefs could violate a person’s privacy. You need to be careful and ask questions in a polite, respective manner, and to protect the information you obtain. Respondents are most likely to provide accurate information when asked for it in a comfortable context with mutual trust, when they believe that serious answers are required from them for a legitimate purpose, and if they believe that their answers will remain confidential. This requires treating all respondents with dignity and reducing any anxiety or discomfort. The researcher is responsible for protecting the confidentiality of data. A second ethical issue involves voluntary participation by respondents. You can never force anyone to answer a survey question and must obtain informed consent. Respondents can withhold information or refuse to participate at any time. Because you depend on voluntary cooperation, you need to ask well-developed questions in a sensitive way, treat respondents with respect, and keep what you learn confidential. A third ethical issue is the misuse of surveys. Because surveys are so common, some people use the survey format for illegitimate purposes. When someone uses a survey format to attempt to persuade others to do something, it is called a pseudosurvey. A charlatan may have no interest in learning information from a respondent but only use a

survey to gain entry into homes, persuade a person to vote in a certain way, or try to sell something in the guise of a survey. The mass media report more on surveys than other types of social research; however, the way mass media report on surveys permits abuse. Few people who read poll results in a newspaper or hear them on television realize that ethical codes require certain details about the survey method. The purpose of providing details on method is to reduce misuse of survey research. Researchers urge the media to include the information, but it is rarely included. In addition to omitting critical details on how a survey was conducted, the media often report on weak, biased, and misleading surveys along with sound, rigorous, professional ones without any distinction. This only increases public confusion and a distrust of all surveys. Making It Practical: Ten Items to Include When Reporting Survey Research Below are

10items that national public opinion organizations recommend to include when reporting any poll or survey: 1. The sampling frame used (e.g., telephone directories) 2. The dates on which the survey was conducted 3. The population that the sample represents (e.g., U.S. adults, Australian college students, housewives in Singapore) 4. The size of the sample for which information was collected 5. The sampling method (e.g., random) 6. The exact wording of the questions asked 7. The method of the survey (e.g., face to face, t elephone) 8. The organizations that sponsored the survey (paid for it and conducted it) 9. The response rate or percentage of those contacted who actually completed the questionnaire 10. Any missing information or “don’t know” responses when results on specific questions are reported

WRITING PROMPT Incomplete Reporting of Survey Research For many years, national public opinion organizations outlined what reports of survey findings should include so that the people would be able to determine whether the survey was of poor quality and unreliable. What three aspects of how a survey was conducted, if omitted, would mislead people the most? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

Summary: What You Learned about Surveys In this chapter, you learned about survey research. It is a process of asking many people the same questions and examining their answers. You also learned some principles of writing good survey questions. There are many things to avoid and to include when writing questions. You learned about the advantages and disadvantages of mail and web surveys, and of telephone and face-to-face interviews. You saw that interviewing, especially face-to-face interviewing, can be difficult. Although this chapter focused on survey research, researchers use questionnaires to measure variables in other types of quantitative research (e.g., experiments). The survey is often called the sample survey because most surveys involve random sampling. Survey researchers try to minimize errors, but survey data often contain them. Errors in surveys can compound each other. For example, errors can arise in sampling frames, from people not answering survey questions, from question wording or order, and from interviewer bias. Do not let the existence of errors discourage you from using the survey. Instead, use care when designing or conducting surveys and exercise caution when generalizing from survey results.

Writing Good Survey Questions 1. Avoid leading questions in surveys, i.e., questions that push respondents toward specific answers. 2. Also avoid vague questions, double-barreled questions, and keep the language of questions simple. 3. Be aware of prestige bias and false premises in survey questions, and ask about separate variables of a hypothesis, not whether respondents accept the hypothesis you are testing. 4. Weigh the advantages and disadvantages of openended versus closed-ended response formats when designing a survey. 5. Keep the answer choices mutually exclusive, exhaustive, balanced, and provide adequate number (over three but less than nine) of choices. 6. Phrase recall questions in ways that make it easy for respondents to remember and phrase sensitive questions in ways to increase honest, accurate answers. 7. Offer a “do not know” or “no opinion” choice most of the time and use a full-filter question when asking about topics about which few respondents may be informed. 8. Be aware of social desirability bias and avoid it when possible.

Quick Review What Is a Social Survey? 1. Survey data come from self-reports. To measure variables, you must convert them into survey questions that people are able and willing to answer. 2. It is possible to test multiple hypotheses at the same time in one survey. 3. In survey research, control variables measure variables that might be alternative explanations in a causal relationship. 4. In survey research, you often must use logic to meet the requirement of time order of variables in a causal relationship.

How Do We Conduct a Survey? 1. You should tailor survey questions to the type of respondent who is in the survey and to the format of the survey. 2. You need to very clear and explicit about the type of information you request in each question. 3. You should create a system to number each respondent’s questionnaire and way to record responses during survey design.

9. Use contingency questions to increase the relevance of questions for specific respondents. 10. Be sensitive to the emotional impact of specific words or phrases.

Shared Writing: What To Do When You Get a Bad Survey Give an example of a survey question that you considered to be highly biased or leading. How would you respond? Choose two of your classmates’ examples and provide advice that you would give to people when they receive a questionnaire or hear an interviewer ask this question? A minimum number of characters is required to post and earn points. After posting, your response can be viewed by your class and instructor, and you can participate in the class discussion.

Post

0 characters | 140 minimum

4. It is always best to pilot test a questionnaire before using it in the research project.

137

Chapter 7

The Experiment

Learning Objectives 7.1 Identify how experiments are the strongest

type of social research for testing cause– effect relationships 7.2 Report the need to make valid comparisons

in experiments 7.3 Analyze the function of independent and

dependent variables as used in experimental research In terms of politics, many observers like journalists, policy makers, informed-educated citizens, and researchers in the United States and in other countries see Americans as a deeply polarized society. Recent news stories report solid red states versus blue states, refusals to compromise, and political gridlock. Poll data show that divisions by political

138

7.4 Describe combining and managing as the

two main tasks in experimental designs 7.5 Report some of the issues that might

weaken internal and external validity 7.6 Identify the importance of comparisons in

experimental research 7.7 Analyze why ethical concerns often arise

inexperimental research party in the 2008 and 2012 elections were sharper than any time in the past 40 years. Others claim that most ordinary people are not polarized and share many common beliefs. What influences perceptions of political polarization? Why do some perceive sharp partisan divides whereas others do not? Van Boven, Judd, and Sherman (2012) thought that

The Experiment 139

people who were intensely committed to a particular point of view and held extreme partisan attitudes might be projecting their own polarization onto others. In short, not everyone is polarized; rather the people who feel most polarized see others in highly polarized terms. The researchers identified three social psychological processes to account for this situation: 1. a general tendency to categorize others 2. a shift from categorizing to a dividing of people between in-groups (“the people like me”) and out-groups (“others”) 3. “naïve realism.” Naïve realism is a tendency people have to see their own attitudes on partisan issues as unbiased and rational, and the attitudes of people in outgroups as influenced by bias, self-interest, and ideology. Essentially, polarized people assume that they are not biased but that others are. The researchers called these combined processes polarization projection. They created a social experiment to test polarized projection using a fictional political issue that had the essential features of political partisan conflict: well-defined groups with polarized attitudes toward a policy concerning the allocation of scarce resources. They told the experimental participants, undergraduate students at a large public university, that the university administration was seriously considering adoption of a fictional Nonresident Attraction and Retention Program (NARP). The NARP would give nonresident, out-ofstate students (who pay higher tuition than in-state students) privileged benefits, such as priority access to desirable dorms, priority class registration, and free printing. The researchers thought resident and nonresident students would develop partisan attitudes toward NARP. In the experiment, 101 undergraduate students (41 men and 60 women) at the University of Colorado-Boulder first completed a questionnaire on the NARP policy using a 5-point Likert Scale. It asked questions such as what students liked or disliked about the policy and whether it violated their sense of fairness. The researchers next randomly assigned participants into two conditions, either the introspection group or the control group. They gave the introspection group time and cues to ponder why they held their views before they were asked to estimate the distribution of all other students’ attitudes on campus. The control group students just went to estimation. The researchers asked participants in the introspection condition to describe their reasoning processes with questions like: What factors, different thought processes, and experiences might have caused you to hold your stance? Did you consider if the proposal was in your self-interest? Did you engage in careful and extensive thought? Afterwards, the introspection participants estimated the distribution of other students’ attitudes toward NARP.

The results revealed that the participants who had introspected about their attitudinal processes showed more polarization projection than did participants in the control condition. The study demonstrated the most highly committed people overestimate the polarization among others. When polarized people are stimulated to think about why they hold their views (i.e., naïve realism), they become even more polarized and see people who hold an opposing position as stereotypic caricatures. This especially happens when they believe they have exercised careful reasoning and believe those holding an opposing position engaged in biased, self-interested reasoning.

7.1: When Are Experiments Most Useful? 7.1 Identify how experiments are the strongest type of social research for testing cause-effect relationships Compared to the other social research techniques, experiments provide the strongest means for testing causal relationships. In many ways, a research experiment is an extension of commonsense logic. The key difference is that everyday experiments are less careful and systematic than scientific ones. In commonsense language, an experiment refers to two situations: Before-and-After Comparison You modify something and then compare an outcome to what existed prior to the modification. For example, one morning, you try to start your car. To your surprise, it does not start. You “experiment” by cleaning off the battery connections and then try to start it again. You modified something (cleaned the connections) and compared the outcome (whether the car started) to the prior situation (it did not start). Your “hypothesis” was that the car did not start because of a buildup of crud on the connections, and once you removed the crud, the car would start. Side-by-Side Comparison You have two similar

things, and then you modify one but not the other and compare the two. For example, you watch a young boy playing with a soft drink can. He vigorously shakes it. You hold a can that is not shaken. You then open both cans for him. He laughs when the shaken can “explodes” with a fizzy mess, but the one you held does not explode. You began with two similar things (soft drink cans) and modified one (shook up a can) but not the other. Before you opened the cans, you hypothesized that the shaken can but not the other one would make a fizzy mess. Compared to other social research methods, experiments best satisfy the three conditions needed to demonstrate causality—temporal order, association, and no

140 Chapter 7 alternative explanations. In order to demonstrate causality in an experiment, we follow three basic steps: 1. Start with a cause–effect hypothesis. 2. Modify a situation or introduce a change. 3. Compare outcomes with and without the modification.

7.1.1: Questions You Can Answer with the Experimental Method Experimental research has a clear, simple logic that offers a very powerful way to focus narrowly and demonstrate causal relations among a few variables in controlled situations. Research questions most appropriate for an experiment fit its strengths and limitations that include the following features: • It can isolate and identify a causal mechanism. • It targets two or three specific variables, and it is narrow in scope. • It is limited by the practical and ethical aspects of the situations you can impose on humans. In general, the social experiment is better for targeted micro-level concerns (e.g., individual or small-group phenomena) than for complex macro-level issues containing many factors that operate together. In other words, they are excellent when you can isolate few variables in a controlled, small-scale setting but are not well suited for questions involving many diverse influences that operate across an entire society or over decades. For example, to track changes in public attitudes about same-sex marriage across the entire society over the past 20 years, a survey would be much better than an experiment. We can, however, provide insights into larger issues when we integrate and synthesize the results from many narrowly focused experiments.

WRITING PROMPT Limitations of Experiments Give an example where you could conduct a survey and experiment on the same topic. What could you learn from a survey that you could not learn from an experiment? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

7.1.2: Limitations of Experimental Research Ethical and practical constraints limit what you can study with the experimental method. Let us say you want to know whether biracial children are more likely to develop an interest in a career in human services than single-race children

are. You can study this issue using a survey or existing statistics research, but not the experimental method. For a true experiment, you would force some parents to have biracial children and not others, force all the parents to raise their children similarly, and then wait until the children grow up to examine their career choices. This is both unethical and impractical. Despite its great strength at demonstrating causal relations, the experiment is limited in the questions you can ask, the variables you can measure, and your ability to generalize from one experiment to larger society. The ideal solution is to study an issue using both experimental and a nonexperimental methods and then combine what you learn from both. Maybe you are interested in attitudes toward people in wheelchairs. You could conduct an experiment asking participants to respond to photos of people in wheelchairs and not in wheelchairs, with the different research participants seeing the same person in or not in a wheelchair. You could ask them survey questions like the following: Would you hire this person? How comfortable would you be if this person asked you for a date? You could also conduct a social survey on attitudes about disability issues. In addition, you could conduct field research and observe people’s reactions to a person in a wheelchair in a natural setting, or you might be in a wheelchair and carefully note the reactions of others to you. Our greatest confidence comes when many welldesigned studies conducted by different researchers yield similar results.

WRITING PROMPT True Experiments A true experiment often requires manipulating situations or aspects of people lives but within practical and ethical limits. What practical or ethical limits can you see restricting your ability to carry out true experiment on social situations or relations? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

7.2: Why Assign People Randomly? 7.2 Report the need to make valid comparisons in experiments The cliché “compare apples to apples, don’t compare apples to oranges” is not about fruit; it is about making valid comparisons. In experimental social research, what you want to compare must be fundamentally alike. You can facilitate comparison by creating similar groups of research participants. To make a valid comparison, you want to compare

The Experiment 141

participants who do not differ with regard to variables that could be alternative explanations to your hypothesis. Let us say you want to learn whether completing a college course affects a person’s skill level. You have two groups of participants: one group completed the course whereas the other did not. To make a valid comparison, you want participants in the groups to be similar in every respect, except the one you are examining in the study—taking the course. Why is it important to have participants in the groups to be similar in every respect? Compare Your Thoughts Let’s imagine that participants who completed the course are also two years older than those who did not. You could not know whether it was completing the course or being older that accounted for group differences in skill levels.

7.2.1: Random Assignment andRandom Sampling Many experiments use random assignment, also called randomization. To compare groups of participants, you do not assign them based on your feelings, their personal preference, or their specific features. Random assignment is unbiased because your desire to confirm a hypothesis or a participant’s personal interests is removed from the selection process. It is random in a statistical or mathematical sense, not in the everyday sense of unplanned, haphazard, or accidental. In probability theory, random is a process

in which each case has a known and equal chance of selection. It obeys mathematical laws, which make precise calculations possible. Both random sampling and random assignment use a mathematically random process. When you randomly sample, you use a random process to select a smaller subset of people (sample) from a larger pool (population). When you randomly assign, you use a random process to sort a collection of participants into two or more groups. You can both randomly sample and randomly assign. You can first sample to obtain a set of participants (e.g., 150 students out of 15,000), then randomly assign to divide the sampled participants into groups (e.g., divide the 150 people into 3 groups of 50 each). Combining random sampling and random assignment is the ideal; however, because of the extra effort required, it is rare in social experiments (see Figure 7.1). If the purpose of random assignment is to create two (or more) equivalent groups, you may ask whether it might be simpler to match participants’ characteristics in each group. Professional researchers occasionally match participants on certain characteristics, such as age and sex, but this is infrequent. True equal matching falls apart as the number of relevant characteristics expand, and soon finding exact matches is impossible. Individuals differ in thousands of ways. You cannot know which traits are relevant and influence your variables. Let us say you compare two groups of 15 students. Group 1 has eight males. To match, group 2 must also have eight males. Two males in group 1 are only children; the rest have one or more siblings. Three

Figure 7.1 Random Assignment and Random Sampling Random Sampling Population (Sampling Frame)

Sample

Random Process

Random Assignment Step 1: Begin with a collection of subjects.

Step 2: Devise a method to randomize that is purely mechanical (e.g., flip a coin). Step 3: Assign subjects with “Heads” to one group

and “Tails” to the other group.

Control Group

Experimental Group

142 Chapter 7 are from a divorced family. One male is tall, slender, and Jewish; another is short, heavy, and Methodist; and so forth. To match groups, you have to find a tall Jewish male only child from a divorced home and a short Methodist male only child from an intact home. True matching on more than two characteristics soon becomes impossible. For this reason, randomization is preferred. With a true random process, over the long run, the odds are that the groups will be equal. If the groups are equal, the experiment’s internal logic is stronger. Making It Practical: How Do You Conduct Random Assignment?

7.3: How Do We Use Variables in an Experiment? 7.3 Analyze the function of independent and dependent variables as used in experimental research Experimental research has its own terms, concepts, and logic.

In the experimental context, the terms independent variable and dependent variable have specific meanings. In experiments, the independent variable is a condition or situation you modify or alter. For this reason, we also call it the treatment, manipulation, stimulus, or intervention. Experimental researchers strive to create realistic conditions to produce specific reactions and feelings within the experimental participants. For the independent variable, an experimenter often manipulates what different sets of participants see or think is happening. He or she may give them different instructions, show them different situations using visual images or elaborate equipment, use different physical settings, or stage different contrived social situations.

Example Study Random assignment is simple in practice. Start with a collection of people, such as 50 volunteers who show up to participate in a study. Next, divide them into two or more equal-sized groups using a true random process, such as tossing a coin or throwing dice. If you toss a coin, you may assign all for whom heads appear in one group and the rest in another. You will probably have about 25 participants in each group. What if you have 80 people and want to assign them to four groups? You can use coin toss (every other toss goes to a group), dice, or another pure random process, such as a randomizing software application. The key feature of the process is that all have an equal, one-in-four chance, of ending up in a group. Nothing about a specific participant or an experimenter’s preference affects who goes to which group. It is entirely due to pure mathematical chance.

WRITING PROMPT Random Assignment or Matching Imagine that you have 100 volunteers to be participants in an experiment, and you want to randomly assign them into four groups of equal size (i.e., 25 each). How would you randomly assign the 100 into four groups? Be very specific in explaining what you would do. The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

Was It a Gun or a Tool?

In many situations, police officers must respond quickly and accurately to a potentially dangerous situation, such as trying to determine if an individual is holding a gun. In 2014, national outrage erupted when a white police officer shot an unarmed black teen in Ferguson, Missouri. A few weeks later, a white police officer shot and killed a 12-year-old black boy in a park holding a toy gun just 2 seconds after pulling up in his squad car. Earlier, in 1999, four white New York City police officers shot and killed an unarmed immigrant from West Africa who was holding out his wallet which the police mistakenly thought was a gun. Payne (2001) created two experiments to learn whether racial stereotypes about “dangerous people” interferes with the person’s ability to make accurate split-second judgments. Payne built on past studies of priming that found people link

The Experiment 143 visual and other images to preexisting negative stereotypes. The images prime or activate a negative response, often automatically and unconsciously, in people who hold stereotypes. If revealing a negative racial stereotype is socially inappropriate, people will try to control or hide its priming effects. However, hiding the expression of a stereotype slows a person’s decision making. For a short time period (perhaps several seconds), people must reconsider their public responses in order to make them more socially acceptable. In the experiment, Payne had 31 white undergraduate students complete an attitude survey. Next, they looked at visual images on a computer. The experimenter told participants he was measuring their speed and accuracy of identifying visual images. Participants practiced using the equipment and classifying 48 photos. Payne then told them that they would see a pair of photos. The first one was a warning that the second would soon appear. The first photo was always the face of either a white or a black man. The second photo was of a handgun or a tool (hand drill or socket wrench). The first photo appeared for 200 milliseconds; the second photo appeared for 200 milliseconds. After a participant responded, the screen went blank. Participants did not have a fixed time by which they had to respond. After the participants gave a response, the next pair of photos appeared. Participants identified 192 photos with the race of the first photo and tool or gun randomly mixed. Thus, Payne randomly mixed what each participant saw instead of randomly assigning the participants. In a second experiment, everything was the same except that Payne added pressure. If participants failed to respond within 500 milliseconds after the second photo, they saw a dramatic red warning and had 1 second to respond. In the first experiment, the participants showed no differences in accuracy. In the second experiment, with time pressure, the error rate was much higher. Many of the participants who first saw the black man’s face mistook a hand tool for a gun. However, their errors of seeing a tool as being a handgun did not increase after they saw a white man’s face. The participants who made the most errors, i.e., mistaking a tool for a gun, were participants who most strongly accepted negative stereotypes about African Americans based on answers to the survey at the start of the study.

In many experimental designs, you measure the dependent variable more than once, both before introducing the independent variable, in a pretest, and again after introducing it, in a posttest. When introducing an independent variable, researchers can use two or more groups that experience different situations or a single participant group that experiences different situations one after the other. If there are two or more groups, it is an independent group design. In the study that opened this chapter, the independent variable was to have one group of participants “introspect” and seriously consider their reasoning for holding a partisan position, whereas the control group did not. The group that introspected was the experimental group. When the independent variable has many different values, you may have more than one experimental group, called comparison groups, one for each level of the independent variable. At other times, experimenters use a repeated measures design. There is one group and the same participants experience multiple situations over time as was the case in the study about viewing images of a black or white face then a gun or tool.

7.3.1: Planning an Experiment To plan an experiment, you must decide on a specific experimental design and plan what participants will experience from beginning to end. Planning includes decisions about the number of groups, how and when to create independent variable conditions, and how often to measure the dependent variable. Overall, you need to plan and configure seven parts of a design. Not all experiments have all seven parts, but a few experimental designs have more than these seven. 1. Independent variable 2. Dependent variable 3. Pretest 4. Posttest 5. Experimental group 6. Control group

Analysis of Was It a Gun or a Tool? In the experiment described in Example Study: Was It a Gun or a Tool?, the independent variable is different priming that the experimenter created by showing participants either a white man or black man’s face. Experimenters measure dependent variables in many ways, including response times, percent accurate scores, social behaviors, attitudes, feelings, and beliefs. In the experiment described in the Example Study, the dependent variable was participant accuracy about whether a second photo showed a gun or a tool. You can measure dependent variables by paper-and-pencil indicators, such as a questionnaire, by interviews, by observing behaviors (making a choice or response time), or by physiological responses (e.g., heartbeat or sweating palms).

7. Random assignment You also develop measures of the dependent variable and pilot test the experiment.

Summary Review

Steps in Conducting anExperiment 1. Begin with a straightforward hypothesis that is appropriate for experimental research. 2. Decide on an experimental design that will test the hypothesis within practical limitations.

144 Chapter 7 3. Decide how to introduce the independent variable or create situations to induce it. 4. Develop a valid and reliable measure of the dependent variable. 5. Set up an experimental setting and conduct a pilot test of the variables. 6. Locate appropriate participants. 7. Randomly assign participants to groups and give them instructions. 8. Gather data for the pretest measure of the dependent variable for all groups. 9. Introduce the independent variable to the experimental group only (or to relevant groups if there are multiple experimental groups) and monitor all groups. 10. Gather data for posttest measure of the dependent variable. 11. Debrief the participants. This means you ask participants what they thought was occurring and reveal the true purpose and situation you deceived them about in any aspect of the experiment. 12. Examine data collected and make comparisons between different groups using statistics and graphs to determine whether the data support the hypothesis.

7.4: How Do We Combine Parts into Experimental Designs? 7.4 Describe combining and managing as the two main tasks in experimental designs Designing an experiment involves two tasks: 1. Combining several parts into a specific configuration 2. Managing the overall experiment The logic of experimental research requires that you carefully manage all aspects of the experimental setting. In order to accomplish this, you must isolate the effects of the independent variable and eliminate alternative explanations. Any aspects of an experimental situation that you do not control might become an alternative explanation for changes in the dependent variable. Alternative explanations weaken your ability to show a causal connection between the independent and dependent variables. One way that experimenters control the experimental setting is with deception. Carefully applied, this can help to prevent participants from altering their behavior or opinions to fit the researcher’s hypothesis. By focusing attention on a false topic, the unaware participants will act more “natural” with regard to the variables of interest. Experimenters have deceived participants by intentionally

misleading them with written or verbal instructions, by using the actions of helpers or confederates, or by arranging a setting. Experimenters also invent false dependent variable measures to keep the true ones hidden. Obviously, using deception raises ethical concerns.

7.4.1: Types of Experimental Design To plan an experiment, you combine the parts of an experiment (e.g., pretests, control groups, etc.) into an experimental design. Experimental designs vary in their components: Some do not use random assignment, some lack pretests, some do not have control groups, and others have several experimental groups. In research reports, experimenters name widely used designs. To understand experimental design, it is helpful to learn the standard designs that illustrate several ways to combine the parts of an experiment. We can illustrate the standard experimental designs by looking at variations on this simple example: You want to test whether servers (waiters and waitresses) receive bigger tips if they introduce themselves by first name before taking an order and return 8 to 10 minutes after delivering food to ask, “Is everything fine?” The independent variable is server behavior. The dependent variable is the size of the tip received. We can divide standard designs into three groups: • True experimental • Pre-experimental • Quasi-experimental designs

7.4.2: True Experimental Designs Classical Experimental Design We will start with the “gold standard” of the true experimental designs, the classical experimental design. All other designs are variations on it. Classical experimental design has random assignment, a pretest and a posttest, an experimental group, and a control group. Example. You give 40 newly hired servers a training session and instruct them to follow a script. They are not to introduce themselves by first name or return during the meal to check on the customers. You randomly divide the participants into two equal groups of 20. You record the average weekly tips for all participants for one month (pretest score). Next, you retrain one group of 20 participants (the experimental group). You instruct them henceforth to introduce themselves to customers by first name when taking an order and to check on the customers 10 minutes after delivering food. They are to smile and ask, “Is everything fine?” You instruct the other participants (control group) to continue without an introduction or checking. Over a second month, you record average weekly tips for both groups (posttest score).

The Experiment 145

Figure 7.2 Example Using Classical Experimental Design Month 1 Randomly assign participants to training sessions

Month 2

Group 1 Serve food without introduction or checking

Pretest (amount of tips)

Independent Variable Present Self-introduction and return to check on customer

Posttest (amount of tips)

Group 2 Serve food without introduction or checking

Pretest (amount of tips)

Independent Variable Absent Serve food without introduction or checking

Posttest (amount of tips)

Two-Group Posttest-Only Design This design

has all the parts of the classical design except a pretest. It is similar to the static group comparison, a type of two-group pre-experimental design. The two-group pretest-only design includes random assignment absent in the static group comparison. Random assignment improves the chance that the groups are equivalent, but without a pretest you cannot be as certain that the groups really began equal on the dependent variable.

ependent variable or improves their posttest score. Richd ard Solomon developed a design to address this issue by combining the classical experimental design with the twogroup posttest-only design and randomly assigning participants to one of four groups. Example. You randomly divide 80 newly hired servers into four groups and give them training. You instruct participants in groups 1, 2, and 4 not to introduce themselves or to return during the meal to check on the customers. You instruct participants in group 3 (experimental group 2) to introduce themselves and to return during the meal to check on the customers. During the first month, you count average weekly tips for groups 1 and 2 only (pretest). After the first month, you “retrain” group 1 participants (experimental group 1) henceforth to introduce themselves to customers by first name and to return during the meal to check on the customers 10 minutes after delivering food. All other groups continue as first instructed. During the second month, you record average weekly tips for all groups (posttest) (see Figure 7.4).

Example. You randomly divide 40 newly hired servers into two groups and give all training. You instruct one group not to introduce themselves by first name or to return during the meal to check on the customers. You instruct participants in the other group to introduce themselves to customers by first name and to check on the customers 10 minutes after delivering food. You record average weekly tips for both groups (posttest score) (see Figure 7.3). Solomon Four-Group Design It is possible that your pretest measure sensitizes participants to your

Figure 7.3 Example with Two-Group Posttest-Only Experimental Design Month 1 Randomly assign participants to training sessions

Month 2

Group 1 Serve food without introduction or checking

Independent Variable Present Self-introduction and return to check on customer

Posttest (amount of tips)

Group 2 Serve food without introduction or checking

Independent Variable Absent Continue to serve food without introduction or checking

Posttest (amount of tips)

Figure 7.4 Example with Solomon 4-Group Experimental Design As you can see, groups 2 and 4 are identical, except that group 2 has the pretest. Groups 1 and 3 are also the same. Both use a self- introduction and checking on customers. Month 1 Randomly assign participants to training sessions

Month 2

Group 1 Serve food without introduction or checking

Pretest (amount of tips)

Independent Variable Present Self-introduction and return to check on customer

Posttest (amount of tips)

Group 2 Serve food without introduction or checking

Pretest (amount of tips)

Independent Variable Absent Continue to serve food without introduction or checking

Posttest (amount of tips)

Group 3 Serve food with introduction and checking

Independent Variable Present Continue self-introduction and return to check on customer

Posttest (amount of tips)

Group 4 Serve food without introduction and checking

Independent Variable Absent Continue to serve food without introduction or checking

Posttest (amount of tips)

146 Chapter 7

Figure 7.5 Example with Latin Square Experimental Design Notice that everything is the same, except that the server’s self-introduction comes before checking on customers for group 1 and after checking on customers for group 2. The only difference between the two situations is the timing of the self-introduction; it is not whether thereis a self-introduction.

Randomly assign participants to training sessions

Customer Arrives

10 minutes After food

After Customer Finishes Meal

Group 1

Self-introduction and take order

Check on Customer

Offer dessert or bill

Amount of tips

Group 2

Take order

Check on customer

Self-introduction and offer dessert or bill

Amount of tips

Latin Square Designs You use this design when

you are interested in how several treatments in different sequences affect a dependent variable. The Latin square design is just like the classical experimental design, but it has multiple comparison groups. You have two or more levels or types of independent variables. All participants get all the levels or types of independent variables but different sequences. You can determine whether the time order of the treatments has an effect. Example. You randomly divide 40 newly hired servers into two groups and give all training. You instruct one group to introduce themselves by first name when they first arrive at a table to take an order, and then return 10 minutes after serving the meal to ask whether everything is OK. They return after customers finish eating to ask whether the customers would like dessert or are ready for the bill. You instruct participants in the other group not to introduce themselves when they first arrive to take an order, but to return 10 minutes after serving the meal to ask whether everything is OK. They return after a customer finishes eating and only then introduce themselves. They also ask whether the customers would like dessert or are ready for the bill. You compare the size of tips for the two groups (see Figure 7.5). Factorial Designs In all previous designs, you only looked at one independent and one dependent variable. However, at times two or more independent variables operate together to influence a dependent variable. Some research questions suggest that you look at the simultaneous effects of multiple independent variables. Perhaps a server’s behavior is not the only factor affecting the size of tips. Maybe gender also plays a role. You could consider server behavior and gender together: Do male and female servers who introduce themselves and those who do not receive the same tips irrespective of server’s gender? In a factorial design, you use two or more independent variables in combination. You look at every combination of the variable categories (called factors). For two categories for two variables (server checking back and not

Posttest

checking back, and male or female), you need four groups. When each variable contains more than two categories, the number of combinations grows very quickly. In this kind of design, the independent variable becomes each combination of the variables and categories, such as the mix of gender and server behavior. At times, the independent variables combine a pre-existing variable situation, such as gender, with one that the researcher manipulates, such as training for a type of service.

Example. Perhaps from previous studies you

know that servers who self-introduce and check back earn 10 percent higher tips than those who do not (a courteous service effect). You wonder whether there is also a gender effect or whether gender makes no difference. You can repeat the classical experimental design (recall Figure 7.2), only now do both with allmale server groups and all-female server groups (see Figure 7.6). Effects in Factorial Design Factorial designs can influence the dependent variable with two types of effects:

• Main effects • Interaction effects Main effects are present in one-factor or single-treatment designs. They are how one variable alone affects the dependent variable (courtesy or gender alone). In a factorial design, combinations of independent variable categories can also have an effect. These interaction effects occur when specific combinations of variables and

The Experiment 147

Figure 7.6 Example with Factorial Experimental Design Example of 2 (Gender) * 2 (Self-Introduction) Factorial Design Month 1

Randomly assign participants to training sessions

Month 2

All Female Group 1 Serve food without introduction or checking

Pretest (amount of tips)

Independent Variable Present RETRAIN Self-introduction and return to check on customer

Posttest (amount of tips)

All Female Group 2 Serve food without introduction or checking

Pretest (amount of tips)

Independent Variable Absent Serve food without introduction or checking

Posttest (amount of tips)

All Male Group 3 Serve food without introduction or checking

Pretest (amount of tips)

Independent Variable Present RETRAIN Self-introduction and return to check on customer

Posttest (amount of tips)

All Male Group 4 Serve food without introduction and checking

Pretest (amount of tips)

Independent Variable Absent Serve food without introduction or checking

Posttest (amount of tips)

c ategories interact to produce an effect beyond that of each variable alone (see Example Study: Communal Values and Assumed Similarity). Example You discover that servers who self-introduce and check back get 10 percent higher tips than those who do not (a courteous service effect). You notice that female servers receive 6 percent higher tips than male servers (a gender effect). All servers work the same hours with the same number of customers. Let us say that a male server who does not selfintroduce averages $1,000 in tips in a month. If only main effects operated, a female who did not self-introduce and checked back averages $1,060 in tips (a gender effect of 6 percent), a male who self-introduced and checked back averages $1,100 in tips (a courtesy effect of 10 percent), and a female who self-introduced and checked back averages

$1,160 in tips (a courtesy plus a gender effect). Interaction effects are an extra boost beyond each separate effect alone. Perhaps females who self-introduce and check back have an extra boost. The interaction is extra over and above what we expect from gender and courteous service effects alone. If the courteous female servers got 9 percent extra due to interaction effects, they would receive $1,250 in tips (10 percent courtesy, 6 percent gender plus the interaction effect of 9 percent). Because factorial designs can be complex, we have a shorthand way to discuss them. A “two by three factorial design,” written 2 * 3, means there are two independent variables. The first has two categories and the second has three categories. A 2 * 3 * 3 design indicates three independent variables; the first has two categories, and the other two each have three categories.

Example Study Communal Values and Assumed Similarity People tend to predict that the attitudes, beliefs, experiences, behaviors, and traits of other people are similar to their own, but they only extend this prediction to certain other people. They assume more similarity with in-group members (i.e., a shared group membership) than with out-group members. More generally, people assume similarity with others with whom they expect communion (such as being liked, feel close, or share group membership) than with people whom they expect a lack of communion. Locke et al. (2012) conducted an experiment to test the hypothesis that people who hold strong communal feelings (i.e., feel very connected to others) assume a greater belief similarity with those with whom they have a connection (ingroup) compared to those who they view as disconnected (members of an out-group). The study’s dependent variable was the degree of perceived similarity. Its independent variables

were strength of communal feelings and whether a person held a belief that they were similar to other people who were members of an in-group or out-group. Study participants were 217 students at the University of Idaho in the United States and 223 students from three South Korean universities (Baejae, Dongkuk, and Hansung). All students completed a 16-item questionnaire measuring strength of communal feelings in social relations from 1 = not important to 4 = extremely important (e.g., When I am with him/her/them, it is important that they support me when I am having problems). The students were then given a set of 60 paired traits (e.g., impatient-indecisive, shy-dramatic, practical-principled) to rate from 1 “extremely undesirable” to 9 “extremely desirable.” They rated traits as being true of themselves, of students at their university (ingroup) and of students at the foreign university (out-group).

148 Chapter 7 Read More The researchers found that participants saw students at their own university as more similar to themselves than students at the foreign university. This alone is not surprising. Participants with strong communal feelings rated other students more similar to themselves than participants with low communal feelings. Looking at the two independent variables together, the researchers found an interaction effect. Students who scored high on communal feelings rated students in their own country

(in-group) as being more similar to themselves (compared to low communal students), and they also rated foreign students (out-group) as being more different. In other words, the strongly communal-oriented participants overestimated their similarity to in-group members and simultaneously overestimated their difference from out-group members. A communal orientation and evaluating an out-group combined to accentuate perceptions of out-group difference and to accentuate perceptions of in-group similarity.

Figure 7.7 Interaction Effect, Communal Values by Assumed Similarity Assumed Similarity Based on Communal Values and In-group/Out-group (own campus or not) Assumed Similarity Coefficient

0.5 Out-group In-group

0.4 0.3 0.2 0.1 0.0 –2

–1

Communal Values (Standardized)

WRITING PROMPT Factorial Designs It is easiest to think of an experiment having one independent and one dependent variable. Nonetheless, an experiment with more than one independent variable (i.e., factorial designs) can be valuable. Explain how adding another independent variable to a simple experiment would help you see important interaction effects that would otherwise be impossible to see. The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

7.4.3: Pre-Experimental Designs Some designs lack random assignment and use compromises or shortcuts. Researchers use pre-experimental designs in situations where it is difficult to use the full classical design. Pre-experimental designs have weaknesses that make it more difficult to infer a causal relationship. One-Shot Case Study Design Also called the one-

group posttest-only design, this design has only one group, a treatment, and a posttest. Since there is just one group, there is no random assignment. The design’s logic has an

implicit, unexamined assumption that dependent variable change is due to the independent variable. Although the design is very weak, we use it in daily life, such as classroom instruction. A teacher gives instruction (independent variable), and we assume that the instruction causes student learning (dependent variable). Example. You take a group of 40 newly hired servers and train all of them. You instruct them to introduce themselves to customers by first name and to check on the customers, asking “Is everything fine?” 10 minutes after delivering the food (independent variable). You record the average weekly tips for all servers for one month (posttest score). You assume that the courteous service caused the level of tipping for the month. One-Group Pretest-Posttest Design This design

has one group, a pretest, a treatment, and a posttest. It lacks a control group and random assignment. It allows you to measure change over time, but you lack a comparison group. Again, despite its weakness, we often use it in daily activities. A teacher might use this design if she measures what her students know on the first day of class and again measures how much they know after instruction for an entire academic term. Example. You take a group of 40 newly hired

servers and train all of them. You instruct them to

The Experiment 149

follow a script in which they are not to introduce themselves by first name or return during the meal to check on the customers. All begin employment. You record average weekly tips for one month (pretest score). Next, you retrain all 40 servers and instruct them henceforth to introduce themselves to customers by first name and to check on the customers 10 minutes after delivering the food (independent variable). You record average weekly tips for the second month. The one-group pretest-posttest design is a big improvement over the one-shot case study because you measure the dependent variable both before and after introducing the independent variable. However, without a control group you cannot know whether something other than the independent variable is causing change in the dependent variable. For example, if a teacher not only measured how much her class knew at the start and end of a term but also checked other students who were not in her class, she could better evaluate impact of her class on learning. If she measured gain in knowledge in both her students and other students not in her class at the start and end of a term, she might discover that the students not in her class had gained just as much knowledge on their own. Without a comparison group, she cannot be certain that her instruction alone is the cause of her students gaining knowledge. Static Group Comparison Also referred to as the

posttest-only nonequivalent group design, this design has two groups, a posttest, and treatment. It lacks random assignment and a pretest. Its weakness is that you cannot know whether group differences prior to the experiment caused differences in a posttest outcome rather than the independent variable. Example. You give 40 newly hired servers a train-

ing session and instruct them to follow a script in which they are not to introduce themselves by first name or check on the customers. After one month, you retrain 20 randomly picked participants (experimental group). You instruct them henceforth to introduce themselves to customers by first name and to check on the customers 10 minutes after delivering the food (independent variable). You instruct the rest of participants to continue without an introduction or checking. Over the second month, you record the amount of tips for both groups (posttest score).

7.4.4: Quasi-Experimental Designs These designs allow you to test for causal relationships in situations where a true experimental design is difficult or

inappropriate. Quasi means “as if” in Latin. In general, you have less control with quasi-experimental designs than in true experimental designs. Many are repeated measure designs. Interrupted Time Series In an interrupted time series design, you only have one group. In this design, you make multiple measures of the dependent variable both before and after the treatment. Example. You get a commitment from 20 newly

hired servers for a six-month experiment. You instruct them to follow a script in which they do not introduce themselves or return during the meal to check on customers. You record their average weekly tips for three months. Next, you retrain all servers and instruct them henceforth to introduce themselves to customers by first name and to check on the customers 10 minutes after delivering the food. You record their average weekly tips for three months. Equivalent Time Series An equivalent time series is

another one-group design that extends over time. It is similar to the interrupted time series, but instead of one treatment it has a pretest, then a treatment and posttest, then treatment and posttest, then treatment and posttest, and so on. Example. You get a commitment from 20 newly hired servers for a six-month experiment. You instruct them to follow a script in which they are not to introduce themselves by first name or to return during the meal to check on customers. You record the amount in tips for month 1. Next, you retrain all 20 servers and instruct them henceforth to introduce themselves to customers by first name and to check on the customers 10 minutes after delivering the food. You record average weekly tips for month 2. Next, you retrain all 20 servers and instruct them to stop introducing themselves to customers and checking back on customers. You record average weekly tips for month 3. Next, you retrain all 20 servers and instruct them again to introduce themselves to customers by first name and to check on the customers 10 minutes after delivering the food. You record average weekly tips for month 4. Next, you retrain all 20 servers and instruct them to stop introducing themselves to customers and not to return after delivering the food. You record average weekly tips for month 5. Next, you retrain all 20 servers and instruct them to introduce themselves to customers and to check on the customers 10 minutes after delivering the food. You record average weekly tips for month 6.

150 Chapter 7

Summary Review Table 7.1 A Comparison of Experimental Designs Random Assignment

Pretest

Posttest

Control Group

Experimental Group

One-Shot Case Study

Yes

One-Group Pretest Posttest

Yes

Static Group Comparison

Yes

Classical

Yes

Two-Group Posttest Only

Yes

Solomon Four Group

Yes

Latin Square

Yes

Factorial

Yes

Two-Group Posttest Only

Yes

Time Series Designs

Yes

Design PRE-EXPERIMENTAL

TRUE EXPERIMENTAL

QUASI-EXPERIMENTAL

Figure 7.8 Summary of Experimental Designs with Notation Observe the notation for many standard experimental designs Name of Design Classical experimental design

Design Notation s s

s s

X X X

s s s s

s s X s s s s X sX s s Xc s Xc s s s Xa s s Xb s s Xa s s Xb s

Pre-Experimental Designs One-shot case study One-group pretest-posttest Static group comparison Quasi-Experimental Designs Two-group posttest only

Interupted time series Equivalent time series Latin square designs

s X

s s R

Solomon four-group design R Factorial designs R

s s s s s s

Xa Xb Xc Xa Xb Xc s s

s s s s s s s s

X Xb Xa Xb Xc Xc Xa X

s s s s

X X1 X1 X2 X2

Z1 Z2 Z1 Z2

s s s s

The Experiment 151

7.4.5: Design Notation

X = independent variable

As you have seen, you can arrange the parts of an experiment to create many designs. Researchers have a shorthand symbol system to express the combined parts of a design quickly. Once you learn design notation, you will find it easier to manipulate and compare designs. The system lets you express a complex, paragraph-long description in five or six symbols arranged in two lines. The symbols are as follows: O = observation of dependent variable

R = random assignment The Os are often numbered with subscripts from left to right based on time order. Pretests are O1, posttests O2. When the independent variable has more than two levels, the Xs are numbered with subscripts. Place symbols in time order from left to right. The R is first, followed by the pretest, the independent variable, and then the posttest. Arrange symbols, with each row representing a group of participants (see Figure 7.8).

Example Study Anti-Marijuana Television Ads independent variable was a set of ads with anti-marijuana messages targeted at high-sensation-seeking teens and aired during programs that teens tend to watch. Most students saw three of the ads per week during a campaign.

Palmgreen et al. (2001) conducted a two-part applied research experiment that combined interrupted and equivalent time series designs to test the impact of anti-marijuana television ads on teenage drug use. Past studies told them that teenagers vary in “sensation seeking.” High-sensation seekers have a stronger need for novelty, fast-pasted drama, risk, and emotional impact than low-sensation seekers. The experimenters focused on high-sensation seekers because they were at greater risk to become habitual drug users. In two similar counties, they randomly selected 100 teenagers from local schools, grades 7 to 10. In both counties, students completed monthly surveys each month for 32 months. The surveys measured level of sensation seeking, television watching habits, specific television ads seen, attitudes about drugs, and drug use. Researchers found that the two groups were similar and that their rates of illegal drug use were close to national estimates for teens. Their dependent variable was a report of marijuana use in the past 30 days. Their

In Knox County, Tennessee (around the city of Knoxville), researchers used an interrupted time series design. Starting in March 1996, students completed the surveys each month. The ad campaign ran for four months, from January to April 1998. Students continued to complete surveys through December 1998. Prior to the campaign, marijuana use among high-sensation students had grown steadily, from 17 to 33 percent. Usage declined as soon as the campaign began. It continued to decline steadily, from May to December 1998, dropping to about 20 percent. Drug usage had been much lower among low-sensation students, and it changed only a little over time. The researchers used an equivalent time series design, with two independent variables (ad campaigns), in Fayette County, Kentucky (around the city of Lexington). Here, too, students completed monthly surveys starting in March 1996. One ad campaign ran from January to April 1997, and then it stopped. It ran again from January to April 1998, the same time as the one in Knox County, Tennessee. Prior to the first campaign, marijuana use among high-sensation students had steadily grown from about 20 to 37 percent. Usage declined as soon as the campaign began. It continued to decline, from May to November 1997, to about 27 percent, and then it slowly increased again. The increase continued for about two months into the second ad campaign (January to April 1998), and then use declined. It continued to decline after the campaign, to about 27 percent by December 1998.

Table 7.2 Interrupted Time Series Design, Anti-Marijuana Ad Campaigns Group

March–Dec 1996

Jan–April 1997

May–Dec 1997

Jan–April 1998

May–Dec 1998

Knox County TN

OOOOOOOOOO

OOOOTV ads

OOOOOOOO

OOOOTV ads

OOOOOOOO

Fayette County KY

OOOOOOOOOO

OOOO

OOOOOOOO

OOOOTV ads

OOOOOOOO

O = Dependent variable measure = a monthly survey of teen pot-smoking behavior.

152 Chapter 7 WRITING PROMPT Evaluating an Applied Field Experiment After reading through the study on Anti-marijuana Television Ads in Tennessee, do you see any alternative explanations for a decline inmarijuana use? How strong do you think the dependent variable measure is? How would you strengthen this experiment? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

7.5: What Is Experimental Validity and Why Is It Important? 7.5 Report some of the issues that might weaken internal and external validity The internal logic of an experiment should be tight, so it makes the causal connection between the independent variable and the dependent variable very clear and removes potential alternative explanations for changes in the dependent variable. For example, you want to make certain that it is the self-introduction and courtesy check on customers, not other factors (e.g., changes in the menu, features of the servers), that causes differences in tipping. Anything other than the independent variable that influences the dependent variable threatens a study’s truthful revealing of a causal relationship. We also have less trust in the truth of a causal relationship when we see the relationship operating in laboratory studies but not in the “real world.”

7.5.1: Internal Validity The core idea of validity is being truthful or trustworthy. Internal validity is showing truthfulness by eliminating potential causes of the dependent variable other than the independent variable by the controlling conditions within the experimental situation and using the logic of the experimental design. Next, we examine possible threats to internal validity in experimental design. Researchers uncovered many potential threats to internal validity, or issues that might weaken internal validity. Here we consider seven common ones. Selection Bias This occurs when participants do not

start the same. For example, in an experiment on physical aggressiveness, all the participants put in the experimental group happen to be football, rugby, wrestling, and hockey players whereas all control group participants are classical musicians, chess players, dancers, and painters. Selection

bias is often an issue in designs lacking random assignment. You can detect selection bias by looking at pretest scores because all groups should begin the same on the dependent variable before the independent variable is introduced. History This is the possibility that an unexpected event unrelated to the independent variable occurs during the experiment and influences the dependent variable. History effects are more likely in experiments that continue over a long time. For example, halfway through a two-week experiment to evaluate participants’ attitudes toward space travel, a spacecraft explodes on the launch pad, killing the astronauts. Maturation This is the threat that a biological, psy-

chological, or emotional process within the participants that is not part of the independent variable induces a change in the dependent variable. Like the history effect, maturation effects are more common in experiments over a long time. For example, during an experiment on reasoning ability that lasts six hours without a break, participants become hungry and tired. As a result, they score lower. Designs with a pretest and control group help you to detect whether maturation or history effects might be present. You may suspect such effects if you see similar changes in both experimental and control groups. Testing Sometimes the act of measuring the dependent variable with a pretest itself affects an experiment. This threatens internal validity because the act of measuring in addition to the independent variable could have an effect on the dependent variable. The Solomon four-group design helps you to detect testing effects. For example, several years ago I wanted to experiment with a new teaching technique (independent variable) to raise student learning (dependent variable). Students were randomly assigned to two sections of the same course that I was teaching. On the first day of class, I gave students in both sections the final examination (pretest). During the academic term, I used the new teaching technique in one section (experimental group) but not the other (control group). On the last day of class, all students got the same final exam as on the first day of class (posttest). Perhaps students remembered the exam questions from the pretest and this affected what they learned (i.e., paid attention to) or how they answered questions on the posttest. If that had occurred, then a testing effect was present. Experimental Mortality This effect arises when

some participants do not continue throughout the experiment. Although mortality does not mean that participants died, if some participants leave partway through an experiment, you cannot know whether the results would be different had all stayed. For example, you begin a six-week weight-loss program with 50 participants. At the end of the program, 30 remain. All who remained to the end lost

The Experiment 153

15pounds and reported no side effects. You do not know whether the 20 who left differed from the 30 who stayed. Maybe the program was highly effective for those who left, and they withdrew after losing 25 pounds. Perhaps they saw no results and quit in disgust. Or maybe the program made them sick, so they quit. To detect this threat, you must report the number of participants in each group from the start to the end and in the pretests and posttests. Contamination or Diffusion of Treatment

This is the threat that participants in different groups communicate with each other and learn about how the other group is being treated. You can reduce it by isolating groups or having participants promise not to reveal anything to others. For example, you place participants in a day-long experiment on a new way to memorize words. During a lunch break, participants from the experimental group tell those in the control group about the exciting new way to memorize what they just learned. The control group participants then start to use the technique. This threat is difficult to detect. To discover it, you need outside information such as debriefing interviews or monitoring participants at all times. Experimenter Expectancy An experimenter ’s behavior, too, can threaten causal logic. We are not talking about purposeful, unethical behavior but indirectly communicating expectations unintentionally. Many

r esearchers are highly committed to the hypothesis. They accidentally and indirectly communicate desired findings to participants. For example, you study the effects of memorization training on learning. You believe that students with higher grades will do better at the training and will score higher. You learn participants’ grades. Through eye contact, tone of voice, pauses, and other nonverbal communication, you unconsciously encourage and train the students with higher grades more intensely. Your unintended nonverbal behavior is the opposite for students with lower grades. The way to detect experimenter expectancy is by hiring assistants who are the only ones who have contact with participants. You might give assistants fake transcripts and records showing that the participants in one group are honor students and those in the other group are on academic probation and failing. In reality, all the participants have equal academic records. Should the fake honor students, as a group, do much better in the learning study than the fake failing students, you uncovered experimenter expectancy as a problem. Researchers control experimenter expectancy by using the double-blind experiment. In it, people who have direct contact with participants do not know some details about the study. It is double blind because the participants and the assistant in contact with them are blind to experimental details (see Figure 7.9).

Figure 7.9 An Illustration of Single Blind, or Ordinary, and Double-Blind

Experiments

Single-Blind Experiment Experimenter

Subjects Who Are Blind to True Hypothesis Double-Blind Experiment Experimenter

Assistant Who Is Blind to Details of Treatment

Subjects Who Are Blind to True Hypothesis

154 Chapter 7 For example, a researcher wants to see if a new drug is effective. Using three colors of pills—green, yellow, and pink—the researcher puts the new drug in the yellow pill, puts an old drug in the pink one, and makes the green pill a placebo, a false independent variable that appears real (e.g., a sugar pill without any physical effects). Assistants who give the pills and record the effects do not know which color pill has the new drug. Only another person who does not deal with participants directly knows which colored pill contains the drug and examines the results.

WRITING PROMPT Placebos and Double-Blind Experiments When placebos are used as a “false” or “empty” substitute for a treatment, researchers often use double-blind experiments. Can you think of ways that experimenters using experiments with placebos might avoid experimenter effects other than using the double-blind design? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

reaction to people of a race, or have participants memorize nonsense words to see how well they learn. Such treatments may not be realistic enough to allow generalizing to daily life experiences. • Reactivity. Research participants respond differently than they would in real life to situations because they are aware that they are part of a study.

Summary Review

Threats to Internal and ExternalValidity Table 7.3 Threats to Internal and External Validity Seven Threats to INTERNAL VALIDITY

Four Threats to EXTERNAL VALIDITY

1. Selection

1. Participants not representative

2. History

2. Artificial setting

3. Maturation

3. Artificial treatment

4. Testing

4. Reactivity

5. Experimental mortality 6. Contamination 7. Experimenter expectancy

7.5.2: External Validity Even if you can successfully eliminate all threats to internal validity, there is another issue. The full truthfulness of a study requires that its findings operate beyond a specific study. If a study’s findings operate only in research experiments, they will be useless to the real world and applied situations and the study will lack trustworthiness as an explanation for how the world works. If we cannot generalize a study’s findings to the real world beyond a study, it lacks external validity. To be a fully truthful explanation, findings must operate in settings external to the experimental situation alone. Four threats might weaken external validity: • Participants are not representative. Participants in the study may not fully represent characteristics such as age, gender, education level, race-ethnicity, and so forth of the population to which you want to generalize results. • Artificial setting. You may conduct the experiments in a very “unnatural” setting, such as a campus laboratory that limits generalizing to more chaotic and uncontrolled natural settings, such as a retail store, a street corner, a playground, a working office, or an informal public social setting like a dance club. • Artificial treatment. You may create a treatment, such as showing a photograph of a person’s face to create a

7.5.3: Field Experiments Thus far, we have focused on experiments conducted under controlled laboratory conditions. You can also conduct experiments in real-life or field settings. The amount of control you have varies on a continuum. At one end is the highly controlled laboratory experiment that takes place in a specialized setting or laboratory. At the opposite end is the field experiment. Unlike in a laboratory experiment where participants usually know they are in a study, those in field experiments may be unaware that they are part of an experiment and therefore react in a more natural way.

Learning from History The Hawthorne Effect Elton Mayo discovered a specific kind of reactivity which conducting a series of experiments at the Hawthorne, Illinois plant of Westinghouse Electric during the 1920s and 1930s. Researchers experimented by modifying many aspects of working conditions (e.g., lighting, time for breaks, heating, etc.) and then measured worker productivity. They

The Experiment 155 discovered that productivity rose after each modification, no matter what it was. Thus, improved lighting and reduced lighting, greater heat and reduced heat, and so forth all had the same effect. This curious result occurred because the workers were not responding to the independent variable. They responded to the additional attention they got from being part of the experiment and knowing that researchers were closely watching them. The results were so surprising that the effect was named after the study, the Hawthorne effect. An experimenter’s degree of control over an experimental setting directly affects internal and external validity. In general, laboratory experiments increase internal validity but reduce external validity. They can be logically tighter and better controlled but may be less generalizable. Field experiments increase external validity but reduce internal validity. They are more generalizable but permit less experimenter control.

are a “natural” arrangement of events with the basic features of an experiment. In such studies, “after the fact” researchers gather data, apply experimental logic, and make comparisons to test causal relations. Internal validity can be an issue in natural experiment because experimenters have limited control. In the study on fast food restaurants posting calories on menus, the independent variable was menu labeling at New York City (NYC) fast food restaurants.

Example Study Labeling Fast Food Menus andCalories Consumed

What are some examples of practical techniques experienced researchers use to increase the effectiveness of experiments? Compare Your Thoughts Three are discussed here. 1.

Planning and pilot tests. During the planning phase, think of alternative explanations or threats to internal validity and ways to avoid them. Also, develop a neat and well-organized system for recording data. You will want to devote serious effort to pilot testing any apparatus (e.g., computers, video cameras, audio recorders, etc.) and prepare a backup. If you use confederates, train and pilot test them as well. After the pilot tests, interview the pilot participants to uncover any aspects of the experiment that need refinement.

Instructions to participants. In most experiments, you give instructions to “set the stage.” Word your instructions carefully and closely follow a prepared script each time so that all participants hear the same thing. This ensures reliability. The instructions are also important in creating a realistic cover story if you use deception.

Postexperiment interview. You should interview participants at the end of the study for three reasons. First, if you used deception, you must debrief the participants and reveal the true purpose of the experiment. Second, you can learn what the participants thought and how their definitions of the situation affected their behavior. Finally, you can explain the importance of not revealing the true nature of the experiment to other potential participants.

7.5.4: Natural Experiments Most experiments are intentional and consciously planned events. Natural experiments, also called ex post facto (after the fact) control group comparisons, are field experiments that people did not intentionally plan as an experiment but

A natural experiment regarding fast food occurred in 2008 when New York City (NYC) mandated that fast food restaurants must include calorie information on menus. Fast food restaurants elsewhere, such as in neighboring New Jersey, did not have such a mandate. The implicit hypothesis behind mandating calorie information on menus is that it would cause customers to alter their behavior and increase the purchase of lower calorie, healthy meals. After the 2008 NYC law became effective, Elbel, Gyamfi, and Kersh (2011) examined restaurants of the largest chains located in NYC and Newark, NJ: McDonald’s, Burger King, Wendy’s, and Kentucky Fried Chicken. They matched a restaurant in NYC with one from the same chain in Newark, aligning restaurants/neighborhoods by several characteristics such as population size, age, race/ethnicity, poverty level, obesity rates, and diabetes rates. The restaurant sample included 5 in Newark (the comparison city) and 14 in NYC. They collected survey and sales receipt data over a two-week period in NYC and Newark before and after mandatory labeling began in NYC. The authors had data on 349 children and adolescents aged 1–17 years who visited the restaurants with their parents or alone. They matched receipt data with food items to analyze nutritional content. They used survey data to verify demographic features (e.g., age, race, gender) and whether respondents noticed the calorie information on menus. The authors reported (Elbel, Gyamfi, and Kersh

156 Chapter 7 2011:246), “We did not observe any differences before or after labeling in NYC or Newark. The same was true for male and female participants of various age groups.” After labeling began in NYC, a little over one-half of customers reported noticing calorie information, but few said that the labels influenced their meal choice. Of adolescents who reported noticing the labels, 16 percent said that the information influenced their food choices. Although one-fourth of adolescents said they made purchases to limit calories, the authors found no differences between those claiming to control calories and those who making no such claim. When asked to estimate calories consumed, most adolescents gave underestimates. In NYC, 63 percent underestimated the total calories before labeling, and 59 percent did so afterward, with the average (mean) underestimate 466 calories before labeling and 494 after labeling. The authors noted four limitations to the study: labeling effects might require a longer time to develop, labeling was changing as NYC authorities started to impose fines on restaurants not in full compliance, only one nearby city, Newark, was a control group, and there was no control over people entered or left the restaurants over the two-week period, so they could not track repeat customers.

WRITING PROMPT Deception in Experiments Other than deception, what are other ways to reduce reactivity? Howdoes avoiding reactivity in a study increase threats to a person’s privacy? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

7.6: How Can We Learn from Making Comparisons of Experimental Results? 7.6 Identify the importance of comparisons in experimental research Comparison is critical in all forms of research. Comparing dependent variable results across different conditions or groups in experimental research enables you to see whether the independent variable affected the dependent variable and to detect possible threats to internal validity. An example of the results from a series of five weightloss experiments using the classical experimental design shows the power of comparing data to see potential threats to internal validity and the independent variable’s impact (see Figure 7.10). In the example, the 30 participants in the experimental group at Enrique’s Slim Clinic lost an average of 50 pounds, whereas the 30 in the control group did not lose a single pound. Only one person dropped out during the experiment. Enrique’s clinic appears to be effective. Susan’s Scientific Diet Plan had equally dramatic results, but 11 participants in her experimental group dropped out. This suggests a threat from experimental mortality. Participants in the experimental group at Carl’s Calorie Counters lost 8 pounds, compared to 2 pounds for the control group. You may notice that the control group and the experimental group began with an average of 31 pounds difference in weight. This suggests a threat from selection bias. Natalie’s Nutrition Center had no experimental mortality or selection bias problems. Yet participants in the experimental group lost no more weight than those in the control group. Natalie’s independent variable appears to be ineffective.

Figure 7.10 Comparisons of Results, Classical Experimental Design, Weight-Loss Experiments Comparing the Results of 5 Different Treatments, all with the Classical Experimental Design Enrique’s Slim Clinic Pretest

Posttest

Experimental

190 (30)

140 (29)

Control group

189 (30)

Natalie’s Nutrition Center Pretest

Posttest

Experimental

190 (30)

188 (29)

Control group

192 (29)

190 (28)

Susan’s Scientific Diet Plan Pretest

Posttest

Experimental

190 (30)

141 (19)

Control group

189 (30)

189 (28)

Carl’s Calorie Counters Pretest

Posttest

Experimental

160 (30)

152 (29)

Control group

191 (29)

189 (29)

Pauline’s Pounds Off Pretest

Posttest

Experimental

190 (30)

158 (30)

Control group

191 (29)

159 (28)

The Experiment 157

Pauline’s Pounds Off also avoided selection bias and experimental mortality problems. Her experimental group participants lost 32 pounds, but so did those in the control group. This suggests that the maturation, history, or diffusion of treatment threats to internal validity could be present. Thus, Enrique’s Slim Clinic appears to offer the most effective treatment without any clearly evident internal validity threats.

experimenters use “manipulation checks” in which they evaluate their independent variable and other experimental parts for reliability, realism, and effectiveness. • What are possible threats to internal validity? Consider all the internal validity concerns discussed in this chapter, such as selection bias or experimental mortality, and check whether they are present.

Examining How an Experiment Was Conducted

7.7: How Can You Be Ethical in Conducting Experiments?

When you read an experimental research study, you should ask questions to evaluate its features:

7.7 Analyze why ethical concerns often arise in experimental research

Tips for the Wise Consumer

• How many participants were involved? Experiments that have very small numbers of participants in each group (10 or fewer) are not as strong as those with 40–50 participants per group. • Who participated? Experiments with participants who are all undergraduate volunteers from a single course at one university are weaker than experiments that use participants who are diverse by age, experience, and social background. Likewise, experiments highly unbalanced by demographic factors such as gender (e.g., 70 percent of one gender) are weaker than those that with a 50/50 gender balance. • How many participant groups are in the experiment and if more than one, was randomization used? Experiments that involve at least one control and one experimental group are stronger than those that lack a control group. Multi-group experiments that use random assignment of participants are stronger than those that do not use randomization. • How was the dependent variable measured? Does it have high reliability and validity? • What was the independent variable? The treatment should be highly realistic and tightly controlled. Many

Because many experiments are intrusive, ethical concerns often arise in experimental research. The independent variable may involve placing people in contrived social settings and manipulating their feelings or behaviors. Ethical standards limit the amount and type of intrusion. Researchers use extra care and consult closely with the Institutional Review Board (IRB) if they ever place participants in physical distress or in embarrassing or anxiety-inducing situations. They must painstakingly monitor events and control what occurs. Some researchers use deception. They mislead or lie to participants. Dishonesty is never condoned; however, deception may be acceptable under certain conditions. It may be acceptable if there is no other way for researchers to achieve a clear research goal. Even for a highly worthy goal, there are restrictions. Researchers should never use more than the minimal amount and type of deception that is required to achieve the experimental effect and only use it for a narrowly explicit research purpose. In addition, researcher must always debrief participants as soon as possible and explain the true situation if they used deception.

Summary: What You Learned about the Experiment You have just learned about random assignment and experimental research. Comparison is critical in experimental research, and random assignment is an effective way to create two (or more) equivalent groups to compare. Experimental research can provide you with precise evidence for a causal relationship. It produces quantitative

results that you can analyze with statistics, and it is often used in evaluation research. You also learned about the parts of an experiment and how to combine them to produce different experimental designs. In addition to the classical experimental design, you learned about pre-experimental and quasi-experimental

158 Chapter 7 designs. You also learned the symbol system of design notation and how the logic of experiments helps strengthen an experiment’s internal validity. Threats to internal validity are sources of explanations for changes in the dependent variable that can be alternatives to an experiment’s independent variable. You also learned about external validity and how field experiments can maximize external validity. The greatest strength of experimental research is its ability to control alternative explanations and its logical rigor to establish evidence for causality. Experiments are easier to replicate, less expensive, and less time-consuming than the other research techniques. Experimental research also has limitations. First, you cannot use it to address some questions because control and experimental manipulation are impossible. Also, you can usually test one or two hypotheses at a time with experiments, which fragments the growth of knowledge. External validity is another potential problem. Too often experimental researchers use very small, nonrandom samples of college students in one location and then generalize to the entire adult population worldwide. You saw how a careful examination and comparison of results can alert you to potential problems. Finally, you learned about practical and ethical issues with experiments.

Quick Review When Are Experiments Most Useful? 1. Comparing outcomes before and after a modification or with/without a modification is central in the logic of experiments. 2. Experiments are the strongest type of social research for testing cause-effect relationships. 3. Experiments are most effective for looking at a sharply focused issue involving only a few variables. 4. Experiments are limited to situations that can be manipulated and that are within basic ethical standards.

How Do We Use Variables in an Experiment? 1. To make valid comparisons in experiments, we try to compare participants who do not differ with regard to variables that can be alternative explanations to the independent variable in the hypothesis. 2. An experiment may have seven parts: independent variable, dependent variable, pretest, posttest, experimental group, control group, and random assignment. 3. Experimental researchers strive to create realistic conditions for the “treatment” or independent variables that produce specific reactions and feelings within participants. They may manipulate what participants see or think is happening.

4. Experimenters often create an experimental group that gets the “treatment” and a control group that does not, but there are variations with several groups possible. Also, experimenters may use a repeated measures design with only one group.

Table 7.4 Summary of Key Term Definitions Key Term

Definition

Pretest

a measure of the dependent variable prior to introducing the independent variable in an experiment

Posttest

a measure of the dependent variable after independent variable has been introduced in an experiment

Experimental group

in an experiment with multiple groups, a group of participants that receives the independent variable or a high level of it

Control group

in an experiment with multiple groups, a group of participants that does not receive the independent variable or receives a very low level of it

Repeated measures design

an experimental design with a single participant group that receives different levels of the independent variable

How Do We Combine Parts into Experimental Designs? 1. In an experimental design, you combine essential elements (i.e., pretest, posttest, randomization) to test the impact of the independent variable. The three types of designs are: true experimental designs, pre-experimental designs, and quasi-experimental designs. 2. All true experimental designs are based on the Classical Experimental design that has randomization into two groups, experimental and control, and a pretest and posttest. The Two-group Posttest only has all elements of the Classical design except a pretest and is used in special situations. The Solomon Four-group design combines the Classical with the Two-group Posttest-only design. The Latin Square is a variation on the Classical but can have multiple groups in which more than one independent variable is used, and the sequence of the independent variables is of interest. Factorial designs are another variation on the Classical with multiple groups and multiple independent variables, but the interest is in the effect of a combination of multiple independent variables together. 3. Pre-experimental designs lack one or more of the elements present in the Classical Experimental design. Although they are weaker, we use them in some situations. The One-Shot Case study is one of the weakest designs and only has a treatment and posttest. The OneGroup Pretest-Posttest design is slightly stronger by adding the pretest. The Static Group Comparison has two groups but lacks pretests. 4. Quasi-Experimental designs are generally stronger than Pre-experimental designs but are tailored to special situations and may lack some elements of the Classical Experimental design. The Interrupted Time Series has one group

The Experiment 159 that has multiple pretests and posttests. The Equivalent Time Series also has one group and extends over time; however, it has multiple treatments, with at least one pretest and posttest between the treatments.

5. A double-blind experiment helps eliminate experimenter expectancy because neither the participants nor an experimenter who works directly with participants knows the details of the treatment.

5. Design notation is a symbol system for expressing various experimental designs by arraying X’s and O’s.

What Is Experimental Validity and Why Is It Important?

6. External validity is highest when we can generalize experimental results to “real world” settings outside the experimental situation. Field and natural experiments have higher external validity than artificial (or laboratory) experiments, but artificial experiments allow greater control by an experimenter and often have higher internal validity.

1. The internal validity of an experiment is strongest when we can rule out explanations for changes in the dependent variable other than the experiment’s treatment or independent variable.

7. Reactivity is when participants act differently solely because they are aware that they are part of a study, as with the Hawthorne effect. Reactivity is often lower in field or natural experiments.

2. Three threats to internal validity include selection bias, i.e., differences between the control and experimental group that may affect the dependent variable; history effects, i.e., events outside the experimental situation that might affect the dependent variable; and maturation, i.e., “normal” processes related to the passage of time that may affect the dependent variable. 3. Two threats to internal validity come from design factors: testing effects, i.e., the measurement of the dependent variable in the pretest influences the posttest measure, and experimental expectancy, i.e., the researcher’s unintended influence on the dependent variable based on expectations of certain results. 4. Two threats to internal validity due to participant behavior during an experiment are contamination of treatment, when participants in the control group are unexpectedly and unintentionally exposed to the treatment, and experimental mortality, when all participants in all groups to not complete the entire experiment at the same rate.

Shared Writing: Experiments A high-quality experiment will include random assignment, pretests and posttests, experimental designs, factorial designs, internal validity, and external validity. Of these six specific parts of an experiment listed here, which two would you argue is most important and why? Choose two classmates’ responses whose selection differs from yours and provide a constructive counterargument. A minimum number of characters is required to post and earn points. After posting, your response can be viewed by your class and instructor, and you can participate in the class discussion.

Post

0 characters | 140 minimum

Chapter 8

Research with Nonreactive Measures

Learning Objectives 8.1 List four nonreactive quantitative research

techniques 8.2 Apply physical evidence to create

nonreactive measures of variables 8.3 Describe the content analysis process 8.4 List the process steps of existing statistics

8.5 Compare secondary data analysis with

other forms of statistics research 8.6 Recognize that the primary ethical concern

of physical evidence analysis is to protect people’s privacy and the confidentiality ofdata

research What do you throw away?

Have you ever gone through your trash to find something accidentally thrown away? Thieves, private detectives, and police investigators comb through trash to gather information. Researchers also study trash to learn about social behavior. Just as archaeologists study bits of broken pottery to learn about ancient cultures and reconstruct social

160

life from long ago, you can learn about people and their behavior by studying what they discard. I sometimes look through trashcans in my classroom and notice many soft drink and beverage containers. Over the years, I have seen changes in the popularity of beverages among my students. Urban anthropologists have studied the contents of garbage dumps to learn about lifestyles based on what

Research with Nonreactive Measures 161

people throw away (e.g., liquor bottles indicate alcohol consumption). They found, based on garbage that people underreport their liquor consumption by 40 to 60 percent (Rathje and Murphy, 1992, p. 71). Examine my family’s trash and that of our neighbor, and you will discover differences in our respective eating habits, lifestyles, and recreation habits. My neighbor’s family eats a lot of carryout fast food and delivered pizza. Their primary form of recreation is to watch television, especially professional sports. They drink a lot of soft drinks and beer. They subscribe to a weekly television guide and a sports magazine. My family eats a lot fresh fruits and vegetables and drinks bottled water and wine. We cook from scratch and almost never eat fast food or drink sodas. Our recreation is to go to the theater and to read books and newspapers. We subscribe to two newspapers, and to one news and one arts magazine. Trash is physical evidence that can become data to inform us about human behavior. It is a form of nonreactive data, the topic of this chapter.

8.1: What Makes a Study Nonreactive? 8.1 List four nonreactive quantitative research techniques Most quantitative social research, including experiments and survey research, is reactive. In short, the people being studied know that they are part of a research study and may react based on that awareness. Reactive research can be problematic because people may modify their actions when they are aware that they are part of a study. Experimenters must often use deception to offset this. The example of studying trash is a type of nonreactive research where the people being studied are unaware that they are being studied. Rather than collecting actively created data, nonreactive research involves gathering passive data. There are four nonreactive quantitative research techniques: • Physical evidence analysis • Content analysis • Existing statistics analysis • Secondary data analysis Quantitative nonreactive research techniques have two advantages over ones that are reactive: • People do not act differently simply because you study them or they are aware that data about them are part of a research study. • Data collection can be faster, of lower cost and easier compared to techniques like the survey or an experiment.

Nonetheless, nonreactive research has three limitations. • At times data are pre-collected and available but you have little control over how the data were collected but must rely on the integrity, consistency, and organization of others. • Often there is no simple, direct way to measure what interests you, so you must be creative and devise indirect surrogate measures. • When using indirect forms of data, you must infer the data’s significance and meaning using logic and considering an array of other, related forms of evidence.

8.2: How Can You Use Physical Evidence inResearch? 8.2 Apply physical evidence to create nonreactive measures of variables You can learn about social life by looking creatively at various types of physical evidence. It begins when you notice evidence that indicates a variable of interest. Discarded beverage containers indicate beverage availability and people’s taste preferences. The critical thing about nonreactive or unobtrusive measures is that people generate data without being aware of their possible use for research study. Researchers used many kinds of physical evidence to create nonreactive measures of variables. For example, they have looked at family portraits in different historical eras to see how the seating patterns reflected gender and authority relations within a family; measured public interest in exhibits by noting worn floor tiles in different parts of a museum; compared the graffiti in male versus female high school restrooms to see whether themes in graffiti showed a different perspective by gender; and examined high school yearbooks to compare the high school activities of people who had psychological problems later in life with those who did not have problems. Nonreactive data can confirm or reveal aspects of people’s lives different from what we would learn from direct, reactive data. For example, we know that a person’s answer to a survey question may not always match his or her behavior. Perhaps I really like country music. In a survey question, I say my favorite music is classical because I want to appear refined and sophisticated. In an experiment, if given a choice to listen to country, rock, or classical, I also pick classical music to create an impression of sophistication. However, you might follow how some researchers studied the music listening habits. They checked the radio stations that drivers had tuned to when cars were taken in for service. If you checked my car radio, you might find that

162 Chapter 8 all presets are to stations that play country music. Going through my credit card purchases, checking music files on my MP3 player or cell phone, and reviewing all music downloads to my Internet IP address, you see no classical music, only country music. Nonreactive measures may provide a more accurate indicator of music preferences than direct, reactive measures like a survey or experiment. What do you need to do when conducting a study with physical evidence? Compare Your Thoughts • Identify a physical evidence measure of a behavior or viewpoint of interest. • Systematically count and record the physical evidence. • Identify and measure the variables of your hypothesis. • Consider alternative explanations for the physical data and rule them out. • Compare the variables of your hypothesis using quantitative data analysis.

8.2.1: Limitations of Physical Evidence Physical evidence measures are almost always indirect indicators, so you must infer or make a cautious “educated guess” from the evidence to a person’s preferences, behaviors, or attitudes. For example, you infer that leaving a radio station in my car to a station that plays country music indicates my music preference. You do not know whether I set the radio to that station because it has the best weather forecasts or a favorite announcer. You do not know whether someone else who also drives the car likes country music, while I never listen to the radio when driving. Perhaps someone else takes the car in for service and changes the radio settings. For these reasons, you must confirm inferences from nonreactive data with additional evidence. When you infer indirectly from data, looking for patterns in a very large sample makes it less likely you will be misled, but even then, you need to be cautious. To confirm the meaning of physical evidence, you need to rule out alternative explanations. For example, you can measure walking traffic by customers in a store by the amount of dirt and wear on floor tiles. To use this measure, you first must clarify what the customer traffic really means (e.g., Is the floor a path to another department? Does it indicate a good location for a visual display?). Next, you systematically measure dirt or wear on the tiles and record results on a regular basis (e.g., every month). Next, you compare the wear and dirt in one area to wear in other locations. Finally, you rule out alternative reasons for the data (e.g., the floor tile is of lower quality and wears faster, or the location is near a restroom). Many types of physical evidence can reveal information about human behavior.

Another limitation of nonreactive data is privacy violation. You could be violating a person’s privacy by noting car radio stations or sorting through his or her trash. The potential for privacy violation means that you must take extra care in collecting data to protect anonymity and confidentiality. For example, you can record the color, model, and year of a car with the radio stations, but not the car owner’s name. You observe whether two types of garbage occur together (pizza boxes and beer cans) or neighborhood characteristics and trash type (high-cost single homes and food packages labeled “organic”) but not an individual’s name and address from their discarded junk mail. Creativity is an important aspect of nonreactive physical measures. A researcher needs to think creatively about what the evidence might indicate. Perhaps you notice that people driving bright red or yellow cars seem to speed more than people who drive black or gray ones. You hypothesize that people who are attracted to bright colors and want to be noticed also feel less constrained by rules or laws. If you obtained a radar gun to measure car speed and record car color, you could document the relationship between color and speed. However, you should be aware of alternative explanations and exercise caution when drawing conclusions. Perhaps younger people like “flashy,” brightly colored cars and younger people speed due to inexperience and tendency high-risk toward behavior. Thus, age could be causal factor linked to speeding, not a person’s color preference or his or her general philosophy opposing constraints on individual freedom.

Example Study Finding Data in a Graveyard

Have you ever walked through an old cemetery and read what was on the tombstones? You can learn a lot from old tombstones. Writing on tombstones provides data about conditions in the past. In addition to official written records, which may be incomplete or destroyed over time, we can look at the physical evidence on tombstones. Foster and colleagues (1998) examined the tombstones in 10cemeteries in an area of Illinois for the period from 1830 to

Research with Nonreactive Measures 163 1989. From the tombstones, they retrieved data on birth and death dates and gender. In total, they gathered information from over 2,000 of the 2028 burials in the 10 cemeteries. They learned how the area differed from national trends. For example, they found that conceptions had two peaks (spring and winter), and females aged 10 to 64 had a higher death rate than males. Younger people tended to die in late summer but older people in late winter. Cemeteries can also reveal information about family size and relations (e.g., a married adult woman is buried with her parents and not her husband) or community religious tolerance (e.g., tombstones with the symbols of different religious traditions—Christian, Jewish, Muslim—intermixed versus placed in separate cemeteries).

WRITING PROMPT Finding Data What could you learn about someone from examining their records of credit card purchases, cell phone calls, and Internet web searches/pages viewed? What are the possible privacy issues involved in researching such “traces” of behavior? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

8.3: How Can You Find Content within Communication Messages? 8.3 Describe the content analysis process Content analysis is a nonreactive technique that lets you examine both hidden and visible content in communication messages. The content can be words, meanings, pictures, symbols, ideas, themes, or any message that the text communicates, directly or indirectly. Text appears in all communication media. It is in books, magazine articles or advertisem*nts, speeches, and legal documents. Text can be audio or visual material in films or DVDs, musical lyrics, photographs, clothing or hairstyles, and works of art. Professionals in many fields use content analysis. Researchers have used content analysis to study numerous topics, including: themes in popular songs and religious symbols in hymns, trends in the topics that newspapers cover, the ideological tone of newspaper editorials, sex-role stereotypes in textbooks or feature films, how often people of different races appear in television commercials and programs, answers to open-ended survey questions, enemy propaganda during wartime, the covers of popular magazines, personality characteristics evident from suicide notes, social

class and identity themes in advertising messages, and gender differences in conversations, to name just a few! Objective, systematic counting and recording procedures produce quantitative data on the symbolic content that is present in the text. There are also qualitative versions of content analysis that emphasize interpreting symbolic meaning. Here, we focus on quantitative data about a text’s content. Content analysis is nonreactive because the creator or author of a communication message produces the words, images, or symbols that comprise the text with little or no awareness that someone someday might study it closely. Content analysis allows you to uncover aspects of the content in a communication medium differently from what you would likely learn from ordinary reading, listening, or watching. With content analysis, you can compare content across many texts and analyze it using quantitative techniques (e.g., charts and tables). In addition, you can reveal difficult-to-see aspects of the text’s content. For example, you might watch television commercials and vaguely feel that nonwhites rarely appear in commercials for high-cost consumer products like luxury cars. By using content analysis, you can document—in objective, quantitative terms— whether your vague feelings based on nonsystematic observation match what is actually occurring. Coding enables us to create repeatable, precise data about the text of a communication medium. With coding, you convert aspects of text content into variables and quantitative data. You can apply statistics to examine variables in the data the same way that an experimenter or survey researcher would examine variables in other quantitative data. When is content analysis especially useful for researching issues? Compare Your Thoughts • When you want to examine large volumes of text. For example, with careful sampling and measurement, you can analyze what has appeared in all television programs of a certain type on five major channels across a five-year period. • When you want to study topics “from a distance.” For example, you can study the writings of someone who has died long ago, or the media broadcasts from a distant hostile foreign country that bans entry. • When recognizing the content is difficult with casual observation. For example, you can detect subtle themes or bias (e.g., power relations, gender stereotypes) present in a text of which the text’s creator and consumers are unaware.

8.3.1: How to Measure and Code inContent Analysis In content analysis, you convert a large mass of text information into precise, quantitative data. To do this, you c arefully

164 Chapter 8 design and document procedures in a manner that makes replication possible. It should be possible for someone else to repeat what you have done. To operationalize variables in content analysis, you create a coding system. This set of rules explains how to categorize and classify observations in text. As with other kinds of measurement, you need mutually exclusive and exhaustive categories. The written rules make replication possible and improve reliability. You must tailor the coding system to the specific type of text or communication medium you are examining, such as television dramas, novels, music videos, photos in magazine advertisem*nts, and so forth. You must also tailor it to your unit of analysis. The unit of analysis varies widely in content analysis. It can be a television commercial, a phrase, a book’s plot, a newspaper article, a film character, and so forth. For this reason, decide on the unit of analysis before you develop measures or record any of the data. If your unit is a full-length film, a television commercial, or a newspaper editorial, you adjust the coding to that type of unit. What Do You Measure?

Example Study Content Analysis and Opposition to Aid for the Poor In the mid-1990s, large political movements urging drastic cuts of public aid and services for poor people swept across the neighboring states of Arizona and California, both that border Mexico. Brown (2012) used content analysis to understand the state actions and determine their similarities. She sampled 500 newspaper articles from the largest newspapers of each state between 1993 and 1997. She and an assistant first reviewed all articles during the timeframe for any that mentioned welfare reform, aid for poor people and the state’s name. Next, they systematically sampled to get 50 articles for each year for each state. Using the paragraph in an article as their unit of analysis, they next coded several variables, including mentions of common stereotypes about welfare recipients (lazy, immoral, wasteful) and references to the race-ethnicity and U.S. citizenship status of the recipients. Intercoder reliability was .91. Brown’s analysis revealed big differences between the two states in anti-welfare rhetoric.

Compare Your Thoughts You begin with a preliminary coding system that has rules you can use to conduct a pilot study on a small amount of data. You use pilot study results to refine the coding rules for the full study. As you develop rules, you can measure five characteristics of variables in the text content you will code: • Direction. Note the positive/support or negative/oppose direction of messages in the text relative to an issue, trait, or question. For example, you devise a list of ways an elderly television character can act. Some are positive (e.g., friendly, wise, considerate), and some are negative (e.g., nasty, dull, selfish). • Frequency. Count whether something occurs in the text and how often. For example, how many elderly people appear on a television program within a week? What percentage of all characters are they, or in what percentage of programs do they appear? How frequently do they have speaking parts? • Intensity. Measure the strength of a variable. For example, the characteristic of forgetfulness can be minor (e.g., not remembering to take your keys when leaving home) or major (e.g., not remembering your name, not recognizing your children). • Space. Measure the size, volume, and amount of time or physical space. One way to measure space in written text is to count words, sentences, paragraphs, or physical space on a page (e.g., square inches). For video or audio text, you can measure the duration of time. For example, a TV character may be present for a few seconds or continuously in every scene of a twohour program. • Prominence. Prominence is related to amount space—is it located in a time or physical location to get a lot of attention? A television show aired in “prime time” versus 3:00 a.m. has greater prominence. An article on the front page of a newspaper has greater prominence than one buried inside.

Read More In California, articles on cutting welfare and public services emphasized undeserving illegal immigrants. In Arizona the issue was framed as resource-absorbing, lazy racial-ethnic minorities versus the virtuous, tax-paying white-Anglos. For example, 15 percent of California articles referred to the raceethnicity of poor people versus 40 percent in Arizona. In California, 83 percent of anti-welfare articles emphasized the citizenship status of welfare recipients but only 14 percent in Arizona did so. Only 17 percent of California articles explicitly referred to undocumented immigrants as being Hispanic versus 85 percent in Arizona. Brown also studied hundreds of letters, legal statutes, records of public speeches and testimony, and she interviewed 17 welfare reform leaders in the two states. This other data reinforced the content analysis evidence that each state had a distinctly different anti-welfare frame. In Arizona lowincome Latinos were harshly criticized, whether or not they were legal immigrants, and they were grouped together with other racial-ethnic minorities who were also demonized. By contrast, in California legal immigrants, Latino and others, were praised as being hardworking and upstanding members of the community. Brown concluded that the distinct political cultures of each state produced very different types of antiwelfare movements. Antipathy toward the poor was highly racialized in Arizona, but race-ethnicity was not a major issue in California. In California, the focus instead was on the poor’s lack of U.S. citizenship. Brown argued that studies of race, immigration, and welfare that remain on a nationwide level may miss large differences regarding these same issues that occur at more localized levels.

Research with Nonreactive Measures 165

WRITING PROMPT What Content Analysis Reveals Because it is systematic, content analysis reveals aspects of content within media messages and symbolic expression that are not apparent from casual observation. Describe a theme or recurrent pattern in media you’ve recently consumed (it can be films you viewed, TV advertisem*nts you saw, or images in print advertising). How might you go about documenting this theme or patterning using content analysis? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

8.3.2: Coding, Validity, andReliability There are two major types of coding in content analysis: manifest and latent. With manifest coding, you count the number of times a phrase or word appears in written text, or whether a specific action (e.g., a kiss, a slap) or object (e.g., a gun, a dog) is in a photograph or video scene. The coding system is a list of terms or actions for you to locate in text. For written text, you may be able to use a computer program to search for words or phrases. Manifest coding is highly reliable because a word, object, or action is either present or absent. A weakness of manifest coding is that it cannot take into account the specific meaning, context, and connotations of the words, phrases, objects, or actions. A word, object, or action has multiple meanings, and this weakens the measurement validity of manifest coding. Can one word have many different meanings/ connotations? Compare Your Thoughts I read a book with a red cover that is a real red herring. Unfortunately, its publisher drowned in red ink because the editor could not deal with the red tape that occurs when a book is red hot. The book has a story about a red fire truck that stops at red lights only after the leaves turn red. There is also a group of Reds who carry red flags to the little red schoolhouse. They are opposed by red-blooded rednecks who eat red meat and honor the red, white, and blue. The main character is a red-nosed matador who fights red foxes, not bulls, with his red cape. Red-lipped little Red Riding Hood is also in the book. She develops red eyes and becomes red-faced after eating many red peppers in the red light district. Her angry mother, a redhead, gives her a red backside. In latent coding you consider the visual or written text as a whole and then decide whether it contains certain themes. Instead of lists of specific words or actions, a latent coding system has general rules that guide how you interpret text and determine whether themes or moods are present.

Compared to manifest coding, latent coding tends to be less reliable because it relies on a coder’s in-depth knowledge of language, subtle clues, and social meaning. Training, practice, and clear written rules can improve reliability, especially if several people do the coding. On the other hand, latent coding can have greater measurement validity. This is because we communicate meaning in many indirect and implicit ways that depend on the context, not just on specific words or actions. Latent coding captures the direct and indirect meanings that may be embedded in a specific text context. The ideal is to use both manifest and latent coding, although it is time-consuming. If the two methods agree, your confidence in the results is stronger than with one method. If they disagree, it signals that you should reexamine the operational and theoretical definitions. In most situations, you will code text from a very large number of units. For example, you might code content from dozens of books, over a hundred hours of video, or several hundred websites. For large number of units, you may need to hire assistants. You must instruct any assistants on the coding system and train them. The coder-assistants need to understand the variables, carefully follow the coding system, and discuss any ambiguities. As the coding progresses, you must create a written record to document decisions about each a new coding situation to be consistent. If you use assistants as coders, always check for consistency across them. It is not difficult. Have several coders code the same text independent of one another. If you have three assistants coding television commercials, have the three coders independently code the same 15 commercials. Check for consistency across the coders, and determine whether they coded the 15 commercials the same. With intercoder reliability, you measure the degree of consistency with a statistical coefficient. Always report the coefficient along with the study results. There are several intercoder reliability measures. All range from 0 to 1, with 1.0 equaling perfect agreement among coders. A coefficient of 0.90 or higher is excellent, 0.80–0.89 is good, and 0.70 is the minimum acceptable. When the coding process stretches over a considerable time span (e.g., over three months), recheck coder reliability by having each coder independently code samples of text that were previously coded. For example, the assistants code six hours of television episodes in April and recode the same six hours in August without looking at their original decisions. Deviations in coding between the two times mean you probably need to retrain coders and recode text.

8.3.3: Content Analysis with VisualMaterial Using content analysis to study visual text, such as photographs, paintings, statues, buildings, clothing, videos, and film, is more difficult than doing so for written text.

166 Chapter 8 Visual material communicates messages or emotional content indirectly through symbols and metaphors. Moreover, visual images often contain mixed messages and operate at multiple levels of meaning. Learning to interpret visual media takes substantial effort and skill. You need to be aware of multiple symbolic meanings and references. To conduct a content analysis of visual text, you interpret signs or symbols and the meanings they convey within the visual text. Reading visual text is not mechanical but depends on cultural context. We attach cultural meanings to symbolic images. Most people in one culture share a common meaning for its major cultural symbols. However, visual text is often multilayered with several meanings. Different people can read the same symbol differently. For example, one person “reads” graffiti as vandalism that defaces a building, another reads it as a work of art, and a third reads it as a mark of a street gang’s territory. To conduct a content analysis of images, you need to be aware of possible divergent readings of images or symbols (see Figure 8.1).

Figure 8.1 Visual Text Is Cultural Bound

the social situation as well as the culture. Smiling at a funeral is appropriate in some cultures but highly disrespectful in others. A smile’s meaning can also change over time. Perhaps you noticed that in very old photographs, the people never smiled. This was not because they were always unhappy but because smiling as a social convention when being photographed only developed later (in the 1920s in the United States). (Also see Rashotte, 2002) Symbol use can be a source of conflict. Sociopolitical groups invent or construct new symbols and attach meanings to them. For example, the Nazis originally used a pink triangle in concentration camps to mark hom*osexuals who were condemned to extermination along with Jews, and other “undesirables.” The pink triangle later came to mean gay pride. Competing sociopolitical groups often wrestle to control the meaning of symbols. Thus, some people want to assign a Christian religious meaning to the Christmas tree; others say it represents a celebration of tradition and family values without religious content; others see its origins as an anti-Christian pagan symbol; and still others see it as a profit-oriented commercial symbol. Because a symbol has complex, multilayered meanings, you must make qualitative judgments about how to code images.

Checking the Meaning of Culture Bound Symbols

Example Study Magazine Covers and CulturalMessages

Many people associate the swastika with the German Nazi government or extreme racist groups, but various cultures had used the symbol for over a thousand years before the Nazi movement adopted it. You will see the swastika on religious buildings across Asia where it is a good luck symbol and is used in decorative art and clothing. This shows how a symbol’s meaning depends on when and where it appears (see Quinn, 1994) Another example is the smile. We treat a smile as a reflex indicating a positive feeling, and it is nearly a cultural universal. However, it is not always a friendly sign. Its meaning varies by how and when it appears. In some cultures, a smile can signal deceit, insincerity, or frivolity. There are several types of smiles, and smiling depends on

Chavez (2001) conducted a content analysis of the covers of major magazines that dealt with the issue of immigration into the United States. Looking at the covers of 10 magazines from the mid-1970s to the mid-1990s, he classified them into sending one of the following messages: affirmative, alarmist, or neutral and balanced. He also examined the mix of people (i.e., race, gender, age, and dress) in the photographs and whether major symbols, such as the Statute of Liberty or the U.S. flag, appeared. Chavez argued that magazine covers are a cultural site. They are a place where media create and communicate cultural meanings to the public. Visual magazine covers carry multiple levels of meaning. When people see a cover and apply their cultural knowledge, they construct specific meanings. Collectively, the covers convey a worldview and express messages about a nation and its people. For example, we usually see the icon of the Statute of Liberty as strong and full of compassion. Its usual message is “Welcome immigrants.” However, when a magazine cover altered this icon to give it Asian facial features, its message shifted to become “Asian immigrants are distorting the U.S. national culture and altering the nation’s racial make-up.” When a magazine showed this icon holding a large stop sign, its message became “Go away immigrants— we do not want you.” The symbolic messages sent by visual images can have a powerful emotional effect on people that is sometimes stronger than written text.

Research with Nonreactive Measures 167

WRITING PROMPT Symbolic Messages in Content Analysis In addition to the analysis of words, we can use content analysis to examine symbolic messages that are in visual images. Look at the photo of a “staged” politician’s speech or the public relations image of a university campus or its students. What “symbolic messages” do you see? How might you go about conducting a systematic content analysis study of such images? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

8.3.4: Content Analysis ResearchSteps In the following sections, we will explore each of the steps for conducting content analysis research: 1. Formulate a research question 2. Identify the text to analyze 3. Decide on units of analysis 4. Draw a sample 5. Create a coding system 6. Construct and refine coding categories

Step 3: Decide on Units of Analysis This is a major early decision in content analysis. There are many possible units in content analysis—the page, the episode, the character, and so forth. The unit of analysis determines the unit size and quantity of text to which you assign a code. For example, for a political campaign, you may code what each newspaper article reports (excluding editorials, letters to the editor, and advertising). In this case, the article is your unit of analysis. You still need to determine how to identify “campaign-related” articles from others (such as by scanning the headlines). Step 4: Draw a Sample Random sampling works

well for most content analysis studies because you often have a huge collection of units to which you want to generalize (such as all newspaper articles) but only have time to code a small proportion of the units. The steps are as follows: • Define the population (e.g., all articles, all sentences). • Select the sampling element (your unit of analysis). • Create a sampling frame. • Use a random selection process. Making It Practical: Sampling in Content Analysis

7. Code the data onto recording sheets 8. Data analysis Step 1: Formulate a Research Question Begin

with a topic that you turn into a research question. Content analysis is appropriate when your research question involves messages or symbols. Conceptualize each variable of the research question. Let us say you want to study how newspapers cover a U.S. presidential political campaign. Refine the idea of “coverage”—do you mean the amount of coverage, prominence of the coverage, or direction of coverage? You must decide how to examine newspaper coverage. You could survey people about what they think of newspaper coverage or examine the newspapers directly using content analysis. Your research question will guide you to the variables to measure. You might have a research question such as: Does the newspaper give greater coverage to one presidential candidate over the other as the date of the election gets closer?

This suggests the variables of amount of coverage for each candidate and dates of coverage relative to the election. Step 2: Identify the Text to Analyze Find the

communication medium that best matches your research question. Your research question and decision about type of text are usually a single process—candidate coverage and the newspaper as a type of text (vs. television, radio, or other media). You still must identify the specific text (such as which newspapers) and its scope (which dates).

We often sample in content analysis because the number of units quickly becomes impossible to manage otherwise. Suppose your research question is: How do U.S. TV commercials that air during different men’s and women’s sports events differ?

168 Chapter 8 In addition to comparing gender, you wonder whether a sport with “mass appeal” like basketball differs from a “highbrow” sport such as golf. Your unit of analysis is the commercial. Your population is all commercials for men’s (NBA) and women’s (WNBA) professional basketball, and men’s (PGA) and women’s (LPGA) professional golf across a time period of interest. At the beginning, you will need to define precisely the unit of analysis, the commercial. Do audio-only statements “sponsored by” without visuals count? Does a logo or banner scrolling across the bottom of the screen count? Does brand logo on or beside a sports announcer count? Is there a minimum time length for a commercial? How soon before the start or after a sports program counts as a sports commercial? How do you count repeats of the same commercial? Next, you examine possible sports programs. You will need to find out how many sports programs for the relevant sports have been televised and on what channels. You check and learn that NBC and ESPN began to televise WNBA games for the 2001 season. NBA sports have been televised on ESPN and another network (NBC, CBS, ABC) over time. The Golf channel began in 1995 and it has had exclusive rights to the PGA for several years, but on occasion other networks such as Fox Sports showed tournaments, so you must decide whether to limit yourself to one network for the sports. Numerous sports events are televised during a year and the number varies by year. For example, in 2015 there were 34 LPGA tournaments television, while the many NBA teams had different numbers of nationally television games from 2 to 32 depending on the team. To keep it manageable, you might limit your study to five sporting events a year for each sport for each gender. To pick the five events for a year, you might consider two at the start of the season—one in the middle and two at the end of each season. If you are curious about trends over time, you might pick sports during four seasons or years, five years apart, such as 2000, 2005, 2010, and 2015. This gives you 80 sporting events (2 genders 3 2 sports 3 5 sporting events 3 4years5 80 total sporting events). You won’t know the number of commercials (based on your definition) to review per selected sporting event until you review each. Let us say each event typically has 20 different commercials. This gives you 80 events multiplied by 20 or 1,600 commercials to consider. It would take a very long time to review and code all units, so you might want to use random sampling. You might stratify by the four event types (the two genders and two sports), and within each type make sure that you have each of the four selected years then take 25 percent of commercials. This would give you approximately 25 percent of 1,600 or 400 commercials to code.

WRITING PROMPT Content Analysis Sampling Describe a type of message/text you want to conduct a content analysis on, specify its boundaries, and estimate how many units there might be in the population from which you might draw a sample for a specific study. The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit Step 5: Create a Coding System Once you decide on variables to measure, operationalize them by creating a coding system. To create a coding system, carefully conceptualize each variable and decide whether you will use a manifest or latent coding method or both. For each variable, decide what to measure (a variable’s presence, direction, frequency, intensity, space or prominence, or all of them). Step 6: Construct and Refine Coding Categories The coding category is a critical aspect of the coding

system. It determines the number and types of distinctions to make within a variable and how to distinguish among them. Let us say you are measuring the variable “violence” in commentary on football games. You must decide the number of levels: Do you want high, medium, and low levels of intensity, or are there more levels? What distinguishes a low from a medium level? Do you want to consider different types of violence (e.g., physical, emotional, sexual) and look at the intensity of each? Making It Practical: Coding Systems and Categories In order to code a unit in content analysis,

you need to review your research question and identify variables. If you wanted to analyze differences in television commercials during male and female sporting events, you must decide what is relevant about content of the commercial. Perhaps you are interested in the type of product being advertised—alcohol, automobiles, clothing, insurance, other television programs, sports equipment, etc. You might consider the relative cost of the product and whether it is basic-consumer produce (e.g., beer or laundry soap) versus prestige product (e.g., $50,000 luxury brand SUV or Caribbean cruise) and if actors are present, their characteristics, such as gender, race-ethnicity, age, etc. While you can change categories during the coding process (and return to re-code the first units you began with), if you fail to code for a variable and code consistently, you will not clearly document what is in the text. When planning a research project, you should calculate the time required to complete the research. For example, during a pilot test, you may learn that it takes an average of four minutes to view a commercial and code all

Research with Nonreactive Measures 169

the variables in it. This does not include the time for locating the commercials. If you have approximately 400 commercials, at 4 minutes coding time for each, it will require of 27 hours of solid coding—not counting breaks or time to verify the accuracy of your coding. At nine hours per week, coding alone will require about three weeks.

WRITING PROMPT Coding in Content Analysis Imagine that you want to create a coding system for a collection of 30 television commercials about standard motor vehicles (e.g., cars, pick-up trucks, or SUVs). Pick three variables about the commercials to a measure in the commercials. How might you go about creating a coding system for the three variables that you could give to others and have them use? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit Step 7: Code the Data onto Recording Sheets

Once you decide the variables, type of coding, and coding categories, you should construct a sheet (paper- or computer-based) on which to record information. Typically, you will have one recording sheet (piece of paper or word processing page) for each unit or case. If you planned to

code 400 television commercials and the commercial was your unit of analysis, you would need 400 sheets, one for each. Put general information about the study (title, who does the coding) on each sheet. Also have space for basic information about the unit (the date and time) and space for each variable and its categories. You then code the text by filling in the information on each unit. Place them into spaces on the recording sheet. Step 8: Data Analysis Content data analysis is like

other quantitative data analysis. You must transfer the data from the recording sheets into a machine-readable form that computer programs use. Typically, this is a grid or spreadsheet. Each row is a unit or case, and the columns represent variables. You can then use standard analysis techniques for quantitative data (see Figures 8.2 and 8.3).

8.3.5: Limitations of Content Analysis Generalizations in content analysis are limited to the cultural communication itself. Content analysis only describes what is in the text and reveals patterns within it. For example, content analysis allows you to say that a certain type of message appears regularly on television, but not to say how that message influences the attitudes or preferences of people who view it.

Figure 8.2 Example of Blank Coding Sheet Empty Coding Sheet for Study of TV Commercials and Sports Programs Professor Neuman, Sociology-Anthropology Department TV Commercials during Men’s/Women’s Basketball and Golf Project Commercial # Length (seconds)

Coder:

Sport Event:

NBA PGA WNBA PGA

Televised Date:

Product Type:

Food/Drink: Beer Soft drink Restaurant: Fast food Pizza House-hold: Garden/lawn Paint Appliances Carpet Sports Products: Equipment Electronics: Cell phone Batteries Health/beauty: Haircare Colds Pharmaceuticals Makeup/shaving Nonprescription medication/vitamins Transportation: Auto Truck Recreation: Resort Cruise Service: Finance/insurance/investment Medical/dental Exercise/gym Education U.S. military Legal services Real estate Children’s Products: Toys Food

Snack Breakfast Other Furniture Cleaning/laundry Shoes Clothing TV/computer Other Personal hygiene/deodorant Toothpaste/mouthwash Weight loss/nutrition Airline travel Music event Sports Internet/cell phone/cable TV Auto repair Hair Hotel Car rental Home repair/cleaning Hygiene/health

Product Cost:

Under $20 $20–$100 $5,001–$40,000 Over $40,000

$101–$1,000 $1,001–$5,000

Actors Shown:

None All male All female All under 30 All 30–50 All 50+ years old Mixed ages

Mixed gender All whites Mixed races

170 Chapter 8

Figure 8.3 Example of Completed Coding Sheet Coding Sheet for Study of TV Commercials and Sports Programs with Data Professor Neuman, Sociology-Anthropology Department TV Commercials during Men’s/Women’s Basketball and Golf Project Commercial # 121 Length 30 (seconds)

Coder: Susan J.

Sport event:

NBA__X_PGA WNBA PGA

Televised Date: 3/22/2010

Product Type:

Food/Drink: X Beer Soft drink Restaurant: Fast food Pizza Household: Garden/lawn Paint Appliances Carpet Sports Products: Equipment Electronics: Cell phone Batteries Health/beauty: Haircare Colds Pharmaceuticals Makeup/shaving Nonprescription medication/vitamins Transportation: Auto Truck Gifts: Flowers Jewelry Recreation: Resort Cruise Service: Finance/insurance/investment Medical/dental Exercise/gym Education U.S. military Legal services Real estate Children’s Products: Toys Food

Snack Breakfast Other Furniture Cleaning/laundry Shoes Clothing TV/computer Other Personal hygiene/deodorant Toothpaste/mouthwash Weight loss/nutrition Airline travel Perfume Candy Music event Sports Internet/cell phone/cable TV Auto repair Hair Hotel Car rental Home repair/cleaning Hygiene/health

Product Cost:

X Under $20 $20–$100 $5,000–$40,000 Over $40,000

$101–$1,000 $1–5,000

Actors Shown:

None X All male All female X All under 40 All 50+ years old

Mixed gender All whites X Mixed races

Unfortunately, by itself content analysis cannot do any of the following: • Determine the truthfulness of an assertion or statement. • Evaluate the aesthetic qualities of literature or visual material. • Interpret the content’s significance or importance to shape events. • Reveal the conscious intentions of the organizations or people who created the text. • Determine the influence of a message on its receivers or power to alter their thinking. In order to understand fully the process of how messages influence people’s beliefs and behaviors, you need to combine content research with other types of studies, such as experiments. You can then see both how widespread a message appears and how it affects message receivers. For example, your content analysis shows that children’s books contain gender stereotypes. Alone, that does not mean that such stereotypes influence children’s beliefs or behaviors to follow those stereotypes. To determine whether repeated exposure to stereotypes affects children, you need to conduct a separate study on how children’s beliefs and behaviors are influenced by what they read. By

putting the two types of research together, you can develop a complete picture of patterns present in the text and how the text influences children.

8.4: How Can You Mine Existing Statistical Sources to Answer New Questions? 8.4 List the process steps of existing statistics research With many research techniques, such as the survey, experiment, or content analysis, you both create a study design and collect original data. However, you are lucky to have mountains of already collected and publicly available data about the social world. Some of it is in statistical documents, other quantitative information is in computerized data files. In either case, you can search through large collections of information with research questions and possible variables in mind. Once you locate relevant data, you can then statistically analyze the information to address the research questions. Existing statistics research involves analyzing previously collected public data to answer new research

Research with Nonreactive Measures 171

uestions. A big difference from most other research techq niques is that you devote time to learning about and locating the data that are available before you develop a research question and hypothesis. Recall that in other forms of quantitative research, the process was to begin with ideas or concepts, conceptualize them into variables with definitions, develop a research question and hypotheses, operationalize variables into specific measures, gather data, and then analyze the data. In short, for the other research techniques, you start with a research question or hypothesis. For existing statistics studies, you start by learning what data are available. The process goes as follows: 1. Search and scan existing statistical information or data with general ideas and interests.

Table 8.1 Examples of Publicly Available Social Indicators See Land, Michalos, and Sirgy (2012) for a more complete list. Turnout to vote in elections Number of hours that people volunteer per year Percentage of the population that is literate Percentage of the population that lacks health care coverage Number of child abuse cases Average length of time people commute to work Number and size of parks and recreation areas Number of crimes reported to police

Making It Practical: Walkability as the New Social Indicator

2. Conceptualize the data you located into specific variables. 3. Identify variables that have the same unit of analysis, examine details of data collect, and verify data accuracy. 4. Organize independent and dependent variables into hypotheses and a research question. 5. Test the hypotheses using statistical data analysis. An experiment is best for topics in which you can control a situation and manipulate an independent variable. Survey research is best for topics in which you can ask questions and learn about reported attitudes or behavior. Content analysis is best for topics in which you look at the content in a communication medium. The best topics for existing statistics research are ones on which large bureaucratic organizations routinely collect and report quantitative information; many measures in such reports are social indicators. During the 1960s, social scientists who conducted studies in various fields noted that the systematic quantitative information available to public decision makers was limited to a few economic measures. They initiated the “social indicators movement” and developed many new measures, or indicators, on a broad range social conditions or forms of well-being. Their goal was to combine data on social conditions with economic indicators (e.g., gross national product, income) to create a more complete picture of social-economic conditions. They created a separate academic organization and journal devoted to social indicators. Soon environmental impact studies included social indicators with other measures, such as clean drinking water, air pollution, and traffic congestion. Social indicators can measure negative aspects of social life (the death rate of infants during the first year of life, child abuse, crime rate, divorce rate, alcoholism, teen suicide) or positive aspects (job satisfaction, volunteering activity, park land, homeownership, housing with indoor plumbing, nearby medical facilities) (see Table 8.1).

Over the past two decades, thinking in the public health and urban planning fields converged. Public health officials were concerned about increasing rates of obesity and people who rarely exercised. They encouraged people to engage in simple daily physical exercise, such as walking. Urban planners became aware of problems arising from an auto-dependent “build environment” of spread out housing and neighborhood design that were prevalent in the 1970s to 1990s. The design required people to use autos constantly while discouraging walking or bicycling. Academics and practitioners developed the idea of “walkability” to describe the degree to which the physical layout of a neighborhood or location was “walking friendly.” They created many scores or indexes of walkability looking at avail-

172 Chapter 8 ability of sidewalks, major highways to cross, distance from shopping or schools, and so forth. One study (Maghelal and Capp, 2011) found 25 indexes to measure walkability. People created score using a Geographic Positioning System (GPS) analysis of maps with Geographic Information System (GIS), surveys of residents, and observations. In 2007, a company called Walk Score created a walkable measure of a house or apartment building location for the real estate industry. Now a “walk score” from 0 to 100 is a social indicator available for many specific locations in real estate listings (see Vanderbilt, 2012). You can enter an address on a website and get its “walk score.” There are also websites for people looking for place to live that offer walkability scores for entire cities and towns. Some cities are rated as very walkable (e.g., Boston, MA; San Francisco, CA; Washington, DC), while others were cardependent (e.g., Austin, TX; Charlotte, NC; Indianapolis, IN). Researchers compared several ways to measure walkability and explored relationships between the social indicator of “walkability” with health outcomes (e.g., obesity, diabetes, heart disease, and mental depression) and social indicators such as life satisfaction, social interaction, and crime rates. By getting the addresses of many people and checking walkability scores, you can see whether a location’s walkability is associated with other variables (e.g., health outcome). Researchers conducted studies using walkability scores in Australia, New Zealand, Japan, North America, and most European countries. Hundreds of public or private organizations have ongoing data collection and reporting activities for internal policy decisions or as a public service. They rarely collect data to address one specific research question. Existing statistics research is appropriate when you have an issue or question about the various social, economic, and political conditions on which organizations regularly gather and report information. Often organizations collect the data over time or across wide geographic areas. The information is often free or nearly free. For example, free, public existing statistics allow you to determine whether unemployment and crime rates are associated in 150 cities across a 20-year period. Some of the initial data collection may have been reactive. For example, a measure of the unemployment rate comes from a survey that asks people whether they are looking for work. Other data collection is without reactive effects, such as recording how many people voted in an election or the number of students who received high school diplomas. Organizations collect the data as part of their routine bureaucratic planning and monitoring activities, not for research. Your use of the data to answer research questions is nonreactive. People on whom the information is gathered are unaware of its research use. You will face three challenges in doing existing statistics research: searching and locating data sources, verifying data quality, and being creative in thinking about how to turn the data into variables that can answer research questions.

WRITING PROMPT Social Indicators Identify three of the most important social indicators for a community’s well-being, in your opinion, and explain why they areimportant. The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

8.4.1: Locating Data Government agencies, private companies, research centers, and nonprofit organizations gather an enormous volume and variety of quantitative information. The quantity is overwhelming and topic choice is vast. If you plan to conduct an existing statistics study, discuss your interests with an information professional—in this case, a reference librarian, who can point you in the direction of possible sources. Many sources are “free”—that is, publicly available at libraries or over the Internet. Nonetheless, it can take a huge amount of time and effort to locate information. There is a paradox: You do not know what information you will find until you look for it, and you do not know what to look for until you begin to search. Professional researchers can spend over a hundred hours searching in libraries, searching on the Internet, or contacting specific organizations with requests for information. Once you locate information, you need to record it. Some information is already available in a computer-readable format. For example, instead of recording voting data from published books, it is already in a format that computers read (e.g., spreadsheets, statistics programs). The single most valuable source of statistical information about the United States is the Statistical Abstract of the United States. The U.S. government has published it annually since 1878. It was available on the Internet until 2012, when, after a federal government cost-cutting move, a private company took over publishing it and free public access was reduced. It is still available in libraries. It has 1,400 charts, tables, and statistical lists and detailed statistics from hundreds of government and private agencies. It is hard to grasp all that the Statistical Abstract contains until you skim through it. Most information is organized by state or county as the unit of analysis, and often it goes back many years. Most national and state governments publish similar statistical yearbooks. If the country is your unit of analysis, the United Nations and international agencies (the World Bank, Organization for Economic Cooperation and Development) have their own statistical publications with information on countries (e.g., literacy rates, percentage of the labor force in agriculture, birth rates).

Research with Nonreactive Measures 173

Example Study Existing Statistics: Becoming a“Wet” or “Dry” County You might have heard about the great U.S. experiment with Prohibition, 1920 to 1933. During those years, it was illegal nationwide to produce, transport, or sell alcohol. After national prohibition ended, 33 states gave county-level governments the option of being “wet” (i.e., allowing alcohol sales) or “dry” (i.e., banning alcohol sales). As of 1970, 582 counties were dry (defined as banning both the retail sale of alcoholic beverages and the sale of alcohol by the drink). The number of dry counties declined to 262 by 2008. Over the years, a little over 19 percent of the local option counties changed their prohibition status; 50 countries shifted from wet to dry, and 363 changed from dry to wet. Frendreis and Tatalovich (2013) used existing statistical data to evaluate three hypotheses about why a country changed its prohibition status. They obtained data for the dependent variable, i.e., sales of alcoholic beverages in a county, from publicly available data in several studies. One hypothesis is that general societal modernization causes a shift away from traditional (dry) values. Modernization is a value change associated with people becoming more educated, earning higher incomes, and living in urban areas. The authors obtained data from U.S. government statistics on the percentage of people in a country with a college education, average incomes, percentage living in urban areas, and population density. A second hypothesis is that secularization, or a decline in traditional religious authority, causes a change. The authors used two publicly available data sets from nonprofit research organizations on the mix of religious denominations and church memberships in a country. An increased mix of religious diversity, decline in church membership, and decline in Evangelical dominance indicated greater secularization. A third hypothesis is that prohibition status changes due to population growth and “turnover” in the form of population growth or decline, birth and death rates, internal migration (residents of one state moving to another state), and immigration from other countries. The authors obtained country-level population data from the U.S. Census Bureau.

Read More After assembling the data, the authors conducted statistical analyses, using the county as a unit of analysis, to determine which hypotheses best predicted a change in prohibition status between 1970 and 2008. The authors saw the three hypotheses as being complementary rather than competing with one another. They found strongest support for the secularization and modernization hypotheses: Counties shifting to become “wet” showed larger increases in religious diversity during the 1970–2008 period, had faster growth in college graduates, and greater population growth. The factor most associated with a country moving in the opposite direction, from being “wet” to “dry,” was a high rate of Evangelical Protestant adherence in a county with low religious diversity, and

low rates of income growth and little in-migration. The authors noted that while religious values do not automatically change public policy, when a particular religious culture (Evangelical Protestantism) is dominant in a locale, it creates a cultural climate that pushes public policy in the direction of prohibition, especially in locales that lack religious diversity.

8.4.2: Verifying Data Quality Despite its availability, six issues can limit the use of existing statistics: • Missing data • Reliability • Validity • Topic knowledge • Fallacy of misplaced concreteness • Ecological fallacy Missing Data Missing data can be a major limitation with existing statistics and documents. Sometimes the data were collected but have been lost. More frequently, no one collected the data you seek. Officials decide whether to collect official information, and political pressures often influence priorities about what data are gathered. Government agencies start or stop collecting data for political, budgetary, or other reasons. For example, the U.S. federal government stopped gathering data on issues that social researchers had found highly valuable, number of work-related injuries. The official reason given for no longer collecting information was cost cutting, but later officials revealed it was really to promote a particular political agenda. If the government no longer gathers information documenting economic inequalities, environmental conditions, or health risks, people will find it more difficult to make claims of discrimination, environmental destruction, or unsafe conditions. Missing information is especially a problem when you want to show trends over a several year period. A related type of missing data is data collected with less refined categories than you want. Perhaps you are interested in racial-ethnic groups. You want to compare the unemployment rates of whites, blacks, Native Americans, Asians, and Latinos in the United States, but the official data have white and nonwhite as the variable categories. Reliability Reliability problems plague many existing

statistics sources. Reliability problems develop when official definitions or methods of collecting information vary over time. For example, the official definitions of work injury, disability, unemployment, and the like have changed several times. Even if you learn of the changes, consistent measurement may be impossible.

174 Chapter 8 For example, during the early 1980s, the U.S. government changed how it calculated the U.S. unemployment rate. Until then, government agencies calculated it as the number of unemployed people divided by the total number of people in the civilian work force. The new method divided the number of unemployed by the civilian work force, plus all people in the military. A similar complication happens when police departments computerize their records. The number of crimes reported in statistical reports may decrease or increase not because actual criminal acts change, but due to changes in recordkeeping.

tion are not always perfect and may introduce errors of which you are unaware (e.g., census employees who avoid poor neighborhoods, or people who put a false age on a driver’s license). People who collect data that end up as official statistics may make errors in organizing and reporting (e.g., a police department that is sloppy about filing crime reports and loses some). Errors can also occur in publishing information (e.g., a typographical error in a table).

Validity You can encounter three kinds of measure-

An error in official statistics data on the number of people permanently laid off from their jobs illustrates the problem. Official U.S. Bureau of Labor Statistics data on permanent job losses come from a survey of 50,000 randomly selected people. The agency counts the percentage saying they were laid off. However, a mistake was made. The percentage was based on the number of questionnaires sent out and unadjusted for people who did not respond. Here is an example of what happened. The agency sent 50,000 people questionnaires. In 1993, 8,000 returned the questionnaire reporting that they had been laid off. This was 16 percent of the 50,000 total questionnaires and reported as the laid-off rate. However, 40,000 people returned their questionnaires making the true rate 20 percent (8,000/40,000). In 1996, 4,500 returned the questionnaire saying that they had been laid off of work. This is 9 percent of the 50,000 sent out. But only 22,500 people returned the questionnaire making the true rate 20 percent (4,500/22,500). The agency mistakenly reported a 7percent decline in laid-off people between 1993 and 1996 when in reality, there was no change. Only a university researcher’s careful detective work uncovered the error (see Stevenson, 1996).

ment validity problems in existing statistics research. First, the agency or organization that collects information uses a different conceptual definition from yours. For example, you define a work injury as including minor cuts, bruises, and sprains that occur while working on a job. The official definition includes major injuries that required a visit to a physician or hospital. Many injuries you define as work injuries would not be in the official statistics. Another example is that you define a person as being unemployed if he or she would work if a job for which he or she is trained is not available, if the person is forced to work part time but wants to work full-time, or if after a year of trying the person has given up looking for work. The official definition includes only people now actively seeking any full- or part-time work at any job. Excluded are people who stopped looking, who work only 15 hours a week because they cannot find more hours, or who are not looking because there are no jobs for someone with their special training (e.g., a dentist can only find a job driving a taxi). To add complications, different countries or states often use different official definitions. A second validity problem can occur when you use official statistics as a proxy for a variable of interest because they are all you have, so you use them. Let us say you want to know how many people were robbed in Philadelphia last year. You use police statistics on robbery arrests as a proxy. However, many robberies go unreported. If half of robberies are not reported to the police, you are not measuring the true number of people robbed by using official statistics. You are measuring only half of all robberies, the number of robberies that are known to police. Such reporting bias can affect hypotheses testing. Let us say people under 35 years old report a robbery to police far less than people over 35 do. Using official data, you find that most robbery victims are older than 35; however, this may not be true but is caused by a reporting bias in the official statistics. A third validity problem arises because you depend on others to collect data. The people who collect informa-

How do errors occur in official statistics? Compare Your Thoughts

Making It Practical: Official Unemployment Rates versus the Nonemployed In most coun-

tries, the official unemployment rate measures the unemployed as a percentage of all working people. It excludes two related categories of not-fully-employed people: involuntary part-time workers and discouraged workers. In some countries (e.g., Sweden and the United States), the unemployment rate would double if these people were included. Most official statistics exclude other nonworking people such as transitional self-employed and the underemployed. Different definitions treat the situation of unemployment differently. An economic policy or labor market perspective sees the unemployment rate as measuring people ready to enter the labor market immediately. It sees nonworking people as a supply of labor that is available to employers, or as an input to the economy. A social policy or human resource perspective sees the unemployment rate as measuring people not cur-

Research with Nonreactive Measures 175

rently working to their full potential. The rate tells us who cannot fully utilize their talents, skills, or time. Nonworking people viewed individuals who are unable to be productive, contributing members of society. Thus, a nation’s official statistics will include a definition that reflects a particular value perspective or theory of what unemployment means for society. Consider the following list of all nonemployed people: Categories of Nonemployed/Fully Utilized • Officially unemployed. People who meet three conditions: (1) lack a paid job outside the home, (2) are actively engaged in looking for work, and (3) can begin work immediately if it is offered. • Involuntary employed. People with a job, but who work irregularly or many fewer hours than they are able or willing to, or who are working part time but desire and are able to work full time. • Discouraged workers. People who are able to work and who had actively sought work for some time, but being unable to find work have stopped looking. • Other nonworking. People not working because they are retired, on vacation, temporarily laid off, semidisabled, homemakers, full-time students, or in the process of moving. • Transitional workers. Self-employed people who are not working full time because they are just starting a business or are going through bankruptcy of a failed business. • Underemployed people. Persons with a temporary fulltime job for which they are seriously overqualified. They seek a permanent job in which they can fully apply their skills and experience. [For information on measuring unemployment see National Center for Education Statistics (n.d.) and Sorrentino (2000)].

WRITING PROMPT The Unemployment Rate The unemployment rate is a widely reported social indicator, yet it is very controversial. Different people propose many different ways to conceptualize and measure it, and it is defined and measured in very different ways from country to country. Provide your own definition of someone who is “unemployed.” Describe how you would go about measuring unemployment based on your definition. The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

Topic Knowledge Because existing statistical data

are easily accessible, you may be able to get a lot of data

on an issue that you know little about. This can lead to erroneous assumptions or false interpretations of results. Before using any data, you should become informed about the topic. You may have a statistic, but do you know what it really means? For example, the widely quote “Consumer Price Index” (CPI) is based on a sample of 200 categories of goods of the typical consumer, including food, housing, transportation, education, sales taxes, health care, and even pets. It constantly changes. It does not include buying life insurance or savings. If the CPI tried to maintain a fixed sample of products, the sample quickly would shrink and become unrepresentative of what consumers were purchasing. Each time that an item in the CPI sample permanently disappears, another replaces it. At one time cell phones, cable TV, and computers were not part of the mix; they replaced old products like landline phones services and other home entertainment (large stereo systems). Also, when the price of one product (e.g., steak) goes up, the index does not assume people will switch to another (e.g., hamburger). This complex issue is far greater when using international statistics because the definitions and situations in different countries can vary widely. Fallacy of Misplaced Concreteness The

f allacy of misplaced concreteness occurs when you quote statistics in excessive detail to give an impression of scientific rigor and precision. For example, an existing statistics report says that in March, 2015, Australia had a population of 23,735,583. It is better to say that it is about 23.7 million, because the exact number is not that precise. The counting of people could easily be higher or lower by a few several thousand people. If you calculated the percentage of divorced people in a town as 15.655951, only report it to one decimal place, or 15.7 percent. The ecological fallacy happens when your research question is about a lower-smaller unit of analysis (such as individual behavior) and you only have data on a higherlarger unit of analysis (an entire state). Published data are often available on a higher-larger unit of analysis. If you are interested in variables on a much lower level than the available data, you may be seriously misled. For example, data may be available on the state level (such as unemployment rate by state), but you are interested in individuals (characteristics of individuals who become unemployed). From existing statistics, you find that states with high unemployment rates also have a high percentage of cigarette smokers. It is an ecological fallacy to say that unemployed people are more likely to smoke. Data for the state as a unit of analysis do not show which individuals smoke. Simply because more smokers are in a state that also has a high unemployment rate does not make the unemployed people more likely to smoke than employed people. If you want to know whether unem-

176 Chapter 8 ployed individuals are likely to smoke, you need data for which the individual is the unit of analysis. Official statistics may not have the data to match your research question, so you must change your question or switch to a different research technique, such as the social survey. You can ask individuals: Do you smoke? What is your employment status? Then see whether the two variables are associated among those individuals.

Tips for the Wise Consumer Using Data from Existing Statistical Sources When you use data from existing statistical documents, do the following: 1. Be certain that the definition and measure of a variable truly fit the variable of interest to you. 2. Watch out for missing data or variable categories that combine distinctions of interest to you. 3. Be aware of the units of analysis used in measures and avoid the ecological fallacy. 4. Investigate the topic area that applies to the data you are using. 5. Read statistical tables very carefully, including the details in footnotes and other explanations, so you know exactly what the table offers.

WRITING PROMPT Finding Statistics Online Mountains of existing statistical information on social, political, and economic indicators are available for free, some of it online, some ofit in printed form. List three best places on the web where you could locate legitimate existing statistical information on social, political, and economic indicators for free. Why did you choose these sites/resources? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

8.4.3: Thinking Creatively about Variables of Interest When you use existing statistical data, someone else already collected data that represent conceptual variables. Often the data do not match the exact variable of greatest interest to you. This means you must be creative to identify a surrogate or proxy that is closest to the meaning of your conceptual variable. After searching through many existing statistical sources looking for data measures that

best match your concept, you need to select and combine measures. Perhaps your conceptual variable is “upperclass neighborhood” but no existing data source has such a measure. You nonetheless find data on various social characteristics by city block. If you are careful and creative, you can build a substitute measure of “upper-class neighborhood.” Let us say you know that upper-class people tend to own their own house that has a high value (over 1 million dollars), completed a college degree or more schooling, send their children to expensive private schools, and belong to exclusive country clubs. You find data by city block on the percentage of owner-occupied houses of a high value and the percentage of residents with a college degree or more schooling. You also obtain address lists of student at expensive private schools and for members of exclusive private clubs, then match the addresses to city block. Using the four data sources, you identity 15 city blocks in which 70 percent houses are owner-occupied and worth over 1 million dollars, and 85 percent residents over 25 years of age have at least a college degree, and have the addresses if at least 15 children attending expensive private schools and at least 12 exclusive private club members. By merging the four existing statistics data measures, you have an indicator of an “upper-class neighborhood.”

8.4.4: Standardizing the Data To make valid comparisons, most variables must be standardized and turned into rates. Standardization lets you compare on a common base by adjusting or removing the effect of relevant but different characteristics making important differences visible. Standardization involves selecting a base and dividing a raw measure by it. The base is simply a number or characteristic that influences how you interpret raw measures. Most, but not all, existing statistics information is standardized. At times, you must standardize it before you can use the information. Without standardization, it is impossible to compare and very easy to misinterpret the information. Population size is an example. It is relevant to the number of people who vote. Let us say you are interested in citizen participation in elections in various states. You find from Federal Elections Commission that 13,038,547 people in California and 713,180 people in Maine voted in the 2012 presidential election. Does the bigger number mean Californians are more politically active than people in Maine? Compare Your Thoughts If California and Maine had the exact same number of people, you would not need to standardize. You might realize

Research with Nonreactive Measures 177

that California has more people (population was 38 million in 2012) than Maine (1.3 million people in 2012), so California should have more voters. To compare the voting rate or percentage of voters in the two states, you must first standardize the data. You can remove the size difference in the two states to make visible the true differences in voting rates. The percentage is a common form of standardization; it standardizes on base 100. We can divide the number of voters in each state by the state’s population size. If 13 million out of 38 million voted in California, about 34 percent voted. If roughly 731,000 out of 1.3 million people in Maine voted, then about 56 percent voted. Of course, California’s population size means it had a bigger impact on the national election’s outcome, but if your interest is in active participation in elections, then you need to adjust for different state population sizes. But wait, not everyone can vote! Recent immigrants, noncitizens, and people under the age of 18 cannot vote but are in the state population; therefore, the entire state population is not the best base to use. A critical question in standardization is deciding what base to use. The choice is not always obvious. It depends on the variable, and you need to think about it. To get an accurate measure of the percentage of people who voted, we need to divide number of voters by the number eligible voters. Political scientists calculate voter turnout based on the voter age population (VAP) and the voter eligible population (VEP), removing prisoner, noncitizens, etc., who cannot vote. In 2012, California had 29 million VAP and 23.7 million VEP. After calculating voting, California had a 55.1% VAP turnout and 45.1% VEP turnout while Maine had a 68.2% VEP and 67.0% VAP turnout. So even after proper standardization, turnout was higher in Maine than California. Different bases can produce different rates, and how you define a variable changes the base. For example, some define the unemployment rate as the number of people in the work force who are out of work. The overall unemployment rate is given by the following fraction: Number of unemployed people Unemployment rate = Total number of people working You can also divide the total population into subgroups to get rates for subgroups in the population, such as white males, Asian American females, African American males between the ages of 18 and 28, or people with college degrees. Rates for subgroups are often relevant to a research question. Perhaps you conceptualize unemployment as an experience that affects an entire household. You can change the base to households, not individuals. The formula for an unemployment rate will look like this: Household Unemployment rate = Number of households with at least one unemployed person Total number of households

How you conceptualize a variable suggests different ways to standardize, but failing to standardize can produce distorted results. A few years ago, a student announced to me that based on the Statistical Abstract of the United States, New York had a terrible child abuse record with over 74,000 child victims, whereas her home state of Utah had only about 13,500. She failed to standardize on population size. Once she adjusted for the number of children in each state, New York had 16.3 victims of child abuse per 1,000 children whereas Utah had 18.3 per 1,000. Thus, the situation actually looked worse for Utah. Of course, the detection and followup of abuse cases may not be the same across the states, but the importance of standardizing data was clear.

WRITING PROMPT Standardization Standardization is needed to compare, but many people including journalists cite numbers without necessary adjustment. In September 2015, Seattle reported 10,004 homeless people, Honolulu said it hadthe highest for a small city at 4,900 and declared a state of emergency, while New York City reported 59,305. How might you standardize the homeless numbers to compare and find out which city had the greatest homeless problem relative to its size? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

8.5: Using Secondary Sources to Answer Research Questions 8.5 Compare secondary data analysis with other forms of statistics research Secondary data analysis is similar to existing statistics research in that you analyze data that someone else collected. The main difference is official statistical information is often released in aggregate form, for large, macro units of analysis, or in descriptive tables that do not have all variables on all units. For example, official statistics will give you the percentage of doctors in a state affiliated with hospitals and the percentage of hospitals with certain medical technology. You will not find out the number of doctors affiliated with each separate hospital and what equipment each hospital has. If you are interested in whether the number of doctors at a hospital is associated with it having certain medical equipment, you need data in which the unit of analysis is the hospital, but this may not be in any of the official statistics available to the public. Most secondary data are available as computer-readable data files with a

178 Chapter 8 disaggregated unit of analysis and many variables. The majority of secondary data was originally collected as a random sample survey. In contrast to primary data research where data collection is emphasized, in secondary data analysis the focus is on statistically analyzing data collected by others. Secondary data analysis facilitates replication and permits asking research questions that the original researchers did not consider. In a way, it is similar to entering the research process after the data collection phase has been completed. Collecting data on a large scale (such as nationally) is very expensive and difficult. A major national survey with rigorous technical support can cost tens of thousands of dollars. Fortunately, several organizations gather, preserve, and share survey data or other types of data. They make their data available for others to analyze. The most widely used source of survey data in the United States is the General Social Survey (GSS). The U.S. government pays for the social survey, and the National Opinion Research Center at the University of Chicago has conducted it nearly annually for over 30 years. Data from it are made available to the public for secondary analysis. Many other nations have started similar surveys, and comparative survey data for multiple years are available in the Eurobarometer, the World Values Survey, Asian Barometer, and International Social Survey Program. The datasets are available to researchers for secondary analysis at no or very low cost. What is the General Social Survey? Compare Your Thoughts The General Social Survey (GSS) is the best-known source of survey data for secondary analysis in the United States. It is available in computer-readable formats. Neither the datasets nor codebooks are copyrighted. Users can copy or disseminate them without obtaining permission. Results using the GSS appear in over 2,000 research articles and books. The National Opinion Research Center (NORC) has conducted the GSS almost every year since 1972. Data come from face-to-face interviews of a random sample of 1,500 to 4,000 adult U.S. residents. The NORC staff carefully selects and trains its interviewers. The interviews are typically 90 minutes long and contain about 500 questions. The response rate has been 71 to 79 percent. Each year a team of researchers selects questions for inclusion. They repeat some questions and topics each year, include others on a four- to six-year cycle, and add other topics in specific years. You can learn more about the GSS by visiting the NORC website. You can order a copy of the GSS data or analyze it online. Online analysis of the data is available through several sources, such as Survey Data and Analysis (SDA) at the Computer-assisted Survey Methods Program of the University of California–Berkeley.

You will want to begin with a list of survey questions or variables that are available at several online sites. To analyze data in detail, you will need statistical analysis software and a raw data file. Several sources sell the data file at low cost, and they pre-format it for major statistics software packages.

8.5.1: Limitations of Secondary DataSources Using data collected by others is not trouble free. As with existing statistics research, you first must locate a data source and see what variables it contains. The most common limitation is that secondary data lack the variables you want for a research question. For example, in the GSS you are dependent on other researchers including survey questions on topics that you find interesting. Even when the data are available on variables of interest, the researchers who designed a study and collected the data may have conceptualized them differently. Before you proceed with secondary data analysis, you need to consider units in the data, the time and place of data collection, the sampling methods used, and the specific issues or topics covered in the data. For example, you may want to examine race/ethnic tensions between Latinos and Anglos in the Southwestern and Pacific regions of the United States. However, you find secondary data that include only the Pacific Northwest and New England states. You will have to reconsider the research question or use other data if they exist. Are there any limitations to using secondary data analysis? Compare Your Thoughts Secondary data analysis seems easier because the data are provided and you only need to know how to use and interpret statistical software programs. However, the other limitations common for existing statistics research, such as validity and topic knowledge, also apply in secondary data analysis. As with existing statistics research, you first need to examine the data and then develop hypotheses and a research question. Let us say you obtain a copy of the GSS. You may discover three limitations to its use: 1. Data are on all adults across the United States and not on specific types of individuals and geographic locations that you need for a research question. Perhaps you are interested in how 18to 22-year-olds in your state feel about an issue, but the GSS does not allow you to examine this topic. 2. GSS survey questions are not on issues in your research question. Perhaps you are interested in a specific issue, such as whether a person engaged in sexual activity with his or her spouse before marriage, but the GSS did not include such questions.

Research with Nonreactive Measures 179 3. GSS questions are worded differently than you wish or have different answer choices. For example, a long-used GSS question about prayer in public schools asks about whether respondents agree with a U.S. Supreme Court decision banning it. You cannot really learn people’s opinions about favoring or opposing specific types of prayer or religious activities in a school. If the GSS offers a two-choice answer, support/oppose abortion under certain conditions, but you want a wide range of choices (strongly support to strongly oppose), you will be restricted.

Example Study Secondary Data and the Immigration-Welfare Question The ISSP (International Social Survey Program) is a major source of secondary survey data publicly available for comparative research. The ISSP grew out of a collaboration between U.S. researchers who oversaw the GSS and German survey researchers in 1982. British and Australian researchers joined in 1984. Since 1984, ISSP expanded to include 53 nations. Like the GSS, the ISSP has a set of core questions included every year, and specific topic questions asked periodically. Not all of the 53 nations participate every year. For example, in 2012 the survey topic was family and gender roles, and 37 countries participated. It was the fourth time for the topic. The first time was in 1988 when only eight countries participated. The ISSP organization coordinates survey questions and methodology, and arranges data archiving and sharing. Data from 1985 to 2011 are housed at archives in Britain, Spain, and the United States. So far, over 5,000 articles and research reports have used ISSP data. Brady and Finnigan (2014) examined ISSP archives and selected years when “role of government” was the topic. They used 1996 and 2006 ISSP data for 17 countries to learn whether public support for social programs in a country is associated with its level of immigration. They noted that migration to high-income countries has surged. The foreign-born percentage of the population more than doubled from 1970s to 2005, and no affluent democracy experienced a decline in the percentage of foreign born.

Read More Scholars and policy makers thought that immigration would reduce public support for social policies, reasoning that as a society became more diverse and less ethnically hom*ogenous, support for public social services (e.g., welfare, housing aid, health care) would decline. Past studies showed that opposition to welfare in the United States is greatest in states with a high level of prejudice and a large proportion of blacks, and studies from several countries found increased antiimmigrant attitudes and the size of migrant populations grew. A simple “reduction hypothesis” states that decreased ethnic hom*ogeneity and fear of competition from recent arrivals will lead native-born people to want a freeze or reduction in social benefit programs. The researchers contrasted it with a “compensation hypothesis.” It posits that more immigration increases public support strong welfare programs than will compensate for and protect current the native-born from economic competition and insecurity resulting from high rates of immigration. They also considered a third “chauvinism hypothesis.” It predicts that greater immigration only reduces support for only those social policies that directly benefit immigrants. The researchers analyzed ISSP data for nationally representative samples from 17 high-income democracies, with over 17,000 respondents. Their dependent variable was a set of attitude questions about six social policies—providing jobs, housing, health-care, pensions, housing, and income assistance. They stated, “These multiple welfare attitudes also advantageously force respondents to think about specific social policies instead of a general sentiment about welfare, which likely has different cultural connotations crossnationally” (p. 24). In addition to looking national-level variables (e.g., generosity of social programs, rate of immigration into a country), they also considered individual-level variables (e.g., a respondent’s gender, education, age, and employment) in the statistical analysis. The researchers found that increases in percentage of foreign born and greater immigration did not reduce public support for social policies in a country. The major exception was that immigration reduced support for the idea that the government “should provide a job for everyone who wants one.” In general, they found greatest support for compensation and chauvinism hypotheses.

Summary Review Table 8.2 Review Strengths and Limitations of Nonreactive Research Nonreactive Research

Major Strengths

Major Limitations

Physical Evidence

Indirect and unobtrusive evidence

Must infer to people’s intentions

Content Analysis

Reveal hidden content in communication

Patterns in text, not its effects on people

Existing Statistics

Quantitative data of wide scope, low cost

Measurement reliability and validity

Secondary Analysis

Large-scale survey data, low cost

Limited variables are available

180 Chapter 8

8.6: How Do You Conduct Ethical Nonreactive Research? 8.6 Recognize that the primary ethical concern of physical evidence analysis is to protect people’s privacy and the confidentiality of data Ethical concerns are not at the forefront of most nonreactive research because the people who you study are not directly involved. In unobtrusive research, content analysis, existing statistic, and secondary analysis, few ethical concerns arise. The primary ethical concern of physical evidence analysis is to protect people’s privacy and the confidentiality of data. The use of official existing statistics raises other issues. They are social and political products. Implicit theories and value assumptions guide which information researchers collect and how they collect it. Measures or statistics that agencies define as being official and collected regularly can be involved in political disputes. If measuring a social condition in one way is official and in another way is not, measuring it in the official way may benefit certain political values over others. What information is gathered and made public also shapes public policy decisions. For example, political activity pressured government agencies to collect information on certain social conditions (e.g., the number of patients who died while in public mental hospitals). Government officials previously did not think the condition was sufficiently important to warrant public attention, or preferred to keep it quiet and hidden. Likewise, information on the percentage of nonwhite students enrolled in U.S. schools at various ages only

became available in 1953. Nonwhite students attended schools before 1953, but court decisions and public interest in racial discrimination caused government agencies to begin to collect the data. Just as organized concern about an issue stimulates the collection of new official statistics, the collection of official statistics on an issue can stimulate public attention. For example, drunk driving became more of a public issue after government agencies began to collect and publish statistics on the number of automobile accidents and on whether alcohol was a factor in the accidents. Most official statistics are collected for top-down bureaucratic or administrative planning, not for a researcher’s purposes or for people who might question the goals of bureaucratic decision makers. A government agency may gather data on the number of tons of steel produced, miles of highway paved, and average number of people in a household. It might not gather information on drinking water quality, contamination at food processing plants, or stress-related job illness because someone decided such conditions are not important ones. Some officials see the gross national product (GNP) as a critical measure of societal progress. However, the GNP omits noneconomic aspects of social life (e.g., time spent playing with one’s children) and ignores some types of work (e.g., housework). To a large degree, official statistics reflect the outcome of political debates about what the public needs to know and the priorities of top officials in government. It is important to recognize that people in positions of power are making decisions about what information to collect and make public. By not collecting and making public certain information, they may be protecting their own position of power or advancing a specific social-political value position.

Summary: What You Learned about Research with Nonreactive Measures You learned about several types of nonreactive research techniques. You can use these techniques to measure or observe aspects of social life without affecting the people who you study. The techniques can produce numerical information that you can analyze to address research questions. You can use the techniques with other types of quantitative or qualitative social research to address a large number of issues. As with other quantitative data, you need to be concerned with data quality and measurement. It is easy to use available information from a previously conducted survey or government document, but this does not mean that it is the best measure of what really interests you. Another limitation of nonreactive research stems from the availability of

existing information. Existing statistics and secondary data analysis are low-cost research techniques. However, you lack control over or detailed knowledge of the data collection process. This can be a potential source of mistakes and errors, and you must be especially vigilant and cautious.

Quick Review How Can You Use Physical Evidence in Research? 1. Nonreactive or unobtrusive measures (i.e., measures that are not obtrusive or intrusive) are information people

Research with Nonreactive Measures 181 generate without being aware that researchers may use it in a study. 2. Researchers have been very creative in finding many kinds of physical evidence to create nonreactive measures of variables about human social behavior. 3. Nonreactive data can confirm or reveal aspects of people’s lives different from what we would learn from direct, reactive data. 4. Physical evidence measures are usually indirect indicators that require you to infer or make a cautious “educated guess” from the evidence to a person’s preferences, behaviors, or attitudes. 5. It is always best to collect other confirming evidence and to rule out alternative explanations when you rely on indirect indicators.

How Can You Find Content within Communication Messages? 1. Content analysis is the study of text, (i.e., any message that is in a written or visual communication medium) by using objective, systematic counting and recording to yield quantitative data. 2. Content analysis is especially useful in examining large quantities of text, studying text at a distance, and revealing elements within the text that may not be easily seen with causal observation. 3. A critical feature in content analysis is to create a coding system for recording information from the text. Coding can measure the direction, frequency, intensity, space, and prominence of features within the text. 4. The coding system must be tailored to the specific communication medium, and can use direct counting of specific words or images (manifest coding) or the indirect counting of themes (latent coding). In general, manifest coding has greater measurement reliability. 5. Because content analysis examines large volumes of text, random sampling is often used. 6. There are many possible units of analysis in content analysis, and the unit of analysis selected greatly influences sampling and coding processes. 7. A content analysis study can tell us what patterns are in the text but not what impact those patterns have on people. To understand the impact of receiving, reading, or viewing text, we need to supplement content analysis with other types of studies.

How Can You Mine Existing Statistical Sources to Answer New Questions? 1. In existing statistics research, you analyze previously collected public data to answer new research questions. Unlike other research techniques, you devote time to learning what data are available before you develop a research question and hypothesis.

2. In existing statistics studies, you first search existing statistical data with general ideas and interests, next conceptualize the data into specific variables and examine data details and accuracy, then organize variables into hypotheses and a research question, and finally test hypotheses using statistical data analysis. 3. Missing data can be a major limitation with existing statistics and documents. Sometimes the data were collected but have been lost. More frequently, no one collected the data you seek. 4. Reliability problems develop in existing statistics when official definitions or methods of collecting information vary over time. 5. Three measurement validity problems can arise with existing statistics research. First, the organization that collects information uses a different conceptual definition from yours; second, you use official statistics as a proxy because they are all you have but they are not what you really want; lastly, the data that others collected contain systematic errors. 6. Published existing statistics data are often available for a larger unit of analysis than you wish. If your interest is in variables on a much lower level than your data, you need to be very careful to avoid the ecological fallacy. 7. Published existing statistics data are often available in a raw, or unstandardized form. Often you must evaluate the data’s meaning and make adjustments with standardization before conducting a statistical analysis of the data.

Shared Writing: Nonreactive Researchand Big Data You might have heard about “Big Data.” It is a kind of nonreactive digital data that businesses and governments have collected, stored, and analyzed since about 2010. Massive amounts of digital data such as every website visited (number of hits), location from which all text messages originate (based of GPS, Global Positioning System), etc., are collected and then statistically analyzed to identify patterns of people’s behavior and preferences. With each web search, call, purchase or text you are leaving a “digital footprint” that someone somewhere can examine for some purpose. List all types of digital information do you leave behind that some company or government agency might use to target you or study your behavior? Choose one item from two of your classmates’ lists (that you did not include in your own list and explain how you could regulate nonreactive research on that type of digital data. A minimum number of characters is required to post and earn points. After posting, your response can be viewed by your class and instructor, and you can participate in the class discussion.

Post

0 characters | 140 minimum

Chapter 9

Making Sense of the Numbers

Learning Objectives 9.1 Identify the three steps to prepare data for

statistical analysis 9.2 Explain the different types of descriptive

statistics 9.3 List the various ways to describe numerical

data for one variable About half of U.S. marriages last 20 years duration, a percentage that has not changed since the 1970s. Today, few Americans object to a man and a woman (or same-sex intimate couples) living together before marriage. About one in three couples who marry first lived together. Living together, or cohabitation, increased from 3 percent in 1982 to 11 percent of the population in 2006–10 (Copen, Daniels, and Mosher, 2013). Almost half (41 percent) of all women aged 15–44 cohabited at some time. The percentage of first

182

9.4 Analyze the three techniques to help see

therelationship between two variables in bivariate statistics 9.5 Evaluate how statistical significance helps

determine whether data supports a hypothesis unions among American women aged 15–44 increasingly became cohabitation instead of marriage between 1995 and 2006–2010 (see Figure 9.1). As more people cohabited, they delayed the age of marriage. The median age of a first marriage rose by nearly six years, from 20.8 in 1970 to 26.6 in 2013 for women and 23.2 to 29.0 for men. An unmarried, oppositesex couple lives together for a mean duration of two years. Eventually about half of these couples marry, about

Making Sense of the Numbers 183

Figure 9.1 Type of First Unions among Women Aged

15–44: United States, 1995, 2002, and 2006–2010

SOURCE: Figure 1 in Copen, Daniels, and Mosher (2013)

compatible with computer programs. Only then can you use the programs to manipulate numerical data in ways that generate various statistical results. You then can examine the results and connect what they reveal to your hypotheses or research questions (see Figure 9.2).

Percent

Figure 9.2 A Flow Chart of Preparing Numerical Data

40 30

Preparation of Numerical Data

Data or numbers you gathered in a study

30 23

Create a codebook, which is a catalog of outlining how to enter numbers into a computer

10 0 No Union

Cohabitation 1995

2002

Marriage

2006–2010

40 percent split up, and the rest continue to live together. Several states (Florida, Michigan, Mississippi, North Carolina, North Dakota, Virginia, and West Virginia) have (unenforced) laws making male-female cohabitation illegal. Living together is popular in many but not all countries. In some countries, over 75 percent of people are cohabitating (such as Sweden), but in others under 4 percent live together (e.g., Japan). This brief look at cohabitation shows numbers and statistics that can help you see the trends in intimate social relationships over time. Most research reports with quantitative data use charts, graphs, tables, and statistics of various types. It is not wise to skip over this part of a report; their purpose is to give you a condensed picture of what is in the data. Graphs and tables can tell a vivid story, but you need to learn their language. Although computers can create many statistical results from data, you must interpret or give substantive meaning to the results to determine what answers they offer for a research question.

9.1: How Do You Prepare Data for Statistical Analysis? 9.1 Identify the three steps to prepare data for statistical analysis The raw information, or data, that you collected with a research technique (e.g., experiment, survey, content analysis) in the form of numbers represents values of variables that measure characteristics of participants, respondents, or other units. The numbers are not the end of a study; rather, they are the start of a process of connecting data to a hypothesis or research question. You usually must reorganize raw data into a format that is

Organize your data based on the codebook

Enter data into the computer

Check and clean the data

Analyze the data using charts and tables and statistics

9.1.1: Organize the Raw Data into aMachine-Readable Format The computer programs that perform statistical analysis, or statistical software, enable us to create charts, graphs, tables, and other results that summarize and reveal features of quantitative data in ways that facilitate connecting the data with hypotheses. If you do secondary data analysis, the data often arrive “prepackaged” in a format designed for the statistics software. This allows skipping over a step of data preparation. When you gather data yourself or have “raw” data in the form initially collected (as mass of raw numbers), you first need to convert the raw data into a computer-friendly format, and engage in data coding. As with the coding process in content analysis, data coding for statistics software requires you to create and consistently apply explicit rules to transfer information from one form into another. In this case, you transfer the raw numbers you have collected into a format that statistical software recognizes. The many specific kinds of statistical software designed for the analysis of quantitative data all recognize numerical information organized as variables and data records. You need to be very clear about the variables and units of analysis because once you have entered data as numerical information into the software, it becomes a mass of numbers detached from substantive meaning as variables and organization as units in the study.

184 Chapter 9 Initially you might record information about variables for each person or unit on a separate file card or other physical record, such as a row of information in a printed spreadsheet or grid. Because each data record contains information for all the variables for a single person, unit, or case, you will have as many data records as there are cases or units. For example, if you had surveyed 200 people, you would have 200 data records, 1 per person; if you gather data on 85 elementary schools, you will have 85 data records. Each data record contains information on all the measured variables for a specific unit. In data coding, you assign numbers to the attributes or categories in each variable. This is because computers operate with numbers, especially for the statistical analysis of quantitative information. Some of your variables may be quantitative information (this can include information like age in number of years, amount of income), making the use of numbers easy. Other variables are qualitative and use categories not usually thought of as numbers. You can assign an arbitrary number to categories of these variables as a way to keep track of category differences. Variables you measured at the ratio level and interval level already are in the form of numbers, but variables measured at the nominal and ordinal levels are ones to which you will need to assign arbitrary numbers. Example Let us say you gathered information on five variables (age, gender, amount of schooling, and two attitudes). You will code, or assign numbers, for each variable category, or level, for all five variables. Age is easy, since we think of age as the number of years since birth and it is measured at the ratio level. For a nominal variable like gender, you might code males as 1 and females as 2. The numbers for male and female are arbitrary and could just as easily be 5 and 7. Amount of schooling is often measured at the ratio-level, as the number of years of schooling completed. Alternatively, you may have measured it at the ordinal level based on the degree or diploma a person earned. In this case, you might assign numbers to levels such as 0 = did not finish high school, 1 = high school diploma or equivalent, 2 = associates degree or technical school certificate, 3 = four-year college degree (usually called a bachelor’s degree in the United States) or equivalent, 4 = advanced, or postgraduate degree beyond the bachelor’s degree level. If you measured attitudes using a five-point Likert scale, which also is at the ordinal level of measurement, you might assign numbers as follows: 1 Strong agree 2 Agree 3 Disagree 4 Strong disagree 9 Did not answer question

WRITING PROMPT Data Entry Discuss some of the possible consequences of being sloppy or careless when recording codes or putting data into data records? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

9.1.2: Prepare a Codebook As you assign numeric codes to each category of a variable, and even items of missing information, it is essential to keep a written record of coding rules. For example, you may forget whether you coded male as 1 and female as 2, or how you decided to code a person’s answer to a survey question. To keep track, every collection of quantitative data that has been prepared for use in statistical software has a codebook. In secondary data analysis, the codebook comes with the data set. For data that you collect, you must create an organized, detailed codebook. Be sure to make multiple copies of it. If you fail to write down the details of the coding procedure, or if you misplace the codebook, you have lost the key to the data and have to start all over again. Once you have assigned numeric codes and organized all the data records, you can transfer the information into a computer-readable form. Nearly all statistical software requires information in the form of numbers organized in a grid format. Most often in the grid, each row is a data record (on one respondent, subject, or case) and column contains information on variables. Computer programs also allow you to enter the codebook details so that the output can be displayed in a clearly labeled form (e.g., 1 = “Strongly Agree” for a survey question answer). Precoding by including codes in a questionnaire or similar data recorder saves time. The best time to think about coding is before you collect any data. You want to anticipate what will occur during later steps in the overall process. As experimenters organize participants and create measures of dependent variables, they need to think about coding the dependent variable. Survey researchers often precode a questionnaire before collecting data. If you do not precode, the first step after collecting data will be to create codes and a codebook. Precoding speeds up and simplifies the process. In addition to coding variables, you want to assign each case or unit an identification number to keep track of them. Each is the data record for a single case. Identification numbers make it possible to go back to the original source of data (such as a questionnaire or recording

Making Sense of the Numbers 185

sheet) if problems arise later in the process. The identification number links information to a specific research participant, but you can protect privacy by only using the identification number for the data. To do this, after you have coded all the data erase the participant’s name

from all records and henceforth only use the identification number. In Figure 9.3 the computer format condenses the information from survey questions for three respondents into three rows of numbers. Raw data often look like a

Figure 9.3 Coded Data for Three Cases and Codebook This figure shows the steps from a survey questionnaire, to a codebook, to putting the data in a computerreadable format. Your task is to transfer information from the questionnaire—what humans can read—into an organized set of numbers—what computers can read. Humans cannot easily read information in a computerreadable format. Without the codebook, such information is meaningless. Excerpt from Survey Questionnaire (precoded) Respondent ID 1. Note the Respondent’s Gender:

Male (1)

Female (2)

2. How old are you? 3. How many years of schooling have you completed? 4. The first question is about the President of the United States. Do you Strongly Agree, Agree, Disagree, or Have No Opinion about the following statement: The President of the United States is doing great job. Strong Agree (1) Agree (2) Disagree (3) Strong Disagree (4) No Opinion (9) 5. What are your overally feelings about the future, the next 2–3 years specifically? Very positive (1) Somewhat positive (2) Neither positive or negative (3) Somewhat negative (4) Very negative (5) No idea/no opinion/refused (9)

Excerpt from Codebook Column

Variable Name

1–3 4

ID Gender

5–6 7–8 9

Age Schooling Presjob

Future

Description Respondent identification number (001 to 250) Interviewer report of respondent’s gender 1 5 Male, 2 5 Female Two digit respondent’s age in years Respondent’s years of schooling 01 to highest years. The president of the United State is doing a great job.

1 5 Strongly Agree 2 5 Agree 4 5 Disagree 5 5 Strongly Disagree Respondent’s feelings about the future 1 5 Very positive 2 5 Positive 3 5 Neither

Data on first three respondents, as you put the data into the computer Case Number Age 20 001 19 002 21 003

Years of Presjob Future Gender Schooling Answer Answer 3 1 14 1 1 5 15 2 4 2 16 1

Excerpt of data records in the computer, first three data records 0012011413 42736302 182738274 10239 18.82 3947461 ... etc. 0021921551 23334821 124988154 21242 18.21 3984123 ... etc. 0032111624 0123982 1137272631 2345 17.36 1487645 ... etc. etc.

186 Chapter 9 block of numbers. For example, a 15-minute telephone survey of 250 students yields a grid of 250 rows, with each row being a data record. If you asked 25 survey questions, the grid will have at least 25 columns plus space for identification number. It becomes a block of numbers with 250 rows by over 25 columns. Often the first three numbers are identification numbers. Example data in Figure 9.3 show information for the first (001), second (002), and third (003) respondents. Notice that zeros are placeholders to reduce confusion. Eventually you will have 001 through 250. Using the codebook, you can also work backward to decode information. Look at the data record for the statistics software with the codebook in the figure. Case 1 is data on a 20-year-old man with 14 years of education. He answered “Strongly Agree” to the Presjob question and “Neither” to the Future question. Case 2 is data for a 22-year-old woman with 15 years of schooling. She answered “Strongly Disagree” to the Presjob question and “Very Positive” to the Future question. Case 3 has information on a 23-year-old man who completed 16 years of schooling. He answered “Agree” to the Presjob question and “Somewhat Negative” to the Future question.

WRITING PROMPT Codebooks Explain the following metaphor and provide an example from your experience: “A codebook is a key with which you can unlock the identity of variables in the mass of numbers in a computer data file. If you lose the key, you cannot locate variables and the numerical data file becomes useless.” The response entered here will appear in the performance dashboard and can be viewed by your instructor.

check on coding is to use computer software. You can use possible code checking. You ask the software to all the array variables and check for codes outside your coding rules. For example, if you coded gender with 1 = Male, 2 = Female, you can review all the data records for codes on the gender variable. If you discover the number 4 has been entered for the gender in a data record, you have found a coding error that must be fixed. If you discover many coding errors using possible code checking, it is best to recode all the data before proceeding. You can have a perfect sample, perfectly valid measures, and no errors in gathering data, but if you make errors coding or entering data into a computer, your entire research project can be ruined.

9.2: How Do You Describe Quantitative Results? 9.2 Explain the different types of descriptive statistics Statistics refers both to a set of collected numbers such as existing statistics in the Statistical Abstract of the United States and a branch of applied mathematics. As applied mathematics, statistics tell us how to manipulate and summarize collections of numbers. Descriptive statistics is a branch of statistics. The easiest way to think of the many types of descriptive statistics is by the number of variables considered at one time: one, two, or three and more or univariate, bivariate, and multivariate. There is a logic to this terminology: Univariate statistics describe one variable (uni- refers to one; -variate refers to variable), bivariate statistics describe two variables, and multivariate statistics describe three or more variables.

Submit

9.1.3: Clean the Data If you analyze secondary data, someone else has already coded and cleaned the data, and you trust that they were very careful and accurate. If you collect your own data, you want the coding to be highly accurate. You can have a perfect sample, perfectly valid measures, and no errors in gathering data, but if you make errors coding or entering data into a computer, your entire research project can be ruined. After you finished coding all data into a computerfriendly format and entered it into the software, you still need to check and “clean” the data. To do this, recode some of the raw data (e.g., a 10 percent random sample) a second time. If you find zero coding errors in what you entered into the computer, proceed with confidence; if you find many errors, you should recheck all coding. An additional

9.3: Univariate Statistics 9.3 List the various ways to describe numerical data for one variable The easiest way to describe numerical data on one variable is with a frequency distribution. You also can present the same information in a visually dramatic or graphic form, such as the pie chart, bar chart, or histogram. You probably are familiar with the pie chart from elementary school math. It divides the total into slices. The bar chart is used for discrete variables and has a vertical or horizontal orientation with a small space between the bars. The terminology is not exact, but histograms are usually upright bar graphs for interval or ratio data. A good representation of a frequency distribution can be seen in Figure 9.4.

Making Sense of the Numbers 187

Figure 9.4 Univariate Statistics on Cohabitation Opinions about cohabitation in three countries. (From ISSP, 2012). Survey question: To what extent do you agree or disagree? It is all right for a couple to live together without intending toget married? United States

10.0%

10.1%

0.9% 2.8%

Denmark

South Korea 3.9%

4.4%

5.7% 31.0%

20.4%

18.3%

16.1%

45.9%

86.1%

13.6%

30.7%

Strongly Agree Agree

Neither Agree or Disagree Disagree

Strongly Disagree

100

United States Strongly Agree Agree

Denmark Neither Agree or Disagree Disagree

9.3.1: Where Is the Middle? You may want something simpler than showing cases for all variable categories, especially for ratio or interval data. You can summarize the information into a single number, or a kind of average, by using measures of central tendency in descriptive statistics. Three main measures of central tendency are the mean, median, and mode. All are called averages, which is a less precise and less clear way to say the same thing. Most of us learned about these long ago, but you may have forgotten specifics about them. The mode is the easiest to compute but the least useful measure of central tendency. For example, the mode is 5 in the following list of eight children’s ages who are waiting at a bus stop: 6, 5, 7, 10, 9, 5, 3, 5. A distribution can have more than one mode. For example, the mode of the following list is both 5 and 7: 5, 6, 1, 2, 5, 7, 4, 7. If the list gets long,

South Korea Strongly Disagree

just look for the most frequent score. If there are two or more cases with the same value, there is a mode and a score exactly equal to the mode. It has some use. For example, if you were selling ice cream and could only stock one flavor, you would want to know the flavor that is the mode to make the most sales.

188 Chapter 9 The median is the middle point and the 50th percentile. You may be able to spot it or can compute it with a two-step process. First, organize the scores from highest to lowest. Next, count to the middle. If you have an odd number of scores, this is simple. For example, seven people are waiting for a bus; their ages are 12, 17, 20, 27, 30, 55, 80. Count in four people from either end. The median age is 27. Note that the median does not change easily. If the 55-year-old and the 80-year-old both got on one bus and the two 31-year-olds joined the remaining people still waiting, the median would be the same. What if there is an even number of scores? For example, six people at a bus stop have the following ages: 17, 20, 26, 30, 50, 70. If you count from either end, you see that the median is somewhere between 26 and 30. To compute the median, add the two middle scores and divide by 2, or 26 + 30 = 56/2 = 28. The median age is 28. Notice that no one is 28 years old. This list of six ages has no mode because each person is a different age. The mean is the most widely used measure of central tendency. Many advanced statistical measures use it. You compute it by adding up all scores and then dividing the total by the number of scores. The mean age in the previous example is 17 + 20 + 26 + 30 + 50 + 70 = 213; 213/6 = 35.5. Notice that no one in the list is 35.5 years old and that the mean does not equal the median. Changes in extreme values (very low or very high) can strongly influence the mean. If the 50-year-old and 70-year-old left and were replaced with two 31-year-olds, we have the following distribution: 17, 20, 26, 30, 31, 31. The median is unchanged at 28. The mean is 17 + 20 + 26 + 30 + 31 + 31 = 155; 155/6 = 25.8. It dropped almost 10 years, from 35.5 to 25.8. Thus, the mean dropped a great deal when we removed a few extreme values. Normal and Skewed Distribution It is often

useful to plot a frequency distribution. The bottom line has values of a variable (from lowest to highest), and a vertical line on the left indicates the number of cases. As you plot each case, the data points appear to pile up over one

another. The plot of data points can take on many shapes. A frequent shape is a bell-shaped curve. Many naturally occurring phenomena fit this form if plotted (e.g., tree diameter, people’s height, weight, IQ), so it is called a normal distribution. The curve is tall in the center and smoothly falls off to the right and left sides as they approach the extreme highest and lowest scores. The curve looks like a bell and is symmetrical. When your data form a bell-shaped curve, the three measures of central tendency will equal each other. Let us say 15 people of the following ages are waiting for a bus: 2, 4, 5, 6, 8, 8, 10, 10, 10, 12, 12, 14, 15, 16, 18. This forms a symmetrical, bell-shaped curve. The mode, median, and mean are all 10 years old. Not all curves form a nice bell shape; some form a skewed distribution. “Skew” means there is an imbalance with more cases in the extreme upper or lower scores. When this occurs, the three measures of central tendency are not equal. If most cases have low scores and few are high scores, then the mean will be the highest, the median in the middle, and the mode the lowest. If most cases have higher scores with a few extreme low scores, the mean will be the lowest, the median in the middle, and the mode the highest. In general, the median is the best measure of central tendency to use for a skewed distribution.

Summary Review

Three Measures of CentralTendency 1234556789 5 Mode = most common in the list above, because there are two number 5s. 5 Median = midpoint in the list above, because it is the middle: 1, 2, 3, 4, 5, 5, 6, 7, 8, 9 5 Mean = arithmetic average of the list above, because 1 + 2 + 3 + 4 + 5 + 5 + 6 + 7 + 8 + 9 = 50; 50/10 = 5

Figure 9.5 Measure of Central Tendency in Normal and Skewed Distributions The 3 Measures of Central Tendency are sometimes all the same

Number of Cases

Normal Mode, Mean, Median

Skewed Distributions

Median

Mode

Mode Median

Mean

Variable Values

Mean

Variable Values

Making Sense of the Numbers 189

WRITING PROMPT Central Tendency While most of us learn about the three measures of central tendency early in our schooling, many people use the term “average,” instead of the mean. Most also apply the “average” (i.e., mean) in all situations. What are the potential problems with using the general term “average” and applying it in all situations? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

salary is $20,000. It is a better one-number estimate of what the vast majority (50 of 57) people earn than the mean, because it is influenced less by the high salary that Joe pays to himself and his general manager.

WRITING PROMPT Skewed Data Why is it important to check the frequency distribution of data in a social research study before calculating various measures, such as a central tendency? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit Making It Practical: Mean Salary Data Why

avoid the mean for a skewed distribution? Consider the example of salary levels. Joe owns a chain of five coffee shops and employs a total 56 employees, all of them work full-time. Here are their salaries: • 50 cashiers and servers make $20,000 a year. • 5 supervisors make $40,000 a year. • 1 general manager makes $100,000 a year. • Joe, the owner and CEO, pays himself $450,000 a year. Calculate the mean salary (50 * $20,000 = $1,000,000) + (5* 40,000 = $200,000) + $100,000 + $450,000 = $1,750,000 (total payroll), divided by number of people employed including Joe $1,750,000/57 = $30,701.75. The mean is deceptive as an average because 50 of the 57 people earn about $10,000 less than it. Only the five supervisors, who earn more than $10,000 above the mean, are even close to it. The median is the mid-point. One half of 57 is 28.5. If you arrayed all 57 people by salary from lowest to highest, then picked the person between the 28th and 29th position, it would be someone earning $20,000 a year. So, the median

Submit

9.3.2: What Is the Spread? Measures of central tendency give you a single number, the center. An equally important characteristic of a distribution is its spread around the center. Two distributions can have identical measures of central tendency but differ greatly in their spread about the center (see Figure 9.6).

9.3.3: Three Common Ways toMeasure Variation Range Range is the simplest. The range for the bus stop

in front of the bar is from 25 to 35, or 35 - 25 = 10 years. If the 35-year-old got onto a bus and a 60-year-old joined the line, the range would change to 60 - 25 = 45 years. Range has limitations. For example, consider two groups of six people, both have a range of 35 years: 30, 30, 30, 30, 30, 65, and 20, 45, 46, 48, 50, 55.

Figure 9.6 Spread of a Distribution For example, seven people are at a bus stop in front of a bar. Their ages are 25, 26, 27, 30, 33, 34, 35. Both the median and the mean are 30. At a bus stop in front of an ice cream store, seven people have the identical median and mean, but their ages are 5, 10, 20, 30, 40, 50, 55. Ages of people in front of the ice cream store are spread farther from the center, or the distribution has more variability. Of course, you might have 70 people rather than 7, but the principle is the same. Mean = 30

Number of cases

Mean = 30

30 33 34 AGES Narrow Spread

5 55.

30 AGES Wider Spread

190 Chapter 9 Percentiles Percentiles tell us the score at a specific

place within the distribution. You already learned about the 50th percentile, also called the median. Some researchers also give the 25th and 75th percentiles or the 10th and 90th percentiles to describe a distribution in simple terms. For example, at the 25th percentile, 25 percent of the cases are at that score or lower. You compute a percentile similar to the median. Let us say you want to know the 25th percentile of 100 people. Rank the scores and count up from the bottom until you reach number 25. If the total is not 100, you simply adjust the distribution to a percentage basis. Many standardized tests for admission to college or graduate schools, as well as medical evaluations, report with percentiles. If you did average on one of these tests, you would be at or near the 50th percentile. If you scored very low on the test compared to others tested, you might be at the 20th percentile. This means that 80 percent of people scored higher than you did on the same test. If you scored very high, such as the 95th percentile, it means you scored higher than 95 percent of others. It also means no more than 5 percent of test takers equaled or exceeded your score.

Standard Deviation Standard deviation is the most comprehensive and widely used measure of variability. It uses the mean and is a kind of “average distance” between all scores and the mean. Alone, it is difficult to use the standard deviation, but it is useful when comparing populations (see Table 9.1). Making It Practical: Looking at Income Spreads in Evanstown and Newville Variabil-

ity has important social implications. In Evanstown, a city of 100,000, the median and mean family income is $24,700, and it has zero variation. Zero variation means that every family has an income of exactly $24,700. Forty miles away is another city of 100,000, Newville. It has the exact same mean family income of $24,700, but 98 percent of its families (98,000 families) are poor with incomes of $15,000 per year, and 2 percent have incomes of $500,000 per year. Evanstown has perfect income equality, whereas Newville shows extreme inequality. If you do not know about the variability of income in the two cities, you will miss very important information.

Table 9.1 The Standard Deviation Let us use the standard deviation to compare populations. The standard deviation of parents’ level of schooling for class A at local elementary school is 3.317 years; for class B, it is 0.812; and for class C, it is 6.239. Comparing the standard deviations, you quickly see that the parents of children in class B are very similar, whereas those for class C are very different. In fact, in class B, the schooling of an “average” parent is less than a year above or below than the mean for all parents. In short, the parents are very hom*ogeneous. In class C, the “average” parent is more than six years above or below the mean. The parents are very heterogeneous in their level of schooling.

Example of Computing the Standard Deviation [8 respondents, variable = years of schooling] Score

Score − Mean

Squared (Score − Mean)

15 − 12.5 = 2.5

6.25

12 − 12.5 = −0.5

.25

12 − 12.5 = −0.5

.25

10 − 12.5 = −2.5

6.25

16 − 12.5 = 3.5

12.25

18 − 12.5 = 5.5

30.25

8 − 12.5 = 4.5

20.25

9 − 12.5 = −3.5

12.25

Mean = 15 + 12 + 12 + 10 + 16 + 18 + 8 + 9 = 100,100/8 = 12.5 Sum of squares = 6.25 + .25 + .25 + 6.25 + 12.25 + 30.25 + 20.25 + 12.25 = 88 Variance = Sum of squares/Number of cases = 88/8 =11 Standard deviation = Square root of variance = 211 = 3.317 years. Here is the standard deviation in the form of a formula with symbols. Symbols: X = SCORE of case Σ = Sigma (Greek letter) for sum, add together X = MEAN n = Number of cases S = standard deviation s =

g 1 x - x2 2 n - 1

There is a slight difference in the formula depending on whether you are using data for the entire population or a sample to estimate the population parameter. If you have a sample, divide by n - 1 as shown in the formula; if you have the entire population, just use n.

Making Sense of the Numbers 191

Learning from History Teen Birthrate Decline Knowing the mean plus the standard deviation is useful when making comparisons. The U.S. Census reported that in 1990 12.8 percent of all births in the United States were to a woman 20 years old or younger, and by 2009 this had dropped to 10.0 percent of all births. However, these are national figures, and the percentage of births to teens varies greatly by state. For example, in 2009 only 6 percent of all births in Massachusetts were by teens; however, teens were 16.5 percent of all births in Mississippi. Perhaps you want to know whether this decline in teen births across the years was concentrated in just a few states, or whether it was uniform nationwide, and spread across the all regions and states of the United States. An easy way to find out is to compare standard deviations in the two time points. In 1990, the standard deviation, or amount of variation, across the 50 states and District of Columbia was 3.433. In other words, the average deviation, or distance of state teen births from the national statistical mean, was 3.433 percent. In 2009, the standard deviation across the 50 states and District of Columbia was 2.596. The smaller standard deviation in 2009 tells us that not only did teen births decline as a percentage of all births, but variation in the teen births across states also declined. This means that the difference among states in teen births became smaller, or that the states became more similar with regard to the percentage of births by teens over time.

WRITING PROMPT Getting the Full Picture Several years ago, a school asked students at the end of the school year to complete a “satisfaction” survey on each teacher with whom they had a class. Several levels of school officials looked at the class average (i.e., mean) score for each class for each teacher to learn how well the teacher was doing and whether the teacher might need retraining. One teacher recommended looking at the standard deviation as well as the mean scores. How would providing both measures provide a fuller picture of student satisfaction than the mean alone? What might be missed by looking only at the mean and never at the standard deviation? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

9.3.4: Z-scores The standard deviation is very useful for indicating the degree of variation, but it is not easy to interpret by itself. Its smallest value is always zero, indicating that there is no variation at all, and all cases have identical values. However, its largest value can vary depending on the range of a

ariable’s scores. A variable with higher mean or median v scores will have a bigger standard deviation. To deal with this issue and provide a clearer picture, researchers use a standardized measure, Z-scores. It takes into account a variable’s standard deviation and the size of its scores. A Z-score expresses the spread of scores in terms of number of standard deviations from the mean. In short, Z-scores adjust for the characteristics of specific distributions. This is very useful whenever the means and standard deviations are different and you want to compare the same variable in two populations. It is not difficult to calculate. Z-scores are also used in other statistical measures, such as calculating correlation coefficients. How do you calculate Z-scores? Compare Your Thoughts Personally, I do not like the formula for z-scores, which is: Z-score = (Score − Mean)/Standard Deviation, or, in symbols, z = X - X/sd where X = score, X = mean, sd = standard deviation I usually rely on a simple diagram that does the same thing and that shows what z-scores really do. Consider data on the ages of schoolchildren with a mean of 7 years and a standard deviation of 2 years. How do I compute the z-score of 5-year-old Miguel, or what if I know that Yashohda’s z-score is a +2 and I need to know her age in years? First, I draw a little chart from -3 to +3 with zero in the middle. I will put the mean value at zero, because a z-score of zero is the mean and z-scores measure distance above or below it. I stop at 3 because virtually all cases fall within 3 standard deviations of the mean in most situations. The chart looks like this: 23

(a) age in years 3 values5of the mean 7 Now,1I label the and9add or 11 subtract 13 standard deviations from it. One standard deviation above the 23 22 21 0 11 12 13 mean (+1) when the mean is(b)7 and standard deviation is 2 years is just 7 + 2, or 9 years. For a -2 z-score, I put 3years. This is because it is 2 standard deviations, or 2 years each (or22 4 years), the11Mean 12 of 7. My13 diagram now 23 21lower than 0 (a) looks like this: 1

7 0 (b)

13 age in years 13

It is easy to see that Miguel, who is 5 years old, has a z-score of -1, whereas Yashohda’s z-score of +2 corresponds to 11 years old. I can read from z-score to age, or age to z-score. For fractions, such as a z-score of -1.5, I just apply the same fraction to age to get 4 years. Likewise, an age of 12 is a z-score of +2.5.

192 Chapter 9 Making It Practical: Z-scores and GPAs

Z-scores are easy to calculate once you have the mean and standard deviation. For example, an employer interviews students from Kings College and Queens College. Suzette is from Kings College and has a grade-point average of 3.62. Jorge is from Queens College and has a grade-point average of 3.64. Both students took similar courses. The colleges both grade on a 4.0 scale. However, the mean grade-point average for students at Kings College is 2.62 with a standard deviation of .50, whereas the mean grade-point average at Queens College is 3.24 with a standard deviation of .40. The employer suspects that professors at Queens College are easy graders, and the grades at Queens College are inflated. To adjust the grades of the two students for the grading practices of the two colleges (i.e., create standardized scores), the employer converts the GPA for each student into z-scores. She does this by subtracting each student’s score from the mean, then dividing by the standard deviation. Suzette’s z-score is 3.62 - 2.62 = 1.00/.50 = 2, whereas Jorge’s z-score is 3.64 - 3.24. = .40/.40 = 1. After this adjustment, the employer sees that Suzette is two standard deviations above the mean GPA in her college. By contrast, Jorge is only one standard deviation above the mean GPA for his college. Suzette’s absolute grade-point average is a little lower than Jorge’s but compared to all the students in each of college her grades are a lot higher than Jorge’s were. Z-scores help the employer see that Suzette’s grades

were well above the average student at her college. Jorge’s grades were only slightly above an average student at his college.

WRITING PROMPT Z-scores Suppose your school wants to hire a new teacher so you look at teaching evaluation scores of two you might hire. On a 5-point scale (5 is best), teacher A has a mean score of 3.9 and teacher B has 3.7. Yet at teacher A’s past school, the mean for all teachers at that school is 4.0 and standard deviation is 1.0. For teacher B’s school, the mean is 3.7 with a standard deviation of 2.0. Based on student ratings, which teacher is better, relative to all other teachers at their past school? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

9.3.5: Alternative Ways to Display Information on One Variable Univariate statistics show us various features of quantitative information for one variable. Sometimes we gather univariate statistics on one variable within two or more

Figure 9.7 Map of U.S. Teen Birth Rates The map shows us that the states with a higher percentage of teen births are located in the south, close to Texas. Five of six states in the highest category are in the south (West Virginia is the nonsouth exception). By displaying a variable in geographic space, we can learn more about it. The map does not explain why more teen births occur in certain areas; that is a topic for further inquiry. Source: Martin, J. A., Hamilton, B. E., Ventura, S. J., Osterman, M. J. K. S.C., & Mathews, T. J. (2015). Births: Final data for 2013. Hyattsville, MD: National Center for Health Statistics.

WA MT OR

ID WY NV

MN WI

IA IL

MO OK

NM TX

WV VA NC

AR MS

NE UT

FL 40.0–49.9 30.0–39.9 20.0–29.9 Less than 20

SC GA

NH MA RI CT NJ DE

U.S. teen birth rate was 26.5 in 2013

Making Sense of the Numbers 193

Figure 9.8 Teen Birth Rates Trends in Teen Birth Rates and Divorce Rates in the United States

Birth Rate per 1,000 Women Aged 15–19

100 80 60 40 20 0 1940 1945 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015 Year 6

Divorce Rate per 1000

5 4 3 2 1

2012

2009

2006

2003

2000

1997

1994

1991

1988

1985

1982

1979

1976

1973

1970

1967

1964

1961

1958

1955

1952

1949

1946

1943

1940

Year

large units (e.g., countries) for purposes of comparison. In addition, we can display information on one variable using geographic space or time to provide greater insights. If you have information by geographic location, you can display data on a map. A map appears simple compared to a table or chart, but it contains more information. In addition to showing different levels of a variable, the map displays its spatial form, or shows how the spread across a geographic area. For example, the National Center for Health Statistics provides information on the rate of births to women aged 15–19 by state in 2013 (see Figure 9.7). Likewise, if you have data that vary by time, you can display the data on a variable along a time line. Compared to a table or chart, a time line shows more information. In addition to showing different variable levels, it shows how the variable varies across time. Suppose you are interested in teen birth rates over time. Information is available over a 70-year period

(1940–2010). Knowing the mean or median is useful, but plotting teen birth rates by year is more informative. We quickly see that teen birth rates began low and grew in the 1950s then declined. When we compare this trend to divorce rates for the same period, we notice that the times when divorce rates were low were the ones when teen birth rates were high and both declined after 1990.

WRITING PROMPT Learning from Univariate Data Identify the pattern in the U.S. map for the variable, births to women aged 15–18. Does the pattern suggest any hypotheses or other variables to consider that might be associated with levels of teen birth rates? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

194 Chapter 9

9.4: Bivariate Statistics 9.4 Analyze the three techniques to help see the relationship between two variables in bivariate statistics You can describe the number of cases in one variable’s categories by looking at univariate frequency distribution or calculating its central tendency or variation, or by placing the variable in a map or along a time line. While the univariate information is valuable and can reveal some patterns, it does not allow you to test or examine a causal relationship. To do that, you need bivariate statistics that allow you to examine two variables together, one as the cause (independent variable) and one as the effect (dependent variable). People often confuse trends in a variable across time with a causal relationship. For example, plotting year by teen births shows you that as time passed the number of teen births declined and rose. The trend is not a causal explanation because the passage of time alone is not what caused a change in the teen birth variable. Many things change over time (phases of the mood, hair styles, etc.), so you need to create an explanation and specify a causal variable. For example, you may hypothesize that a social condition changed over time (e.g., increased schooling, greater access to contraceptives, or heightened value for chastity) and teen births also changed over the same time period. This allows you to start to examine whether the two variables are related. With bivariate statistics, we consider two variables together simultaneously and evaluate the relationship between them—that is, whether and how the variables operate together. Bivariate statistical relationships rest on two ideas: covariation and statistical independence. If two variables covary, the cases with values on one variable tend to have certain values on a second variable in a patterned way. For example, people with higher values on the income variable also tend to have higher values on the life expectancy variable. Likewise, people with lower incomes have lower life expectancy. To say this in a shorthand way—income and life expectancy are related to each other, or they covary, i.e., they vary together. If we know income levels for a collection of people, we also know something about their likely life expectancy. If there is a relationship, knowing the income levels allows us to predict general patterns of life expectancy for a large collection of people with modest accuracy. In short, life expectancy (the dependent variables) depends on income levels (the

independent variable). We can restate this as income level has a causal effect on a person’s life expectancy. This is a simple causal relationship and a testable hypothesis. It is only part of a full explanation that could include explaining why and how the two variables are related in a certain way. Statistical independence is the opposite of covariation. For example, Rita wants to know whether there is a relationship between number of siblings and life expectancy. If the variables have statistical independence, then people with many brothers and sisters will have the same life expectancy as those who are only children. In other words, knowing the number of siblings tells Rita nothing more about the person’s life expectancy. Life expectancy does not depend on number of siblings. A first task in the statistical analysis of two variables is to find out whether they covary, or whether there is statistical independence between them. Three techniques will help you see the relationship between two variables: 1. A scattergram, or a graph or plot of the bivariate relationship; 2. Cross-tabulation, or a percentage table of two variables; and 3. Measures of association, or statistical techniques that indicate the amount of covariation between the variables with a single number (e.g., correlation coefficient).

9.4.1: Scattergrams We put numeric information into charts, graphs, and tables to reveal patterns in the data and to communicate what is in the data to others. If you do not know how to read data patterns in a chart, graph, or table and see variable relationships in them, you will not be able to set up charts, graphs, and tables that communicate this to other people. A scattergram, or scatterplot, is a graph that allows you to see bivariate relationships in data. Usually the independent variable (often symbolized by the letter X) goes on the horizontal axis and the dependent variable (often symbolized by Y) on the vertical axis. In the graph, you want to place the lowest value for each in the lower left corner and the highest value at the top or to the right. Scattergrams are ideal for interval- or ratio-level data and do not work with nominal data. For ordinal-level data, you should have many levels or categories (10 or more). Scattergrams do not work well with few cases (under 10) but are ideal with many cases (50 or more).

Example Study Cohabitation and GenderEquality How does premarital cohabitation influence the gender-based division of household labor within a household?

Batalova and Cohen (2002) examined the relationship among couples in 22 countries using data from the International

Making Sense of the Numbers 195 Social Survey Program. They created a “gender division of labor index” by combining answers to four survey questions about household tasks—who usually does laundry, cares for sick family members, does shopping, and plans dinner. They assigned a score of 1 (wife always does the task) to 5 (when the husband always does the task). They added the scores for the four questions and divided by 4 to create a new index ranging from 1 to 5, with high scores indicating more contributions by the husband. Their primary independent variable was whether a couple cohabited before marriage. They also controlled for factors found to influence the household division of labor in past studies. In general, they found that women more than men do routine household tasks in all countries. They also looked at cohabitation rates by country and at a Gender Empowerment Measure for each country. The United Nations provides an Empowerment Index on

women as a percentage of occupants of parliamentary seats, a percentage of administrators and managers, a percentage of professional and technical workers, and an indicator of gender pay equity for each country. Batalova and Cohen’s scattergrams of household division of labor by cohabitation rates and by gender empowerment both showed that the more widespread cohabitation was in a country, the more husbands contributed to

Figure 9.9 Married-Couple Division of Labor, Cohabitation Rate, and Gender Empowerment Measure in 22 Countries

Figure 9.10 Predicted Household Division of Labor by

household tasks. Also, the higher a country was on the Gender Empowerment Index, the more husbands contributed. They also compared married couples who cohabited before marriage with those who had not cohabited and learned that those who cohabited had a more equal division of housework overall. In addition, they used a statistical model to examine the cohabitating versus noncohabitating difference with the effects of national cohabitation effects and

Premarital Cohabitation

These bivariate graphs show that cohabitation rates are highest in countries where gender equality is stronger.

Couples in countries where many cohabitate have a more gender equal division of household labor, plus those who cohabitated themselves have even higher levels of gender equality

Adapted from Figure 1 on Page 750 in Batalova and Cohen (2002).

Adapted from Figure 2 on Page 752 in Batalova and Cohen (2002).

Russia

Canada Norway U.S.A. W. Germany Great Britain Austria New Zealand Bulgaria Israel E. Germany Netherlands Hungary Slovenia Australia Poland Italy N. Ireland

25 20 15 10 5

1.6

1.7

1.8 1.9 2.0 2.1 2.2 2.3 Division of Household Labor

0.80

Sweden

Gender Empowerment

0.75 0.70

Austria

0.65

2.4

15 10

.80

Norway

.75

Italy Japan

Great Britain

Poland Hungary Bulgaria

Israel Slovenia

1.8 1.9 2.0 2.1 2.2 2.3 Division of Household Labor

Cohabiting Couple Non-Cohabiting Couple

.65 .60 .55 .50 .45 .40

Russia 1.7

1.85 1.90 1.95 2.00 2.05 2.10 2.15 2.20 2.25 Division of Household Labor

.70

Ireland Czech Repb.

0.55

1.6

0 1.8

2.5

New Zealand Canada W. Germany E. Germany Netherlands U.S.A. N. Ireland

0.60

0.40 1.5

Australia

0.50

Ireland

Japan

0 1.5

0.45

National Cohabitation Rate

Sweden

Czech Repb.

Gender Empowerment

National Cohabitation Rate

Cohabiting Couple Non-Cohabiting Couple

2.4

2.5

0 1.8

1.85 1.90 1.95 2.00 2.05 2.10 2.15 2.20 2.25 Division of Household Labor

196 Chapter 9 gender employment. They found that household equality was greater for the couples who had cohabited at each level of national cohabitation and gender empowerment, although the

How Do You Construct a Scattergram?

STEP 1. Note the highest and lowest values (the range) of the two variables. Mark each axis with the name of the variable it will display and divide each axis into appropriate number ranges (graph paper is helpful). For example, Cohabitation Rate in a country may be one variable. You find data for a collection of 17 countries that it ranges from 0.2 percent to 14.4 percent. You can label the horizontal line (x axis) and divide it into interval categories from zero to 14.5 dividing it into 15 equal lines. For the same 17 countries, you have Divorce Rate as a second variable. It varies from 0.8 to 3.7. You can label the vertical line and divide it into interval categories from zero to 4 subdividing it into eight categories. STEP 2. For each case, find the value of each variable and mark the graph where the two values intersect. For example, you put a data point at where a rate of 8.9 Cohabitation and 2.3 for Divorce intersect (Australia), the one where 6.5 on Cohabitation and 2.4 on Divorce intersect (Austria).

size of the equality boost from a couple’s cohabitation varied depending on whether national cohabitation or gender empowerment was being considered.

Examples of Scattergrams

Figure 9.11 Scattergram Examples A visual image of the various patterns on a scattergram

Not Wrong

Cohabitation Rate (2006–11)

Australia

8.9

2.3

Austria

6.5

2.4

Canada

8.9

2.23

Czech Republic

2.9

3.0

11.5

2.7

Finland

11.8

2.5

France

14.4

2.1

Germany

5.3

2.3

Greece

1.7

1.2

Ireland

5.9

0.8

Italy

2.0

0.9

Japan

2.1

2.0

Norway

10.7

2.1

Poland

1.3

1.7

Turkey

0.2

1.4

United Kingdom

8.7

2.4

United States

5.5

3.7

Source: Data from OECD

STEP 3. Continue plotting data points for each country by placing a dot on the grid at the intersection of the two variables until you have plotted all cases on the graph.

Non-Linear

Linear

Divorce Rate (2008)

Denmark

Sometimes

Forms of a Scattergram: Linear versus Non-Linear

Table 9.2 Comparison of Cohabitation Rates and Divorce Rates for 17 Countries Country

Almost Always

Direction of Scattergram: Positive versus Negative

Positive

Negative

Precision of Scattergram: Precise versus Imprecise

Precise

Imprecise

What can you learn from scattergrams? Compare Your Thoughts A scattergram reveals three aspects of a bivariate relationship: its form, direction, and precision.

Making Sense of the Numbers 197

• Form. We can classify relationships between two variables into one of three basic forms: independence, linear, or curvilinear. • Independence usually indicates the absence of a causal relationship between the variables. On a scattergram, the plotted data points appear to be random scatter without any pattern. Although rare, independence can also look like a straight line connecting all the data points that is exactly parallel to the horizontal or to the vertical axis. • A linear relationship indicates there is a possible causal connection between the two variables. In the scattergram data points align in a straight line that goes from one corner to another. This diagonal line does not have to connect all the data points but is a clear pattern. • A curvilinear relationship (also called nonlinear) indicates a complex causal pattern. The data points in the scattergram do not follow a straight line but may look like a pattern of data points that generally forms a U curve, right side up or upside down, or an S curve. • Direction. Linear relationships can have a positive or negative direction. • The plot of a positive relationship looks like a diagonal line of data points that begins in the lower left and goes to the upper right. Higher values on X tend to go with higher values on Y, and vice versa. The income and life expectancy example described a positive linear relationship. • The plot of a negative relationship looks like a line of data points from the upper left to the lower right. It means that higher values on one variable go with lower values on the other. For example, people with more education are less likely to have been arrested. If we look at a scattergram plotting data on a group of males where years of schooling (X axis) by number of arrests (Y axis), we see that most men with many arrests are in the lower right because most of them completed few years of school. Most cases with few arrests are in the upper left because most have had more schooling. • Precision. Bivariate relationships also vary in their degree of precision. • A very precise relationship occurs when all the data points align tightly along the line that indicates the relationship and direction. • A relationship with little precision shows data points scattered about the line. The pattern of a line is evident but many data points are some distance from an imaginary line that organizes points in the scattergram. The degree of spread among the points on the graph indicates the degree of precision. Measures of association (discussed later in this chapter) get larger or stronger when there is more precision.

WRITING PROMPT Learning from Scattergrams How can plotting the data points of two variables of main interest help as a first step when looking at data you collected? What might the scattergram reveal as you attempt to make sense of the data? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

9.4.2: Bivariate Tables A scattergram provides a picture of a relationship for data at the interval or ratio level of measurement but does not work well for variables at the nominal or ordinal levels. Bivariate tables allow you to see relationships between two nominal- or ordinal-level variables. Bivariate tables can show the same patterns in data as a scattergram does but in a more condensed form. Use scattergrams for variables measured at the ratio or interval levels that have many categories and bivariate tables for data at the nominal or ordinal level of measurement with variables that have a few categories (two to five). Before you can use a bivariate table for data measured at the interval or ratio you must first collapse the data into a few categories, essentially converting them to an ordinal level of measurement. To create a bivariate contingency table, you use the process of cross-tabulation. It organizes, or tabulates data in a table that has two variables. Tables show data where categories of two variables intersect, or cross, one another. It is the same process as plotting data points on two variables in a scattergram but in a table format. We call the tables “contingent” because the cases in each category of one variable are distributed by, and contingent upon, the categories of a second variable. After you have entered all cases into table categories for both variables, the table will reveal how many cases are in each combination variable categories for the two variables. Instead of a horizontal and vertical axis as in a scattergram, bivariate tables have columns and rows. Columns are vertical, are spread along the top of a table. Rows are horizontal, listed down the left side of a table. The data within contingency tables can be displayed as a count, called raw frequency, or as percentages. Making It Practical: Cross-Tabulation Let us say you are interested in the relationship between bright car colors and the marital status of their owners. Car color is the dependent variable and marital status is the independent variable. Both are measured at the nominal level. Perhaps you hypothesize that single people (both never married and divorced) are more likely to drive bright red

198 Chapter 9 cars compared to people who are currently married. How might you determine whether these variables are related?

You can gather data to test the hypothesis in several ways. You can check existing statistical records at a Department of Transportation for car color and marital status if available. Alternatively, you can go to several large parking lots and note the license plates and colors of all cars, and then check marital status associated with car registration. Let us say that you found that 13 percent of the 320 vehicles parked in a sample from parking lots are bright red. Official statistics indicated that about 22 percent of the adult population is single, never married. If car color and marital status are not related, you would expect to find 22 percent of red car owners to be single and about 13 percent of single people owning red cars. This would indicate no connection, i.e., they are statistically independent of one another. Another way to test the hypothesis would be to conduct a survey. For example, you surveyed 188 people about their marital status and car color, then placed the data into a raw count bivariate table of car color by marital status (see Table 9.3).

Table 9.3 Red Car Ownership and Martial Status, RawCount Table Color of Car

Single

Not Single

Total

Red

Other

124

158

Total

138

188

The totals for each variable category, the same as in a variable’s frequency distribution, are along the right side (for the row variable) or across the bottom (for the column variable). Each cell of the table shows a count of the number of cases that are at the intersection of two variable categories (a cell is a small box in the center part of a table where each category of two variables crosses one another). Creating a raw count bivariate table, either manually or with statistical software, is straightforward. However, interpreting patterns in a raw count table is difficult when the rows or columns have different totals. Patterns in the data are revealed by comparing the amount of data in some cells relative to others in the table. Yet, unequal

ariable totals make comparing relative cell size very diffiv cult. For this reason, we nearly always convert the raw count table into a percentage table. It is easy to be misled by a raw count table when the row or column totals of variable categories are different sizes. Using percentages adjusts for unequal totals and standardizes information to permit easy comparison. For example, the table indicates that 16 single people and 14 nonsingle people own red cars. To compare the relevance of these numbers is difficult because there is a total of 50 single and 138 nonsingle people. For a meaningful comparison, you must first standardize the data in the cells based on the totals. The easiest way to do this is to use percentages in the bivariate tables. This converts information in the cells of the table into comparable units. Calculating a Percentaged Table You can per-

centage the data in a table in three ways: by row, by column, and for the total. The first two are often used and can show relationships. The last way is almost never used and nearly impossible to interpret. You may ask: Is it better to percentage by rows or columns? Calculating percentages by row or by column can be appropriate; it all depends on the question you want to answer with the data. It also depends on how you place the independent and dependent variables in a table, i.e., which variable goes into the row and which into the column. Whether you calculate manually or use statistical software, the mechanics are the same. To calculate column percentages, compute the percentage for the number of cases in each cell based on the column total of a variable category (located at the bottom of the column). In the example, the total of the first column is 50. This means 50 single people are in the study. The first cell of that column indicates that 16 of the 50 single people own red cars. The percentage of red car owners among all the single people is 16/50 = 0.32 or 32.0 percent. Now look at 14 red car owners who are not single (in the second column, first row of Table 9.3). The table shows that 138 nonsingle people are in the study. So, red car owners are 14/138 or 10.1 percent of all nonsingle people. By reading across and comparing the two percentages in the first row of Table 9.4, you see that 32 percent (red car owners among singles) is larger than 10.1 percent (red car owners among nonsingles). This indicates that single people are nearly three times more likely to own a red car than a nonsingle people, supporting the hypothesis.

Table 9.4 Red Car Ownership and Martial Status, ColumnPercent Table Color of Car

Single (%)

Not Single (%)

Red

32.0

10.1

Other

68.0

89.9

Total (N)

100 (50)

100 (138)

Total (%) 16.0 84 100 (188)

Making Sense of the Numbers 199

Computing percentages by row follows a similar process in Table 9.5.

Table 9.5 Red Car Ownership and Martial Status, RowPercent Table Color of Car

Single (%)

Not Single (%)

Red

53.3

46.7

Total (N) 100 (30)

Other

21.5

78.5

100 (158)

Total

26.6

73.4

100 (188)

To compute the percentage of each cell, take the number of cases in the cell as a percentage of the row total. The row totals are on the far right side of the table. Using the same cell with 16 in it (for single people who own a red car), you can determine it as a percentage of the row total, or the 30 red car owners. The row percentage for the cell is 16/30 = 0.533 = 53.3 percent. This tells you that a little over one-half red car owners are single. Computing percentages by row or column produces different percentages for the same cell, unless the row totals are the same as the column totals—something that rarely occurs. Reading a Percentaged Table Once you understand how to make a table with percentages, reading it becomes easier. To read a table, first look at the title, the variable labels, and any background information. Next, look at the direction in which percentages have been computed—in rows or columns. Notice that Tables 9.3 through 9.5 share much the same title because the same variables are used. Many titles do not indicate the direction of percentages. Some researchers use abbreviated tables that omit the 100 percent total. This can create confusion. It is best to include all the parts of a table with clear labels. The row or column percentages in a bivariate table enable you to address different questions. Looking at the upper left cell in the column percentage table answers the question: What percentage of the single people own a red car? You see that 32 percent of single people own a red car. The row percentage table addresses the question: What percentage of red car owners are single? You can see that 53.3 percent of all red car owners are single. The first question draws attention to the proportion of red car owners among people of different marital statuses. Looking across the top row, you can compare percentage of singles who own red cars (32 percent) versus nonsingles who own red cars (10.1 percent). This is what your hypothesis would predict—red car ownership is more common among single people. The second question draws attention to the owners of red cars. It does not address your hypothesis. Computing percentages in different ways enables you to address similar but different questions. Keeping them

straight is important. Each question that a table can answer can have practical uses. Knowing that about one in three single people drive red cars is useful to a car salesperson. It tells the salesperson that there is about a one in three chance that a single customer will be attracted to a red car, while nearly 90 percent of nonsingles prefer color other than red. Knowing that about half of all red car drivers are single or nonsingle may be helpful to a company selling car finish protection products. The company knows marital status is not a distinguishing feature among red car owners, but can focus marketing nonred car finish products to nonsingles, since 78.5 percent of all customers without red cars will be nonsingle. A hypothesis guides you to look at row or column percentages. At the beginning, you might want to calculate percentages both ways and practice interpreting what each says. If your hypothesis is “a person’s marital status influences his or her choice of car color,” place marital status as the column variable, then percentage by column. If you put the independent variable in the column, calculate percentage by column. There is no “industry standard” for putting independent and dependent variables as row or column. Most researchers place the independent variable as the column and percentage by column, but a large minority put the independent variable as the row and percentage by row. When reading a percentage table, make comparisons in the opposite direction from direction in which percentages are computed. This sounds complex, but it means the following: If a table’s percentages total by column, compare across the rows. For example, compare percentages 32 percent of singles versus 10.1 percent of nonsingles across the top row to see whether categories of the column variable differ. The example indicates that about one-third of single people own red cars versus 10 percent of nonsingle people. If a table’s percentages total by row, compare down/up the columns. For example, compare 53.3 percent versus 21.5 percent to see that nearly half of all red cars are owned by single people, yet the only about one in five nonred cars have single owners. Seeing Relationships in a Percentaged Table

With practice, you will be able to quickly see relationships in a percentage table. If there is no relationship among variables in a table, the cell percentages look equal across all rows or columns. The relationship will have a direction (positive/negative) only if the data are at an ordinal level of measurement, or if interval/ratio-level data have been grouped into ordinal-level categories. Nominal-level data have no direction. When both variables are ordinal, a table can reveal a linear relationship and you can determine its direction. A linear relationship among variables appears as larger percentages arrayed along the diagonal cells. If it is a curvilinear

200 Chapter 9

Table 9.6 General Social Survey 2004, Hours of TV Watching per Day by Educational Degree Amount of Education Daily TV Watching

Less than High School (%)

High School (%)

Some College (%)

4-yr College Degree (%)

Graduate Degree (%)

Total (%)

8+ hours

18.9

7.8

3.8

4.3

7.5

5–7 hours

12.6

7.6

4.3

7.3

3–4 hours

36.0

30.0

26.6

20.6

18.5

27.8

1–2 hours

29.7

49.4

59.5

58.9

66.3

51.1

2.7

5.3

2.5

12.1

10.9

6.3

None Total

100

(N)

(111)

(476)

141

(92)

(899)

Table 9.7 General Social Survey 2004, Hours of TV Watching per Day by Age Age Daily TV Watching

Under 30 (%)

30–40 (%)

41–50 (%)

51–65 (%)

66+ (%)

Total (%)

8+ hours

6.7

7.1

2.9

7.9

13.5

7.4

5–7 hours

9.6

4.7

5.3

6.9

12.0

7.4

3–4 hours

24.7

24.2

26.5

28.1

39.1

27.8

1–2 hours

51.7

53.6

58.8

54.2

32.3

51.3

7.3

10.4

6.5

3.0

6.3

None Total

100

(N)

(178)

(211)

(170)

(203)

(133)

(895)

relationship, the largest percentages form a curve pattern across cells. For example, the largest cells might be the upper right, the bottom middle, and the upper left. It is easiest to see a relationship in a moderate-sized table (9 to 16 cells), where most cells have several cases (a minimum of five) and the relationship is strong and precise. You can transfer the skills applied to read data patterns in a scattergram to see relationships in a bivariate percentage table. This only works if the lowest values on both variable categories are in the bottom left corner of the table. Imagine a scattergram that has been divided into 12 equal-sized sections, as if a grid was placed on top of it. The cases in each section correspond to cases in the cells of a table superimposed onto the scattergram. In other words, a percentage table with ordinal variables looks like a condensed form of the scattergram. The bivariate relationship line in a scattergram corresponds to the diagonal cells in a percentaged table. To see strong relationships in percentage tables quickly, circle the biggest percentage in each row (for column-percentaged tables). Does a diagonal line appear? Consider two ordinal variables, education level and TV watching in Table 9.6. There is a negative relationship between the variables—more highly educated people watch TV fewer hours per day than less educated people. How does the

table show this? Among people who did not finish high school, 18.9 percent watch TV eight hours or more per day. As levels of schooling increase, TV watching declines until at a graduate degree no one watches that much TV. The table shows a negative and linear relationship. By contrast, the bivariate table (Table 9.7) of age by TV watching shows a positive linear relationship. As age increases so does TV watching. Circling the biggest percentage in each row helps to display a relationship. If differences are small (under 6 percent), circle the two biggest cells.

9.4.3: Measures of Association A measure of association condenses information about a bivariate relations into a single number. It expresses the strength, and often the direction, of a relationship. There are many measures of association. The level of measurement in the data determines the proper one to use. Most are symbolized by letters of the Greek alphabet: lambda, gamma, tau, chi (squared), and rho. The emphasis here is on interpreting the measures, not on their calculation. To calculate and understand them fully, it is best to complete a beginning statistics course. Some people use the term “correlation” and confuse the general idea of association with one specific statistical measure of association (see Figure 9.12).

Making Sense of the Numbers 201

Figure 9.12 Six Measures of Association You can see that the six commonly used bivariate measures of association have a lot in common. Most range from -1 to +1, with negative numbers indicating a negative relationship and positive numbers a positive relationship. Measure

Greek Symbol

Cramer’s V

Type of Data

Highest Association

Nominal

1.0

Independence 0

Lambda

Nominal

1.0

Gamma

Ordinal

+1.0, -1.0

Tau (Kendall’s)

Ordinal

+1.0, -1.0

Rho

Interval, ratio

+1.0, -1.0

Nominal, ordinal

Infinity

Chi-square

9.4.4: Results with More than TwoVariables

The interpretation of most measures of association is straightforward. If a measure of association is zero, it indicates statistical independence. If it is a large nonzero number, it indicates a relationship. As the number becomes larger, it indicates a stronger relationship. A very strong relationship between two variables means you would make few errors predicting a dependent variable based on knowing the independent variable. In short, knowing the independent variable causes a big reduction in prediction errors. You can use measures of association with bivariate table, and measures of association provide more precise information about the strength of relationships than possible by just looking at the tables. The two tables of TV watching categories with educational degree and age appear to be similar. Comparing the two gammas for the tables, the negative relationship of education degree by TV watching has a stronger gamma (-0.384) than the positive relationship of TV watching with age (0.163). Remember, the negative sign indicates direction, not strength of a relationship.

Showing an association between two variables is valuable. However, it is not sufficient to say that an independent variable causes a dependent variable. In addition to temporal order and an association, before you can say “cause” you must eliminate alternative explanations. Alternative causes can make a relationship that you see in the data spurious. Experimental researchers control for alternative causes of the dependent variables by choosing a research design that physically reduces the threats to internal validity. Spurious results occur when a third variable (representing an alternative explanation) is the true cause, even if there appears to be an association. Failure to consider such a variable means that a bivariate relationship in the data could be a mirage. Experimenters control for alternative explanations by improving internal validity. Nonexperimental researchers use control variables that measure possible alternative explanation. They must anticipate possible

Figure 9.13 Strength of Measures of Association A score of 1.0 means a 100 percent reduction in errors, or perfect prediction of the dependent variable based on the independent variable. Looking at the Strength of Measures of Association

21 Strongest Negative

2.50

Moderate Weak Negative Negative 2.30 2.10

Strong Negative

Weak Moderate Positive Positive .10 .30

Independence

.50 Strong Positive

1 Strongest Positive

Here are measures of association for data you have seen in this chapter: • The table with marital status and car color has two nominal-level variables. Cramer’s V 5 .264 • The scattergram of education years by TV watching hours has two ratio-level variables. Rho 52.293 • The table of education degree by TV watching categories has two ordinal-level variables. Gamma 5 .384 • The table of age by TV watching categories has two ordinal-level variables. Gamma 5 1.163

202 Chapter 9

Table 9.8 Color of Car by Age (Percentaged by Column) Color of Car

Under 30 Years Old (%)

30 to 60 Years Old (%)

Over 60 Years Old (%)

Total (%)

Red car

34.6

10.5

3.2

16.0

Other color

65.4

89.5

96.8

84.9

Total (N)

100 (52)

alternatives and collect data on control variables that indicate the alternatives at the same time as collecting other data. Only then can they test whether a bivariate relationship is real or spurious. Multivariate tables and statistics enable you to test whether a bivariate relationship is spurious. Another advantage is that they reveal the relative influence on a dependent variable from several independent variables simultaneously. When two independent variables both influence the same dependent variable, often you want to learn which of them has a bigger impact. To see how multivariate tables work, let us return to the example of single people and red cars. Single, never married people are usually younger than married or previously married people. In fact, the mean age of adults is 33.1 years old, whereas the mean for nonsingle adults is 45.57 years old. Younger people also like red cars. Notice in Table 9.8 that over one-third of young people own a red car, versus 10 percent of the middle-aged and about 3 percent of older people.

100 (105)

100 (31)

100 (188)

Does Age or Being Single Have a Bigger Impact on Owning a Red Car? We can compare the two

Cramer’s V numbers. For color of car by age V is .323, and for marital status by color of car V is .264. These are relatively close. Another way to check is to use age as a control variable. To use it as a control variable, we can create three tables, one for each age group (see Tables 9.9, 9.10, and 9.11). This lets us look at the relationship of marital status on car color among the people in each age group. If it is equally strong within each age group, then age group is not important. If it does not appear any longer, then age must be the most important factor.

Table 9.9 People under Age 30 Only Car Color Red car Other color Total (N)

Single (%)

Not Single (%)

77.8

44.1

Total (%) 55.8

22.2

55.9

44.2

100 (18)

100 (34)

100 (52)

Cramer’s V = .322

Table 9.10 People Age 30–60

Tips for the Wise Consumer

Car Color

Single (%)

Not Single (%)

Total (%)

Noticing Statistical Significance

Red car

18.25

13.85

14.35

Other color

81.85

86.25

85.75

100 (11)

100 (94)

100 (105)

Single (%)

Not Single (%)

Total (%) 19.4

Do not let a maze of numbers, tables, charts, and statistics in a report intimidate you when you look at the results. Remember, even if the author is reporting advanced statistics that are beyond what you have learned, results only document what is in the data. Most advanced statistical techniques use a form of statistical significance that tells you which variables or parts of the results are most important. This is how researchers identify which of many variables are most powerful. Many authors use superscripts or asterisks to indicate the level of statistical significance, such as one * for .05 level, ** for .01 level, and *** for .001 level. You can see them described in a key or footnote. Other times they use the symbol p for probability. They may give you something such as p 7 .05. This indicates that p is greater than the .05 level of statistical significance. When you see a list of variables and some have a symbol or asterisks, you can interpret those as the most influential ones (i.e., ones with the biggest impact on the dependent variable). Many researchers try to explain the meaning of their results in simple language, but this not always easy to do because the results often include advanced statistical ideas and terms.

Total (N) Cramer’s V = .038

Table 9.11 People over Age 60 Car Color Red car

20.0

Other color

100

80.0

80.6

Total (N)

100 (1)

100 (30)

100 (31)

Cramers V = .089

A red car is more popular for people under 30 years old than older people, and the impact of being single on wanting a red car is strong for people under 30 years old. Over three-fourths of single people under 30 own a red car. Being single makes little difference in red car ownership among people over 30. This suggests that of the two factors, age and marital status are working together, with age being somewhat more important of the two. Had we seen no relationship in the under 30 table as for the other two age groups, this would suggest that the relationship

Making Sense of the Numbers 203

between marital status and car color was spurious. If the martial status–car color relationship were not spurious, it would persist even after we considered control variables. Finding a bivariate relation to be spurious is very valuable. Three-variable percentage tables can reveal a lot, but looking at many of them can become very complex to detect spurious relationships. Looking beyond three variables at one time becomes very difficult with contingency tables.

Example Study Is the Cohabitation–Divorce Relationship Spurious? For many years, researchers noted that persons who cohabitated had a higher divorce rate than persons who had never cohabitated. The reason for a strong bivariate relationship between the experience of cohabitation and subsequent divorce was unclear. Some religious leaders called it evidence of the immorality of cohabitation, while others thought cohabitating people might be generally more carefree and less conventional. Kuperberg (2014) thought that most people begin to cohabit at a young age (e.g., 18 to 22) and marry while still young. In contrast, people who never cohabited tended pick a partner and marry later (e.g., older than 25). Numerous studies document a very strong relationship between young age of marriage and high risk of divorce. Kuperberg thought that the age at which a cohabiting person selects a partner for co- residence might operate similar to an early age of marriage because when young, a person’s incomplete developmental maturity and limited life experience can lead to unstable intimate relationships. She hypothesized that the true cause of a high risk of divorce was entering into a co-residence relationship at a young age, not the cohabitation experience itself. She conducted a multivariate analysis on secondary data from three samples (1995, 2002, and 2006–2010) of the National Survey of Family Growth on U.S. women ages 15–44. The 1995 sample had 10,847 respondents, the 2002 one had 7,643 respondents, and the last sample had 12,279 respondents. Cohabitation questions were first included in the survey in 1995. She examined cohabitation experiences, first marriages, and marriage dissolution as well as many related factors. The key control variable she considered was age of when cohabitation or co-residence began. She examined the relationship between cohabitation and divorce risk using various statistical tests with controls for age of co-residence. She discovered that the widely reported bivariate relationship between cohabitation and divorce was spurious. If a person began coresidence, as cohabitation or marriage, while still young then the risk of divorce was high. If they began co-residence, as cohabitation or marriage, when they were older then the risk of divorce was much lower. In short, whether a person’s first coresidence was in the form of cohabitation or marriage made little difference. The key cause of high divorce risk was age of co-residence. She concluded, “This research also suggests that young couples wishing to avoid divorce would be better

served by delaying settling down and forming co-residential unions until their mid-20s when they are older and more established in their lives, goals, and careers, whether married or not at the time of co-residence, rather than avoiding premarital cohabitation altogether” (p. 368).

9.4.5: Multiple Regression Analysis One of the most widely used statistical techniques for nonexperimental data analysis in professional research reports is multiple regression. It is a powerful technique that takes the idea of statistical control to a higher level. The calculation of multiple regression is beyond the level of this discussion; however, most statistical software can produce multiple regression results, and the results are not difficult to read or interpret. The technique is designed for intervaland ratio-level data, although nominal- or ordinal-level variables can be independent or control variables. The true power of multiple regression is an ability to control for many independent and control variables simultaneously and give precise estimates of the size of each variable’s impact on a dependent variable. What do multiple regression results tell you? Compare Your Thoughts Multiple regression results tell you two things: 1. R-squared (R2) or the percentage of overall prediction accuracy. It indicates reduced errors when predicting the dependent variable based on information from the independent variables. With several independent variables, you might account for, or explain, a large percentage of variation in a dependent variable. For example, an R2 of .50 means that knowing the independent and control variables give a 50 percent accuracy in predicting the dependent variable. In short, you would make half as many prediction errors as you would without knowing the independent variables. An R2 of .20 is considered very strong in professional social science. It means that independent variables explain 20 percent of change in the dependent variable. The R2 size may not seem impressive, but we judge its size not based on 100 percent perfect prediction, rather on improvement over random prediction (i.e., zero). 2. Multiple regression results tell you the direction and numerical size of each independent variable’s impact on a dependent variable. Multiple regression analysis may indicate how five independent variables simultaneously affect a dependent variable, with all variables controlling for the effects of one another. This is especially valuable for testing theories that state that multiple independent variables cause one dependent variable. Effects on a dependent variable are measured indicated by a standardized regression coefficient, symbolized by the Greek letter beta (β). The interpretation of β is similar to a correlation coefficient, r. In fact, for a multiple regression with only two variables, the beta coefficient equals the r correlation coefficient. A beta coefficient tells you the size and direction of effects.

204 Chapter 9 Let’s say you have a dependent variable, family income (ratio level), and four independent variables: age, years of schooling, hours worked, and hours of TV watched each day. Which one is most important for predicting income? You may say schooling, but what about the other variables? Multiple regression results show that schooling, age, and hours worked each have large positive effects. Watching TV has a small and negative effect. Combined, all the independent variables together are 21.5 percent accurate in predicting a person’s family income. Another way to say this is that for U.S. adults, only knowing four facts about them allows you to predict their family income level and be correct one in five times. One in five is a lot better than zero (see Table9.12).

Table 9.12 Multiple Regression of Family Income

(Dependent Variable) with Education Level, Age, Hours Worked per Week, and Hours of TV Watching per Day forAdults Aged 18–65 Independent Variable

Beta Coefficient

Years of schooling

.306

Age

.242

Hours worked Daily hours watch TV

.169 -.064

(Based on GSS, 2004 data)

9.5: How Do You Go beyond Description toInference? 9.5 Evaluate how statistical significance helps determine whether data supports a hypothesis In most studies, you want to move beyond just describing sample data to test hypotheses and say whether results from sample data hold true in a population. You also need to decide whether differences in results (e.g., between the mean scores of two groups) really indicate that a strong relationship among variables exists. Inferential statistics relies on probability theory to enable you to make inferences from a sample to the population, assess the strength of relationships among variables, and test hypotheses. Inferential statistics assumes a random sample and offers a precise way to specify how confident you can be when inferring from sample data to variable relationships in a population. You may have heard about aspects of inferential statistics such as “statistical significance” or results “significant at the .05 level.” These are ways to decide whether to accept or to reject a null hypothesis. There are many

s tatistical “tests” such as a T-test or F-tests that use statistical significance to help determine whether data support a hypothesis with a lot of confidence or little confidence based on statistical criteria.

9.5.1: Statistical Significance Statistical significance indicates the probability of finding a relationship in sample data when there is none in the population. Probability samples involve a random process, so sample results might differ from a population parameter. Using probability theory and specific statistical tests, statistical significance tells you the results could be produced by random error. If results were unlikely to be the results a random process, you gain more confidence that the results are showing a relationship that is really in the population. Statistical significance can only tell you what is likely. It cannot prove anything with absolute certainty. Building on probability theory, it tells us that particular outcomes are more or less probable. Statistical significance is not the same as practical, substantive, or theoretical significance. To determine practical, substantive, or theoretical significance, you use other criteria in addition to statistical processes. A relationship can have statistical significance but be theoretically meaningless or trivial. For example, two variables may have a statistically significant association with no logical connection between them, like the length of fingernails and ability to speak French. The mathematics and calculations of statistical significance is beyond the level of this discussion, but reading and interpreting statistical significance are not difficult.

Figure 9.14 Looking at Statistical Significance Statistical significance: If you discover a relationship among variables in your sample data, what are the chances that such a relationship actually exists in the population? Population (unobserved relationship)

Random Process

Sample Statistic (observed data)

9.5.2: Levels of Significance Statistical software quickly calculates statistical significance and produces probability estimates indicating how likely chance factors alone produced a relationship among variables in the data. Following the logic of a null hypothesis, if a result is probably not due to chance, then we gain confidence that the population has a real relationship.

Making Sense of the Numbers 205

To say that a relationship is statistically significant, and likely to be “real” rather than due to chance factors, the probability that the results are due to random process must be 5percent or smaller. In other words, only if the odds are 5 percent or less that a relationship is due to chance (i.e., a 95 percent or greater chance something other than randomness produced it) will we call a relationship statistically significant. Statistical software generates specific probabilities, but researchers simplify things and express statistical significance as one of a few levels (e.g., a test is statistically significant at a specific level) rather than stating a specific probability. The level of statistical significance tells you the likelihood that results are due to chance factors—that is, that a relationship appears in the sample when there is none in the population. If results are statistically significant, it is at one of three levels: .05 level, .01 level, or .001 level. To say that results are significant at the 0.05 level means all of the following: • Results like these are due to chance factors only 5 in 100 times. • There is a 95 percent chance that the sample results are not due to chance factors alone but reflect the population accurately. • The odds of such results based on chance alone are .05 or 5 percent. • One can be 95 percent confident that the results are due to a real relationship in the population, not chance factors. These all say the same thing in different ways. They may sound like a discussion of sampling distributions. This is no accident. Both use probability theory. In both, you link sample data to a population. Probability theory lets us predict what is likely to happen in the long run over many events. It does not predict for a specific situation. We cannot know for certain that a relationship we find in a particular sample is what occurs in the population. However, we can state it in probability terms—how likely it is that the sample shows one thing whereas something else is true in the population. You may ask: Why use the .05 level? It means a 5 percent chance that randomness could cause the results. Why not use a more certain standard—for example, 1 in 1,000? This gives a smaller chance that randomness versus a true relationship caused the results. The simple answer is that the scientific community has agreed to use .05 as a pragmatic “rule of thumb” for most purposes. Being 95 percent confident of results is an accepted standard for explaining the social world. It is “good enough” to take seriously, although in some cases (e.g., the chances of a serious side effect for a new drug) you may want to be surer that results are not due to chance.

Making It Practical: Is There a Significant Difference by Gender?

Let us say you have survey data on a random sample of 1,000 college students showing that male and female students differ in the number of hours they study. The data show that men study an average of 14.5 hours a week and women 15.9 hours per week. Is the gender difference due to chance factors in the sample, or is it real? If you had the entire population, would it show no difference, or is there a true gender difference in the population? One way to find out is to conduct a statistical test of differences between two means (called a t-test) and check its statistical significance. If the difference between means (based on sample size, degree of variability, and the size of the difference) is statistically significant at the .05 level or higher, you can say with some confidence a real gender difference is likely in the population. Recall the previous discussion of age and red car preference. Statistical significance also tells you how likely it is that the measure of association between age and red car ownership, gamma, is due to chance factors alone. In that example, the gamma was statistically significant at the .001 level. This tells you that you can be 99.9 percent confident that the age difference in red car owners in the population is real and is not just due to chance factors.

WRITING PROMPT Confidence Intervals and Statistical Significance You learned about confidence intervals in the chapter on sampling. Recall that it means you create a zone or interval around a specific number and can have far greater confidence (e.g., 95 percent) saying that the true population parameter lies within the interval rather than being at a specific number or point. With the statistical significance, we again express estimates using the language of amount of confidence. How do you think the ideas of confidence interval and statistical significance relate to each other? What do they have in common? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

206 Chapter 9

Table 9.13 Type I and Type II Errors The rule of thumb of a .05 level is a compromise between Type I and Type II errors. What Is Actually Happening in the Population What You Say Is Happening Based on Your Sample Data

No Relationship

A Relationship Exists

No Relationship

No Error

Type II Error (false negative)

A Relationship Exists

Type I Error (false positive)

No Error

Accuracy in Inferential Statistics We face a

logical dilemma when we try to be as accurate as possible in inferential statistics. If we demand a very high standard, such as 1 in 1,000 chance of being wrong before we state that a relationship exists in the population, we are more likely to make the opposite type of mistake—saying there is no relationship in the population when in reality one exists. We can make two types of mistakes or errors: false positives and false negatives. A false positive occurs when we say something is true or present when in reality it is not. A false negative is saying something is false or not there when in reality it is. Ideally, we do not want to make either type of mistake. The dilemma is that as we reduce the odds of avoiding a false positive, we increase the odds of a false negative. In other words, there is a trade-off. This dilemma with false positives and false negative appears outside of the research settings. In fact, the term false positive comes from medical settings. Let us say you go in for a cancer test, it can be accurate, a false positive, or a false negative. A false positive is a mistake saying you have cancer but you really do not. A false negative is a test that says you are cancer free but in reality you have cancer. Another example is a jury trial. A jury can err by deciding that an accused person is guilty when in fact he or she is innocent (false positive), or decide that a person is innocent when in fact he or she is guilty (false negative). The medical field and jury do not want to make either error. We do not want to tell a person that she has cancer, when she does not, or that she is cancer free when she has it. Likewise, we do not want to jail the innocent or to free the guilty. Type I and Type II Errors We can combine the ideas

of statistical significance and the two types of error (false positive or false negative). If you are an overly cautious person and set a very high level of significance (.0001), you are likely to make a false negative error—i.e., you say no relationship exists but there is really one. In short, you falsely accept the null hypothesis when there really is a relationship in the population. The formal name for this is a Type II error. By contrast, you may be a risk-taker and set a low level of significance, such as .10. You might say there is a relationship in the data, even if chance factors alone could produce it 1 in 10 times. You are likely to make a false negative error; that is, say that a causal relationship

exists when in reality it is being caused by random chance (e.g., random sampling error). In this case, you are falsely rejecting the null hypothesis when there is no relationship. The formal name for this is a Type I error (see Table 9.13). In research we use inferential statistics and levels of significance to balance avoiding both types of errors. Inferential statistics allow you to say that a relationship is present in a population with specific degrees of certainty. If a relationship is statistically significant at the .05 level, it is saying that sample results are not likely to be due to chance factors. Indeed, there is a 95 percent chance that the population (or the social world to which you want to generalize) contains the relationship. Remember, inferential statistics assume the data are from a random sample. Another limitation is that nonsampling errors (e.g., a poor sampling frame or a poorly designed measure) are not considered. Do not be fooled into thinking that statistical tests offer simple, final answers. Most statistical software can produce inferential and descriptive statistics. Making It Practical: Statistical Software User Beware Almost every social researcher calculates

statistics with statistical software. Even basic spreadsheet programs, such as Microsoft Excel® produce some statistics. Unfortunately, spreadsheets were designed for accounting and bookkeeping functions. Their statistics tend to be clumsy and limited. The marketplace of statistics software can be confusing to a beginner, for products evolve rapidly with changing computer technology and market forces. Overt time the software has become easier to use. The most popular programs in the social sciences are Minitab®, Microcase®, and SPSS® (Statistical Package for the Social Sciences). Others include SAS® (Statistical Analysis System), STATISTICA® by StratSoft, and Strata®. Many began as simple, low-cost programs for research purposes. SPSS is the most widely used statistics program in the social sciences. Its advantages are that social researchers used it extensively for about five decades, it includes many ways to manipulate quantitative data, and it has most statistical measures. A disadvantage is the time it takes to learn the many options and complex statistics. Also, it can be expensive to purchase outside a discount program. For example, the SPSS Grad Pack sold for $4,600 in 2015 with two-year license, but a basic student version of the same software sold for $110.

Making Sense of the Numbers 207

As statistics software has made it quick and easy to use a statistics program, people who do not understand the basics of statistics can now use the software. If someone does not really understand the meaning of statistics, he or she might violate basic assumptions required by a statistical procedure, and use the statistics improperly. The results produced might be total nonsense yet look very technically sophisticated. To avoid such mistakes, you need to have one or more courses in statistics or get assistance from someone who knows the meaning of the statistics than the implications of the results.

WRITING PROMPT Making Errors How is the “trade-off” between Type I and Type II errors in statistics similar to dilemmas outside of the world of research? Provide an example other than the ones given in the text. The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

Summary: What You Learned about Making Sense of the Numbers You learned about organizing quantitative data for analysis. You also learned how to organize data into charts or tables or summarize them with statistical measures. You can describe results, test hypotheses, and answer research questions with univariate or bivariate statistics. Bivariate analysis allows you to test causal hypotheses, but because bivariate relationships might be spurious, you may wish to use control variables and multivariate analysis. You also learned some basics about inferential statistics that tell you whether a relationship is statistically significant or not. Beginning researchers sometimes feel their results must support a hypothesis. Keep in mind that there is nothing wrong with rejecting a hypothesis. The goal of scientific research is to produce knowledge that truly reflects the social world, not to defend an idea or hypothesis. Hypotheses are predictions about relationships based on limited knowledge, so we need to test them. Excellent-quality research can find that a hypothesis is wrong, and poor-quality research can support a hypothesis. Good research depends on high-quality methodology, not on supporting a specific hypothesis. As you do research, you want to avoid making errors about inferences from data to the social world. Errors can enter into the research process and affect results at many places: research design, measurement, data collection, coding, calculating statistics and constructing tables, or interpreting results. Even if you design, measure, collect, code, and calculate without error, you must properly interpret the tables, charts, and statistics. Finally, you need to answer the question: What does it all mean?

The only way to assign meaning to facts, tables, and statistics is to use some type of theory. You can answer few research questions by looking at data, tables, or computer

output alone. The facts do not speak for themselves. As a researcher, you must return to theory, concepts, relationships, assumptions, and theoretical definitions to give the data results substantive meaning. Before we leave our discussion of quantitative research, there is one last issue. Journalists, politicians, and others increasingly use statistical results to make a point or bolster an argument. This has not produced greater accuracy and information in public debate. More often, it has increased confusion and made knowing what statistics can and cannot do even more important. The cliché that you can prove anything with statistics is false. Yet people can and do misuse statistics. Through ignorance or conscious deceit, some people use statistics to manipulate others. The best way to protect yourself from being misled by statistics is not to ignore them but to understand the research process and statistics, think about what you hear, and ask questions. We turn next to qualitative research. The logic and purpose of qualitative research differ from those of the quantitative approach of the past chapters. Qualitative research is less concerned with numbers, hypotheses, and causality and more concerned with words, norms and values, and meaning.

Quick Review How Do You Prepare Data forStatistical Analysis? 1. When collecting data yourself (i.e., not secondary data), you need to reorganize the data from a raw form into a machine-readable format for statistical software.

208 Chapter 9 2. You usually need to code raw data and put information for all variable categories or levels into the form of numbers, with all variable information on each case, unit or participant a separate data record. 3. You need to create a codebook to keep a written record of all coding procedures and decisions, and should assign a unique identification number to each data record. 4. Machine-readable data ready for statistical software are usually in the form of a grid of numbers. You organize variable information into the columns. The rows are data records, one for each unit or case. 5. After you coded and entered all data into statistical software, you need to check and clean the data before proceeding to statistical analysis.

Univariate Statistics: Looking atResults with One Variable 1. Univariate statistics refer to descriptive statistical techniques that examine characteristics of one variable. 2. You can display the values of one variable using a frequency distribution or visual graphics with a pie chart or bar chart. 3. Often we wish to summarize all the information of a variable into a single number, a measure of central tendency, sometimes called the average. Three measures of central tendency are mean, median, and mode. Each has its specific uses. 4. The mean is the most widely used measure of central tendency, but it requires ratio-level data and is most influenced by extremely high or low values in the distribution of a variable, i.e., a skewed distribution. 5. In addition to measuring the middle or center of a variable’s values, it is just as important to indicate the variability or spread of values in a variable around the center. The standard deviation is the most common way to indicate the spread of a variable’s values around its mean. 6. Z-scores allow us to standardize information about the center and variability of a variable. It is very useful when comparing two or more variables that differ with regard to their mean and standard deviation. 7. In addition to statistical measures of one variable, it is also helpful to use maps and timelines to display information on a variable. While the statistical measures available for nominal- and ordinal-level variables are limited, maps and timelines can display information for variables at any level of measurement.

Bivariate Statistics: Looking at the Results for Two Variables Together 1. To test or examine a causal relationship, you need to examine two variables together, one as the cause

(independent variable) and one as the effect (dependent variable). 2. Bivariate statistical relationships rest the idea of covariation and statistical independence. 3. Covariation means to vary together and shows whether the cases with values on one variable also have certain values on a second variable in a patterned way. Independence means there is no relationship between the two variables. 4. A scattergram is a graph that displays bivariate relationships in interval-level or ratio-level data with the independent variable on the horizontal axis and the dependent variable on the vertical axis, organized with the lowest value for each in the lower left corner and the highest value at the top and to the far right. 5. A scattergram shows the form, direction, and precision of relationships in data. The three forms are independence, linear, or curvilinear. With independence the plotted data points appear to be random scatter without any pattern. With a linear relationship, the plotted data points align in a straight line that goes from one corner to another. For a curvilinear relationship, the pattern of data points forms a U curve, right side up or upside down, or an S curve. Linear relationships have either a positive or negative direction. A positive relationship is a diagonal line of data points from the lower left to the upper right, while for negative relationship looks the points go from the upper left to the lower right. In a very precise relationship, all the data points align tightly along the line that indicates the relationship and direction, while widely scattered data points indicate little precision. 6. Bivariate tables can work with variables having few categories (two to five) and data at the nominal or ordinal level of measurement. The tables are created using a process of cross-tabulation, which means organizing data into a table with two variables. Tables show data where categories of two variables cross intersect, similar to plotting data points on two variables in a scattergram. 7. Bivariate tables reveal patterns in the data that we can see by comparing the amount of data in some cells relative to others in the table. We nearly always convert the raw count table into a percentage table because it is easy to be misled by a raw count table when the row or column totals of variable categories are different sizes. Using percentages adjusts for unequal totals and standardizes information to permit easy comparison. 8. Tables can be percentaged by rows or columns. A hypothesis guides you to look at row or column percentages. Relationships are revealed by comparing the size of cells in a table, and a linear relationship appears when the largest cells are along a diagonal in a table.

Making Sense of the Numbers 209 9. Several measures of association indicate the strength of a bivariate relationship with a single number. In most measures of association a value of zero indicates statistical independence, and a large nonzero number, it indicates a relationship, with a larger number indicating a stronger relationship. 10. Showing an association between two variables is not sufficient to say that an independent variable causes a dependent variable. Unless possible alternative causes have been removed, relationship in the data could be spurious. 11. You must measure possible alternative causes as control variables, and examine a bivariate relationship along with control variables to determine whether the bivariate relationship is spurious. 12. Multiple regression is a powerful technique for statistical control but is designed for interval- and ratioleveldata. Multiple regression has an ability to control for many independent and control variables simultaneously. 13. Multiple regression can indicate the overall predictive power of a set of independent and control variables on a dependent variable and provide the direction and size of each variable’s impact on the dependent variable.

Shared Writing: Meaningful StatisticalSignificance Many people treat the term “statistically significant” as a “magic bullet” that provides a simple answer to a complex issue. A relationship can be statistically significant but trivial or meaningless. Statistical significance might come from a statistics software program in which the data fail to meet the logical assumptions required for valid results. Discuss how the common view that statistical significance in scientific research provides fixed, unquestionable answers compares with what statistical significance actually means. Does the rejection of an oversimplified view of statistical significance by many people mean that determining statistical significance is worthless or unimportant? Why or why not? Read through your classmates’ responses and list at least three of the best supporting evidence/arguments provided by other students who are opposing your position. How did the opposing arguments impact your own view? A minimum number of characters is required to post and earn points. After posting, your response can be viewed by your class and instructor, and you can participate in the class discussion.

Post

0 characters | 140 minimum

Chapter 10

Observing People in Natural Settings

Learning Objectives 10.1 Give examples of some of the areas where

field research is effective 10.2 Describe how naturalism is an integral part

of field research 10.3 Outline the progression of stages in field

research 10.4 Apply the five parts of ongoing observation

and research in the field Many occupations require “emotion work” of employees. The “work” is labor or effort you expend, while the “emotion” refers to feelings, or at least a public display of inner feelings. In emotion work, you are expected to show certain feelings as part of a job. Since you must outwardly display certain emotions, you may have to deny, suppress, or

210

10.5 Analyze ways to generate good qualitative

field data 10.6 Apply strategies for planning the exiting

process for a professional field researcher 10.7 Describe how a researcher resolves the ethical

issues that come up during field research 10.8 Analyze the advantages and limitations of

focus groups hide your true feelings. In the travel, hospitality, and entertainment industries; in nursing, social service, and counseling fields; in grief counseling and funeral services; and in many retail sales jobs, the display of emotions is a large part of “doing a good job.” Emotions create an atmosphere and help others experience certain feelings. Emotion work

Observing People in Natural Settings 211

does not always require projecting “warm” emotions of joy or care, it can also demand displays of stern authority, seriousness, and controlled threat in prisons, policing, or border control (see Crawley, 2004). In most jobs, emotion work requires you to appear cheerful, friendly, positive, upbeat, or to be empathetic, concerned, and caring. Some employers have written rules about emotion work, and supervisors may monitor and enforce proper emotional displays/attitudes. Of course, we may do a kind of emotion work on our own as we negotiate with friends or intimate partners, get along with coworkers, or simply take part in daily events. Emotion work that entirely voluntary may differ from when it is a required part of a job. In the Managed Heart, Arlie Hochschild (1983) outlined emotional labor and described how she studied it among airline attendants. She and others learned that employees put forth serious effort to display a friendly smile and cheerful manner. An employee must do this even when feeling tired or depressed or is thinking about a serious personal, family, or work problem. Emotion work drains people physically and emotionally. Acting pleasant, warm, and comforting toward demanding clients or customers can extract a personal toll, especially when the customers or clients are hostile, rude, or obnoxious. Some businesses, such as Disneyland, carefully plan, manage, and coordinate staff members’ emotion work as a central part of creating a total customer experience (Van Maanan, 1991). To study emotion work, researchers gather first-hand qualitative information by directly participating and observing in natural work settings (i.e., they performed fieldwork and conducted in-depth informal interviews). Often they got a job that required them to perform emotion work themselves. They spent many hours in daily contact, both closely observing and talking with people whose jobs involved a lot of emotion work. The field research method is especially well suited for studying emotion work.

If you have heard about field research, it may sound as if you simply hang out with a group of people who are unlike you and it is easy. There are no statistics or abstract deductive hypotheses. You observe a setting and chat informally with the people you are studying. However, unlike quantitative research, you must personally engage in direct, face-to-face social interaction with real people in a natural setting. Professional field researchers often devote months or years to studying people in a setting. They learn about the daily activities, life histories, hobbies and interests, habits, hopes, fears, and dreams of the people they study. Meeting new people and discovering new social worlds can be fun, but it can also be tedious, time-consuming, emotionally draining, and sometimes physically dangerous. A good field researcher combines great empathy and interpersonal skills with superb attention to detail, plus he or she has an ability compartmentalize personal views. This enables the researcher to see with greater clarity all that is occurring in complex social settings or interactions. Field research is ideal for examining the micro-level social world and everyday interactions of people up close. It is well suited for studying issues such as student classroom behaviors, street gangs, retail customer perceptions, or patient reactions that maybe difficult to study using methods like surveys or experiments. In field research, you can get a close-up look and develop insights that may be impossible to achieve using other methods. Many areas use field research, such as anthropology, education, health care, marketing, hospitality-tourism, human service delivery, recreation, and criminal justice (see Figure 10.1). Students in my classes conducted successful shortterm, small-scale field research studies in a barbershop, beauty parlor, daycare center, bakery, bingo parlor, bowling alley, church, coffee shop, laundromat, police dispatch office, fast food restaurant, elementary school classroom, nursing home, tattoo parlor, gay bar, bridal shop, hospital waiting room, and athletic weight room, among other places.

10.1: What Is Field Research?

10.1.1: Ethnography

10.1 Give examples of some of the areas where field research is effective Unlike the quantitative research, field research involves a distinct approach to research and yields qualitative data. Field researchers directly observe and participate in a natural social setting. Actually, there are several kinds of field research, including ethnography, participant-observation research, informal “depth” interviews, and focus groups.

Researchers use ethnography to construct a highly detailed description and up-close understanding of a way of life from the standpoint of its natives/members/insiders. Ethnography builds on the idea that humans live in cultures, both micro-cultures (a single family or small friendship group) and macro-cultures (entire societies or world regions). We continuously learn, use, and modify cultural knowledge. Cultural knowledge includes familiarity with assumptions, symbols, sayings, facts that “everyone knows,” ways of behaving, and artifacts (cell phones, popular drinks, etc.). People learn it by reading, watching films or television, listening to parents or friends,

212 Chapter 10

Figure 10.1 Examples of Field Research Sites/Topics Field researchers have explored a wide range of social settings, subcultures, and aspects of social life. Small-Scale Settings

Nude beaches

Passengers in an airplane

Occult groups

Battered women’s shelters

Prostitutes

Camera clubs

Street gangs, motorcycle gangs

Laundromats

Street people, homeless shelters

Social movement organizations Social welfare offices

Community Settings

Television stations

Retirement communities

Homeless shelters

Small towns

Waiting rooms

Urban ethnic communities Working-class neighborhoods

Occupations Airline attendants

Children’s Activities

Artists

Children’s playgrounds

co*cktail waitresses

Little League baseball

Dog catchers

Youth in schools

Door-to-door salespersons

Junior high girl groups

Factory workers Gamblers

Medical Settings and Medical Events

Medical students

Death

Female strippers

Emergency rooms

Police officers

Intensive care units

Restaurant chefs

Pregnancy and abortion

Social workers

Support groups for Alzheimer’s caregivers

Taxi drivers Leisure Industry Social Deviance and Criminal Activity

Bowling alleys

Body/genital piercing and branding

Hotel lobbies

Cults

Restaurants and bars

Drug dealers and addicts

Resorts

Hippies

observing others, and the like. Cultural knowledge is both explicit and tacit. Explicit knowledge is what we easily see and directly know. It is written down and talked about. Tacit knowledge remains unseen or unstated. We rarely acknowledge it and are often only indirectly aware of it. Tacit knowledge includes unspoken cultural norms. It requires making inferences (going beyond what a person explicitly does or says to what is meant or implied). For example, smoothly conforming to the proper etiquette and polite behavior depends on possessing proper tacit knowledge. Ethnography is a study of how people activate and display culture—what they think, ponder, value, or believe to be true. People do this by what they say and do in specific natural settings. To figure out or infer the meaning of people’s actions, you need to become very familiar with the cultural and social context. You must quickly shift from what you actually hear or observe to an action’s meaning in a situation. For example, an adult brother and sister are

sitting next to one another at a family Thanksgiving meal in the United States. During the entire event, they never exchange a single word. Observing this, you might infer cool social relations between them. Often people only become aware of tacit knowledge, such as the proper distance to stand from others, when it is violated. The violation of social custom or practice can cause unease or discomfort, but people may find it difficult to pinpoint the source of unease. A good ethnographer describes both explicit and tacit cultural knowledge with highly detailed accounts. He or she also analyzes social situations by taking them apart and reassembling them. Recognizing explicit and tacit levels of knowledge, recording detailed descriptions, and analyzing actions in a specific setting are key ethnographic techniques. Often an ethnographer pays great attention to specific details and may consider events by very attentive silence or by entering the flow of ordinary interaction. An ethnographer notes how people construct social meanings as ongoing processes in specific settings. He or

Observing People in Natural Settings 213

she looks for the “unwritten scripts” or routines. Since social activities and events unfold in real time, they are not always predictable. Not everyone in a setting sees everything identically. The ethnographer tries to grasp multiple perspectives in a setting by constantly switching perspectives in order to see activities from several points of view simultaneously. This is difficult to do at first. It is a skill that most people can learn and improve upon with practice.

WRITING PROMPT Tacit Knowledge Describe a social setting and identify 2–3 parts that are tacit and 2–3 others that are explicit. When do tactic features become explicit? Can you think of a situation when an explicit feature later becomes tacit? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

Example Study Ethnographic Inference in a Hotel

Rachel Sherman (2006) conducted field research on luxury hotels (room rates in the range of $600 and up per night). She studied emotion work, workplace relations, and service provided by hotel employees to the wealthy hotel guests. Two luxury hotels in New York were her field sites. She interviewed, observed, and got a job working at the hotels. In the study, she noticed many hundreds of tacit and subtle exchanges among hotel guests and workers. For example, she noticed a service code for employees to provide “unlimited labor” to hotel guests. The employee was supposed to perform the labor in a way that made it appear to be a voluntary favor that he or she wanted to do. A tacit rule was that workers should display emotions of being sincere and truly caring about a guest’s needs or desires. She gave an example of when she worked as a housekeeper (2006, p. 42). Hotel guests asked her to wrap up a large bouquet of flowers for them to take home. She quickly smiled and responded to the request, “I’ll deal with it.” The guest repeated the phrase “deal with it.” She instantly realized that she made a mistake. She should have expressed her joy at the task. The words deal with it implied that it might be undesired work. A proper response would have been to smile and say, “I’d be happy to wrap the flowers for you right away, sir.”

MAKING IT PRACTICAL: SEEING CULTURE IN THE AMERICAN THANKSGIVING HOLIDAY

Ethnographers study social gatherings and socio-cultural events, like the U.S. Thanksgiving holiday. Thanksgiving is part of explicit knowledge. It is on a Thursday in late November. Most businesses and all schools close on the day. It has some visible symbols and traditions. People say it is a celebration of abundance and dates back to the early settlers. The central focus of this holiday is people joining with family members or close friends to share a large feast that usually includes eating roast turkey. Related activities include parades with floats in some large cities, football games on television, the arrival of Santa Claus for children, and the beginning of Christmas holiday shopping. Tacit knowledge at the Thanksgiving meal, says to use forks and knives when eating turkey meat, expect to see common dishes and make comments on their appearance, aroma or flavor, wait until finishing the main meal before starting dessert, eat with others and not off alone in a corner, and converse with others at the meal. Eating large quantities or overeating is expected behavior. The timing of the main feast varies by family tradition, but it is rarely very early in the morning, as a breakfast, or extremely late at night after 10 p.m. Most people eat Thanksgiving dinner between noon and 7 p.m. It is not the same as ordinary dinner. In addition to the turkey, other foods rarely seen on other occasions may be served. Pumpkin pie is the most common dessert. The common color scheme is orange and brown. Often there are

214 Chapter 10 few seasonal decorations (e.g., autumn leaves, pilgrims) but they are usually minor. Gift exchanges are rare. In some rural parts of the United States, men hunt deer or other wild animals on or around Thanksgiving Day.

WRITING PROMPT Holiday Events The social activities and multiple events surrounding special occasions, such as holidays, offer observation opportunities for fieldresearchers because they involve a complex configuration ofcustoms and practices that “everyone knows.” Actually, only people who share the same culture may “know” what to expect but events may appear strange and confusing to anyone unfamiliar with the culture. Describe a holiday or other special social occasion in which you participated that depends on many people all knowing what “should happen” although the informal rules are not made explicit. The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

pportunities, “play it by ear,” and adjust to fluid, chango ing situations. Instead of starting with a hypothesis and then following a fixed sequence of steps rigidly, a field researcher selects techniques based on their value for acquiring information in a specific setting and will change direction to follow interesting new leads as they appear. Flexibility has a down side. Without fixed steps, it is easy to become sidetracked and drift directionless without a focus. Beginning field researchers often say they have little control over events, need focus, and that data collection is unmanageable. The research task can feel overwhelming. As a new field researcher, you may think you are collecting too little, too much, or the wrong data and feel emotional turmoil, isolation, and confusion. This makes it essential to be organized and prepared.

10.3: How Do You Begin to Conduct Field Research? 10.3 Outline the progression of stages in field research

10.2: How Do You Study People in the Field? 10.2 Describe how naturalism is an integral part of field research Field research is as much an orientation toward doing research as a specific set of techniques to apply. Field researchers use a wide variety of techniques in the field to gain information. A principle guiding field research is naturalism. Unlike quantitative research in safe, controlled settings such as an office, laboratory, or classroom, you conduct field research in natural settings. It requires your personal involvement in the less predictable real world and you become a part of the social world you study. If you are a beginning field researcher, you will want to start with a relatively small group (20 or fewer people) who regularly interact with each other in a specific place (a street corner, church, barroom, beauty parlor, baseball field, etc.). Most field research is conducted by one person alone, although sometimes small teams have been effective. Direct, personal involvement in the field can have an emotional impact. Field research can be fun and exciting, but it may disrupt your personal life, unsettle your physical security, or strain your sense of well-being. More than other kinds of social research, field research can reshape your friendships, family life, self-identity, and personal outlook. Direct, personal involvement carries both risks and rewards. Flexibility is another principle of field research. Field research is less structured than quantitative research is. Field researchers know how to recognize and seize

After you learn the basics of a field setting, you slowly develop a specific focus for the inquiry and move through a progression of stages. As your focus develops, it will guide you to the qualitative data you need.

Example Study Two Field Research Studies ofWork in Nursing Homes

Observing People in Natural Settings 215 Rodriquez (2011) conducted field research in two for-profit nursing homes that were part of large multi-facility corporations. Most care workers are low-paid middle-aged females with limited training. He visited the facilities two or three times per week for a year, interviewed staff, observed a variety of settings, and gathered data about the character and scope of care work. He spent time in and around nursing stations, shadowed nursing assistants and licensed nursing staff in and out of residents’ rooms, and occasionally helped with serving meals or escorting residents. He observed staff in breakrooms, at holiday parties and other staff functions, and during lunch. He also spent time at the outdoor “butt hut” where the staff took smoking breaks. He recorded his observations in detailed fieldnotes written soon after he left the facility. He also conducted semi-structured interviews with the staff members, including certified nursing assistants, nurses, physical and occupational therapists, social workers, activities assistants, unit managers, directors of nursing, and the administrators. Rodriquez asked everyone a core set of questions, but tailored interviews to each individual, occupational group, or organization. He asked about daily job tasks, work history, emotional bonds created with residents, how the documentation and reimbursem*nt process shaped work, challenges at work, and feelings about caring daily for people whom they expect to die soon. Rodriquez learned that despite organizational demands and a profit drive that often blocked staff attempts to forge strong emotional connections with residents, the care workers formed emotional attachments to residents. Managers tried to organize and control the use of emotions, viewing them as a resource to exact compliance from residents. Yet, the low level care workers would push back and recast emotions as a strategy through which they gained workplace autonomy and created dignity in the “dirty work” of caring for others during the deeply personal and intimate exchanges that occur daily for months or even years. He discovered that care workers at times produced emotions in ways consistent with organizational demands, but at other times, they did so in ways inconsistent with those demands. The staff regularly used emotions to manage the strains created by their work and to achieve a sense of dignity and pride in what they had accomplished.

Read More In his field research at a nonprofit nursing home with a religious affiliation, “Heartland Community,” Lopez (2007) spent 1,000 hours in participant observation over 9 months. At the facility, the overwhelming majority of staff were female and black, about one-fourth recent immigrants from West Africa or the Caribbean. Lopez also undertook formal training and worked as a part-time as nurse’s aide over the last five months. He revealed his researcher role to all in the field and recorded observations in detailed notes at the end of each day. Lopez focused on workplace relations. He noted the emotional dimension of care work—washing, bathing, dressing, and feeding frail elderly, with staff worrying about ill residents or even attending the funerals of residents who died on their days off. His primary interest was how the staff found it impossible to accomplish all the necessary care tasks in the allotted

amount of time while following the official procedures. The care workers routinely broke rules to get the work done. As he stated, “No one ever followed procedures for washing residents, which involved two washbasins. . . . Rules against leaving residents on the toilet alone were also routinely violated: no one could afford to waste five or ten minutes waiting . . .” (p. 235). In what he termed mock routinization, there was a systematic mismatch between official work routines and actual work practices. The care staff informally acquired a repertoire of skills, many in violation of official procedures, to complete tasks in allotted time and provide an acceptable standard of care. The management made “bad faith” public statements about the importance of following official procedures for health and safety of residents and staff, yet enforced time demands that made it impossible for staff to get the work done had they adhered to official procedures.

MAKING IT PRACTICAL: THE PROCESS OF DOING FIELD RESEARCH Field research does not proceed in a

set of fixed steps, so this outline of stages is only a general guide based on how successful field researchers have proceeded. Stage 1 • Preparing for a field research study • Self-awareness • Background investigation • Practice observing and writing Stage 2 • Starting the research project • Getting organized • Selecting a site • Gaining access • Entering the field • Presentation of self • Amount of disclosure • Selecting a social role Stage 3 • Being in the field • Learning the ropes • Building rapport • Negotiating continuously • Deciding on a degree of involvement Stage 4 • Developing strategies for success in the field • Building relationships • Performing small favors • Being inept but accepted • Avoiding conflict • Adopting an attitude of strangeness

216 Chapter 10 Stage 5 • Observing and taking field notes • The researcher as instrument • What to observe • Physical setting • People • Routines, events, and activities • Learning to listen • Using Serendipity • Sampling • Becoming a skillful notetaker • Types of notes • Supplements Stage 6 Conducting field interviews Stage 7 Leaving the field Stage 8 Writing the field research report

10.3.1: Preparing for a Field Study We can divide the early preparation for doing a field study into three parts. SELF-AWARENESS Human and personal factors have a

role in all social research projects, but they are crucial in field research. Direct, personal involvement with other people means your emotional make-up, personal biography, and cultural experiences are highly relevant. There is no room for self-deception. Your personal characteristics (including your physical appearance, gender, age, racialethnic background) are often relevant in the research process. You need to be candid and have a solid understanding of who you are, because you are in the middle of real, ongoing events, not safely hidden away in a laboratory, office, or computer room. Good field researchers have a well-developed sense of self but are not self-absorbed. You must have an ability to notice details around you and experience empathy for others. You must also be aware of your own personal concerns, commitments, and inner conflicts. Expect anxiety, self-doubt, frustration, and uncertainty in the field, especially in the beginning. You may feel doubly marginal: an outsider in the field setting, and increasingly distant from friends, family, and professional researchers. BACKGROUND INVESTIGATION As with all research,

reading the scholarly literature helps you learn useful concepts, potential pitfalls, data collection techniques, and strategies. Field researchers also use films, novels, or journalistic accounts about the type of field site or topic. They read autobiographies or diaries of people similar to those in the field site to gain familiarity and prepare themselves emotionally. More than a standard literature review, field researchers conduct a wide-ranging background investigation.

PRACTICE OBSERVING AND WRITING Field research depends on skills of careful observing and listening, shortterm memory, and writing. A good field researcher is observant and notices many details. He or she can also “pull back” to see the whole and grasp what occurs “between the lines.” Much of what a field researcher does is to notice the ordinary details of situations and write them down. Attention to subtle details and short-term memory can improve with practice. Before you enter a field site, practice and refine your observation skills. Many people find that keeping a daily diary is good practice for writing field notes. Beyond personal strength and strong social skills, a field researcher needs to be a compulsive, organized note taker.

WRITING PROMPT Becoming Self-Aware Explain why “knowing thy self,” being self-aware, is important in certain types of field sites. What places would you be most comfortable and least comfortable studying as field sites? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

10.3.2: Starting the Research Project We can divide the process of starting the research project into four parts. GETTING ORGANIZED Adjusting your mindset and

attitude is an essential aspect of field research preparation. To begin, you defocus by being open-minded and clearing your mind of preconceptions. Do not begin with a fixed set of ideas, stereotypes, or notions about the topic, field site members, or field setting. Be open to what you actually learn in the field, rather than imposing preconceptions. You need to maintain a balance between being attentive and informed yet open to new experiences or ideas. You begin the field research process with a general topic, not with a narrow question or specific hypotheses. Find a balance: Avoid locking yourself into a narrow focus quickly, but do not become directionless. It takes time in the field and effort to develop research questions that fit a field situation. Often, you can only develop questions after you know more about a site or situation. Patience is another valuable field research skill. SELECTING A FIELD SITE Field projects often begin with a chance occurrence or personal interest. Many studies began with a researcher’s experiences, such as working at a job, having a hobby, or being a patient, client or an activist. The field site is more than a single, fixed physical

Observing People in Natural Settings 217

location. It is a socially defined territory with fluid and shifting boundaries because social groups can interact across several physical sites. For example, a college football team interacts on the playing field, in the locker room, in a dormitory, at a training camp, or at a local hangout. The field site to study a football team includes all five physical locations. You need to recognize the interdependence among the physical sites, the topic, the interacting members, and you, the researcher. Selecting a field site is a major decision. Researchers often try several sites before settling on one and take notes on the site selection processes. What factors affect your choice of a field research site? Compare Your Thoughts • Containment. It is easier to study field sites in which small groups engage in sustained social interaction within a bounded space. Field research in large, open public spaces where many strangers pass through with little social interaction (such as a shopping mall, large discount store, large outdoor park or parking lot) is much more difficult. • Richness. The most interesting data come from sites that have overlapping webs of social relations among people with a constant flow of activities and diverse events. • Unfamiliarity. Despite some initial unease, you can more quickly see cultural events and relationships in a setting that are new to you (see “Acquiring an Attitude of Strangeness” later in this chapter). • Suitability. Consider practical issues, such as your personal characteristics and feelings. Also important are your time and skills, your physical safety, ethical protections, conflicts in the site, and your ability to gain access.

GAINING ACCESS Personal characteristics, such as age,

gender, race, and physical appearance, can facilitate or limit access. You may find that you are unwelcome or not allowed on a site, or that you face legal and political barriers to entry. We can array field sites on a continuum. At one end are open access public areas (public restaurants or check-in airport waiting areas) and at the other end are closed, private, or semiprivate settings (exclusive clubs, activities in a person’s home). Laws and regulations in public and private institutions like public schools, hospitals, and prisons can restrict access. In addition, institutional review boards may limit field research on ethical grounds. Almost all field sites have a gatekeeper. The gatekeeper can be the thug on the corner, an administrator of a hospital, a manager of a restaurant, or the owner of a beauty parlor. Even open public areas often have informal gatekeepers. Formal organizations have authorities from whom you must obtain explicit permission. Whether or not it is required, it is good practice to identify and ask permission of gatekeepers.

Expect to negotiate with gatekeepers and bargain for access. In addition to being flexible, a field researcher sets nonnegotiable limits to protect research integrity. For example, a gatekeeper who demands you only say positive things or insists on reading and censoring field notes makes research at a site impossible. If there are many restrictions initially, often you can reopen negotiations later. Gatekeepers may forget their initial demands as trust develops. Dealing with gatekeepers is a recurrent issue as you enter new levels or areas of the social setting.

Example Study Negotiating with Gatekeepers In her study of luxury hotels, Sherman (2006) described gaining access to two five-star hotels. At one hotel, she received permission from the general manager and human resources manager. When she was assigned to be an intern at that hotel, she had to work with other gatekeepers who were managers or supervisors in various areas (such as guest services, “doorman” and parking, front desk). She found that gaining access from a gatekeeper at one level or area did not automatically transfer to other levels or areas. In addition to formal gatekeepers, she had to negotiate with informal gatekeeper-workers as she worked with the hotel employees. At another hotel, the human resource manager was the primary gatekeeper and there was less negotiation with each area. The internal arrangements and authority within each hotel altered access issues.

ENTERING THE FIELD When entering a field site, adopt

a flexible plan of action. Each site is different, and how you enter will depend on your prior experience, contacts, commonsense judgment, social skills, and often just plain luck. Three issues to consider are presentation of self, amount of disclosure, and social role. PRESENTATION OF SELF When you begin any new social relationship, including entering a field site as a researcher, you display the type of person you are or would like to be. Consciously or not, you do this through your physical appearance, what you say, your use of tone and mannerisms, and by your actions. The message may be, “I’m a serious, hard-working student,” “I’m a warm and caring person,” “I’m a cool jock,” or “I’m a rebel and party animal.” You can show more than one self, and the self you present can differ by the occasion. Self-presentation processes in the field site are critical. Your demeanor—your manner of speaking, the way you walk or sit, your facial expressions and eye contact, your hairstyle and your clothing—speaks for you. Be aware of what they convey to members in the field.

218 Chapter 10 How should you dress in the field? Compare Your Thoughts The best guide is to respect both yourself and the people you are studying. A professor who studies street people does not have to dress or act like one; to dress and act informally is sufficient. Likewise, if you study corporate or school officials, formal dress and display a professional demeanor are expected. Dressing and behaving informally may block your entrée. Being very aware of how self-presentation affects field relations takes skill. You want to fit in, but it is difficult to present a very deceptive front or to present a self that deviates sharply from your ordinary self. Your discomfort and awkwardness will show through and can impede developing smooth relations in the field. DISCLOSURE A field researcher decides how much to dis-

close about self and the research to gatekeepers and members in the field. Revealing details of your personal life, such as hobbies, interests, and background, can build trust and help to create intimate relationships with people in the field. It also can result in a loss of privacy. We can think of disclosure on a continuum. At one end is covert research, in which no one in the field is aware that research is taking place. At the opposite end is fully open research, in which everyone becomes familiar with the researcher and is aware of research project details. The degree and timing of disclosure depend on your judgment and the particulars of a setting. Disclosure may unfold over time, as you feel more secure. SOCIAL ROLES You play many social roles in daily life— daughter/son, student, customer, sports fan, friend, and so forth. You can switch roles, play multiple roles, or play a role in a particular way. You choose some roles, while others are provided to you. Few of us have a choice but to play the role of son or daughter, although you have some control over how you play the role. Some social roles are formal (e.g., bank teller, police chief), others are informal (e.g., flirt, elder states person, or buddy). Field researchers assume and play various social roles in the field. At times, they adopt an existing role, such as the role of housekeeper that Rachel Sherman adopted in her study of hotels. Some existing roles provide greater access to all areas of the site. They let you observe and interact with all members and give you freedom to move around the site. Other roles are more restrictive. For example, the role of bartender when studying a tavern permits access to all areas, but it may limit freedom because it requires providing service, protecting the business, and collecting money. At other times, researchers create a new role or modify an existing one. Fine (1987) created a role of the “adult friend” and performed it with little adult authority when studying preadolescent boys. He was able to observe parts of the boys’ culture and behavior that were otherwise inaccessible to adults. It may take time to adopt some roles, and you may adopt many different field roles over time.

Your skills, time, and personal features—such as age, race, gender, and attractiveness—influence the roles open to you. You can only control some of these. Such factors can influence gaining access or can restrict available roles. Since many roles are sex-typed, your gender is an important consideration. Female researchers have encountered difficulties in a dangerous setting where males are in control. They may be shunned or pushed into limiting gender stereotypes. Gurney (1985) reported that being a female in a maledominated setting required extra negotiations and “hassles.” Nevertheless, her gender also provided insights and created situations that would have been absent with a male researcher. Age, race-ethnicity, and physical ability-disability or stature can have a similar impact on role selection.

WRITING PROMPT Gatekeepers Think of an occasion when you encountered an official or unofficial gatekeeper who controlled who could be in a setting. How did the negotiation for gaining entry or access occur? Where were you successful in gaining entry? Why or why not? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

10.4: What Should You Doin the Field Site? 10.4 Apply the five parts of ongoing observation and research in the field Eventually, after you gain access, you will find yourself in a field site with a preliminary social role. You have advanced to the longest phase of field research. During it, you must manage field relations as you systematically observe and record data.

Learning from History Taxis, Customers, and Tips

Observing People in Natural Settings 219 In a famous article, Fred Davis (1959) analyzed what he learned from field research when he worked as a taxi driver for six months in 1948 in Chicago. He focused on the social relationship between the driver and passenger or fare. He observed that drivers have limited control over the job. They have little choice of passengers and only a few regular clientscustomers. Mobile and without a fixed business location, the low-status cabdriver devoted long hours to short-term encounters with many diverse individuals. The encounters were highly variable and contained risk (such as fare jumpers, women in labor, robbers). Customers often ignored the driver or acted as if he or she was not there. Davis noticed that taxis drivers, like others in similar low-status service jobs, “sizedup” and quickly classified their customers. Customer categories were a large part of the taxi driver’s daily social world. Using a customer category system added a small degree of control and predictability in a highly uncertain job. It also helped when evaluating risk. Drivers also used it to gain some control and estimate the likelihood of tips. The driver had little discretion over how to perform a service. Beyond careful driving, opening doors, or speedy delivery, using a customer classification enables the driver to assert some control and engage in strategic emotion work—smile, make small talk, display kindness—to improve tips.

WRITING PROMPT Where to Start While many researchers study a new social activity or setting, others conduct field research on a job site, leisure interest, or social activity with which they are an active participant. Think of a personal interest or activity of yours that might make a good field research project. What about it makes it good for doing a field research study? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

10.4.1: Being in the Field Once you have selected a site, gained access, and assumed a social role, you need to settle in. We can divide ongoing observation and research in the field site into five parts. LEARNING THE ROPES New researchers often face

embarrassment, experience discomfort, and become overwhelmed by the details in the field. Maintaining a “marginal” status is stressful. It is difficult to be an outsider who is not fully involved, especially when studying settings full of intense feelings (political campaigns, religious conversions). Do not be surprised if you feel awkward until you “learn the ropes” and acquire an understanding of microlevel norms, rules, and customs of a field site. As you learn the ropes and fit in, you learn how to cope with stress and how to normalize the social research.

NORMALIZING SOCIAL RESEARCH A field researcher not only observes and investigates members in the field, but he or she is also being observed and investigated. Frequently, field site members are initially uncomfortable with the presence of a researcher. Many people may be unfamiliar with field research and confuse sociologists, psychologists, counselors, and social workers. They may see you as an outside critic or spy, or as a savior or all-knowing expert. You need to normalize. To help members adjust to the research process, present your own biography, explain field research a little at a time, appear nonthreatening, and accept or ignore minor rule violations in the setting. BUILDING RAPPORT AND TRUST As you overcome

your initial bewilderment to an unusual way of talking or system of meaning, you can slowly build relations of trust and establish rapport. This takes time and repeated positive social interactions. To do this, you need to “get along with” members in the field. This may require listening sympathetically to complaints, sharing experiences, swapping stories, and laughing or crying with field site members. Many factors influence building trust and rapport— how you present yourself; your role in the field; and the events that facilitate, limit, or destroy trust. Gaining rapport and trust is a developmental process that builds up through many social nuances (e.g., sharing of personal experiences, storytelling, gestures, hints, facial expressions). Trust and rapport are easier to lose once established than to gain in the first place. Maintaining good rapport requires repeated social engagements with almost daily reinforcement. For example, if you establish rapport with a group of people and then disappear for three months, do not expect the same level of rapport instantly when you return. Creating trust often requires taking risks or passing mini social “tests.” It also requires continuous reaffirmation, and you re-create it anew as you enter new topics, issues, physical locations, or social groups. In his field study of restaurant kitchens, Gary Fine (1996, pp. 236–237) noted “One way in which my ambiguous presence was handled was through a continuing stream of joking, transforming me from a researcher to worker, from observer to participant . . . Most attempts to involve me were jocular, indicating both my closeness and distance . . .” Besides observing, he found himself peeling potatoes, stirring soup, or getting food from shelves. It is not easy to build rapport and trust in some field sites. Certain sites are full of fear, tension, and conflict, such as a prison or alley where illegal drugs are sold and used. Field site members may be unpleasant, untrustworthy, or untruthful. They may do things that disturb or disgust you. An experienced researcher is prepared for a range of events and relationships. However, it is sometimes impossible to penetrate a setting or get really close to members.

220 Chapter 10

Figure 10.2 Research Role by Degree of Involvement in the Field Site Research Roles in a Field Site

Total Observer

Settings in which cooperation, sympathy, and collaboration are impossible may require unusual techniques. Building rapport and trust is a step toward obtaining an understanding of the social life of a field site. Rapport helps you to understand the members. The understanding is a precondition for gaining greater access to a deeper level—a member’s inner worldview and perspective. The step after seeing an event from a member’s point of view is getting inside and grasping how the member thinks, feels, and reacts. You move beyond surface understanding toward deeper empathy. Do not confuse empathy with sympathy, agreement, or approval. It is simply the ability to see, feel, and think as another person does. Rapport helps to build understanding and ultimately empathy. At the same time, empathy facilitates greater rapport. NEGOTIATING CONTINUOUSLY Most field sites have multiple levels or areas. Entry can be an issue for each. Field site entry is more analogous to peeling the layers of an onion than to opening a door. Moreover, bargains or promises made at initial entry may not be permanent, requiring fallback plans and renegotiation. Perhaps you got permission from the principal and parents to observe young children. After you arrive at the school, you find that two teachers control the children’s time closely. They block your access and give you no chance to observe the children playing or interacting spontaneously. Your fallback plan may be to shift to study and spend more time with the two teachers. You could try to see the world from their viewpoint. Once you better understand them and win their trust, you might be able to renegotiate with them to observe the children more. Frequently, the focus of field research only emerges mid-way in the research process, and it can change. This means you must negotiate for access to new areas as your research focus develops. Also, as you encounter new people in the field site, you may have to negotiate terms again with each new person until a stable relationship develops or has to be renewed. Expect to negotiate and explain what you are doing repeatedly. Although you may feel it, try to suppress displays of your irritation or impatience with the repetition. DECIDING ON A DEGREE OF INVOLVEMENT In addition to a social role in the field, you adopt a researcher role.

Complete Participant

This kind of role is on a continuum by the degree of direct involvement with members in the field and their activities. At one extreme is a detached and remote outsider. You only observe events from the distance and do not directly interact with people. At the opposite extreme is an intimately involved and engaged participant. You fully engage and begin to become just like another member in the field site (see Figure 10.2). The outsider role is faster, easier, and necessary at the start of field observation. At the outsider end of the continuum, you need little time for acceptance and overinvolvement is unlikely. Often, people find it easier open up to the detached, outsider visitor. This role also protects your self-identity. However, you may feel marginal and lack access to an insider ’s experience, thereby increasing the chances that you will misinterpret events. To understand local social meaning fully, it is usually necessary to participate activities and engage others in a field site. Roles at the insider end of the continuum facilitate empathy, sharing experiences with field site members, and fully experiencing their intimate social world. However, as social distance shrinks and sympathy for field site members grows, overinvolvement is possible. Intimate contact can make serious data gathering difficult; it might compromise the distance needed for research analysis. The field researcher’s feelings of loneliness and isolation in the field may combine with a desire to gain rapport and create the situation of going native. This can destroy a study because becoming “one of the gang” or the same as field site members can take priority over careful observation or analysis. A good field researcher judges the best level of engagement for a specific setting and maintains a balance between being involved without being overly involved.

10.4.2: Strategies for Success in the Field All field researchers have strategies that they tailor to the specifics of a field site and their own background. In this section, we look at six of the strategies that many field researchers have used.

Observing People in Natural Settings 221

BUILDING RELATIONSHIPS Over time, you develop

social relationships with people in the field site. This requires putting in time, remembering personal details, and making “small talk” such as saying “good morning” or commenting on the weather. Members who are cool at first may warm up later. Alternatively, they may put on a front of initial friendliness, with fears and suspicions surfacing later. When you know little about a field site, you may resist forming a close relationship with particular people because you are unfamiliar with the social scene. If you develop one or two close friends quickly, they can become allies and valuable sources of information, explaining events to you, introducing you to others, defending your presence, and helping with access. However, first friends can also exercise unwanted control over your field observation, block establishing relations with others, or limit your access to certain areas in a field site. A good field researcher quickly notices social cliques, friendship ties, disputes or tensions, and power dynamics in a setting. For example, after two days in a field site you recognize that Samantha and Judy seem to avoid one another and do not get along. George and Roberto appear to be close friends. Minsook looks quiet, shy, and uncertain of herself and has few friends. April appears to be outgoing and confident and seems to be in charge. As you acquire awareness of the social “lay of the land,” you also need to monitor how your own actions or appearance affects others. For example, a physically attractive researcher who interacts with members of the opposite sex often encounters crushes, flirting, and jealousy. You need to be aware of these field relations and manage them. You must maintain some level professional detachment and discipline but not hurt feelings, harm rapport, or close off access. At times, you need to be able to break off relations or withdraw from them. You may discover that to forge social ties with some people, you must break off close social ties with others. As with the end of any friendly relationship, social withdrawal can cause emotional discomfort or pain for both you and the other person. As a field researcher, you must constantly balance social sensitivity with your research goals. PERFORMING SMALL FAVORS You have probably

noticed that social life is full of exchange relationships. You do something for someone else and he or she usually returns the favor. Exchange relationships also develop in a field site. People exchange small favors, including deference and respect. A research strategy for gaining acceptance in the field is to help in small ways but not expect anything in return. As you repeatedly help and perform “small acts of kindness,” people in the field incur an informal social obligation. They may feel they should reciprocate by offering you help in the field. APPEARING INTERESTED AND EXERCISING SELECTIVE INATTENTION At times, you may feel bored or

distracted in the field site. Looking bored is a nearly certain

way to break off or weaken a developing social relationship. You should learn to “act” and maintain an appearance of interest. Appearing interested in the people and events of a field site with your words and actions (e.g., facial expression, going for coffee, attending a party, listening to jokes), even if you are not truly interested, is an important field research strategy. Putting on a temporary front of involvement to be sociable is a common small deception of daily life. It is a part of being polite. Of course, selective inattention is also part of acting polite. If someone makes a social mistake, the polite thing to do is to ignore it. Although outward inattention should never mean a failure to note what is occurring in the field. BEING AN EARNEST NOVICE When you are new to a setting and first forming relationships, you may be curious and ask many questions. One field strategy is to extend this mode of operating for some time. Behaving like an “expert” is not the way to learn about a field site or win friends. As the outsider, your mission in the field is to observe, listen, and learn from others. It is not to brag, talk a lot, promote your opinions, or correct others’ mistakes. If anything, you should act slightly less informed and knowledgeable than you really are. This is a good way to learn from people in the field. Being humble and adopting a learner role pays research dividends. You learn more by careful listening and asking questions. Treating field site members as experts encourages them to share confidences with you. There is a balance: You want to be eagerly inquisitive but not appear unnecessarily stupid or naïve. It is best to try to ask for explanations rather than assume that you know what is happening or why people do things in certain ways. By adopting a novice role as someone who wants to learn, you both learn more and show respect for field site members. AVOIDING CONFLICTS Fights, conflict, and disagree-

ments can erupt in the field. You may study people with opposing positions. In such situations, you may feel intense pressure to take sides. People may test you to see if you can be trusted or are on the enemy’s side. In such occasions, it is usually best to stay on the neutral sidelines and try to walk a tightrope between opposing sides. Once you align closely with one side, you will be cut off from access to the other side. In addition, you will only see the situation from one point of view and may get a distorted picture of events. ADOPTING AN ATTITUDE OF STRANGENESS Life

contains thousands of details. If you paid attention to everything all the time, you would suffer from severe information overload. We tend to manage life by ignoring much of what is around us and by engaging in habitual thinking. We overlook what is familiar and assume that most other people experience reality the same as we do, treating our way of living as natural or normal. We rarely recognize what we

222 Chapter 10 take for granted. For example, someone hands you a wrapped gift. You probably say “thank you,” open the gift, and then praise it. Yet gift-giving customs vary by culture. In some cultures, people barely acknowledge the gift and quickly put it aside unwrapped. In others, people expect the gift receiver to complain that the gift is inadequate. Habitual thinking and not noticing what is familiar makes doing field research in a familiar setting difficult, instead field researchers adopt an attitude of strangeness. When we first go to a new, unfamiliar place as a stranger, we are extra sensitive to the physical and social surroundings. This occurs because we do not know what will happen and most things are new to us. When we visit a very different culture, we may encounter different assumptions about what is important and how people accomplish daily tasks. This confrontation of cultures, or culture shock, makes it easier for use to see cultural elements, and it facilitates self-discovery. Adopting an attitude of strangeness enables us to retain such an extra-sensitive perspective and use it to look at field events in new ways. The perspective of the stranger makes the unspoken, tacit culture of a setting more visible, and makes it easier for us to notice assumptions and details that we frequently overlook. In addition, the stranger perspective can make reflection and introspection easier and more intense. A good field researcher adopts both a stranger’s and an insider’s point of view, learning to see the ordinary both from an outsider-stranger’s perspective and as an insider-member does. The ability to switch back and forth between perspectives quickly is a critical field research skill to acquire.

10.5: How Do You Collect Data in the Field? 10.5 Analyze ways to generate good qualitative fielddata Data collection begins before you enter the field site, as you plan, identify a field site and gain access. Once in a field site, as you begin to conduct observations and interact with participants, you must become compulsive in seeking out and preserving field data for subsequent analysis. You want to develop the essential skill of acquiring large amounts of high quality data while you are in the field site.

10.5.1: Observing and Taking FieldNotes In this section, we examine how to get good qualitative field data. Field data include what a researcher experiences, remembers, records in field notes, and makes available for systematic analysis.

The good field researcher is a resourceful, talented individual with ingenuity and an ability to think quickly on his or her feet while in the field. In quantitative research, you may use various instruments like questionnaires or response reaction measures in a computer, to acquire data. In field research, you are the instrument for acquiring field data. This has two implications. First, you must be alert and sensitive to what happens in the field and disciplined about recording data. Second, your social relationships as well as your personal feelings and subjective experiences are field data. They are valuable in themselves and for interpreting events in the field. Instead of trying to be objective and eliminate personal reactions as in quantitative studies, reflect on your reactions and feelings and then record them as data. For example, a field researcher visits a p*rnography shop to study what happens inside. He notices an increase in personal unease and anxiety. He reflects on sources of the unease—is it a fear of discovery, is it excitement from transgressing a moral line, or is it uncertainty over someone approaching him in the store to solicit for sex? The researcher’s inner anxiety and feelings that have accompanied his observations of customers and clerks at the shop are legitimate field data. He should notice and record them, as Karp (1973) did in his study of p*rn shops. WHAT SHOULD YOU OBSERVE? In the field site, you must pay very close attention, watch all that is happening, and listen carefully. You use all your senses. Notice what you see, hear, smell, taste, and touch to absorb all sources of information. In the beginning, you want to scrutinize the physical setting and capture its atmosphere. If indoors, notice the color of the floor, walls, and ceiling. How large is the room? Where are the windows and doors? How is the furniture arranged, and what is its condition (e.g., new or old and worn, dirty or clean)? What type of lighting is there? Are there signs, paintings, or plants? What are the sounds or smells? If outdoors, notice buildings, signage, landscaping, street containers and utility poles, empty lots and alleys, night lighting, parking and traffic, odors, graffiti, litter, and the soundscape. Are the sidewalks cracked or new looking with freshly painted indicators? Is it a lively crowded street filled with people walking or sitting on benches situated next to planters, or is it silent and empty except for those lurking in shaded doorways or peeking from windows?

Example Study Noticing Details In one of the luxury hotels she studied, Sherman (2006) noticed that management sent contradictory messages to workers: (1) You are part of a family or community and have free choice; (2) you are under surveillance and required to p erform your job.

Observing People in Natural Settings 223 How did she arrive at this conclusion? She noticed many details in the employee handbook, at training sessions, in worker–manager interactions, and many small things in general. For example, she saw a sign posted for an upcoming employee meeting. It said, “come enjoy the refreshments,” and under this message it said “attendance is mandatory” (2006, p. 75). Through her careful attention to details that she recorded and on which she later reflected, she could build a picture of the corporate culture sustained in each luxury hotel.

WHY BOTHER WITH SUCH DETAILS? Stores and restaurants often plan lighting, colors, and piped-in music to create a certain atmosphere. Used-car salespeople spray a new-car scent into cars. Shopping malls and stores intentionally send out the odor of freshly baked cookies to attract customers. To sell a house, people keep a lawn trim, add a fresh coat of paint, or leave pleasant aromas linger. Urban designers have documented how a large sterile, hard-surfaced open space heightens tension, encouraging quick passage without making eye contact, while other urban physical arrangements encourage relaxed social interaction and informal gatherings.

Compare Your Thoughts Many subtle, unconscious signals influence human behavior. As a field researcher, you want to notice, capture, and record anything in the surroundings that could influence social relations and contributes to creating an “atmosphere.” A good field researcher is intrigued with details. Field observing can be trying, tedious work but a genuine curiosity about details provides the motivation to continue. The mass of ordinary details can reveal “what’s going on here.” By noticing and assembling hundreds of ongoing mundane, trivial, and daily minutiae, a researcher can uncover important features of a field site. What many people often overlook in a setting is what a field researcher learns to notice and from which he or she can discover a great deal. In addition to physical surroundings, you observe people in the field. Notice each person’s observable physical characteristics: age, sex, race, and stature. Why? People socially interact differently depending on whether someone is 18, 40, or 80 years old; male or female; white or nonwhite; short, thin, and frail or tall, heavyset, and muscular. For example, a heightened sensitivity to the racial composition of a setting can be beneficial. A white researcher in a multiracial society who fails to notice that everyone in a setting is also white is not only being racially insensitive but is missing a potentially critical feature for understanding the dynamics in the setting. For example, a 20-year-old field researcher observes a crowded restaurant and sees nothing out of the ordinary until an elderly couple enters then suddenly becomes aware that no one in the restaurant

is under 30 years old. The researcher had almost overlooked a potentially significant feature of the setting, the narrow age range of all the staff and customers. In her study of luxury hotels, Sherman (2006) noticed a “stark” racial/ethnic division of labor in one hotel—“front office workers, many of whom were European, were white; the only nonwhite worker at the desk was Inga, a young Swede of Asian descent. Three of the four doormen were white. . . . A striking and important feature of the front office workers . . . was their youth. Annie, Jackie, Betsy and, Ginger, white front desk/concierge workers, were all under twenty-three” (pp. 86–97). You record all the details because something of significance might be revealed. You may not become aware of a detail’s relevance until later. You want to err by including everything rather than missing potentially significant details. Highly specific, descriptive details capture the setting and events. For example, “a tall, white, muscular 19-year-old male with a crewcut in designer jeans sprinted into the brightly lit room just as the short, heavyset greyhaired black woman in her late sixties wearing a faded tailored blue dress and sequined hat eased into a battered plastic chair” says more than “one person entered, another sat down.” PHYSICAL APPEARANCE AND SOCIAL INTERACTIONS

Physical appearance—such as makeup, neatness, clothing, or hairstyle—sends messages that can influence social interactions. Some people devote a lot of time and money to selecting clothes, styling and combing their hair, putting on makeup, shaving, ironing clothes, and using deodorant or perfumes. This is part of their self-presentation. People who do not groom, shave, or wear deodorant are also presenting themselves and sending messages. No one dresses or looks “normal.” Such a statement suggests that you are not seeing the social world through the eyes of a stranger or are insensitive to social signals. Beyond appearance, people’s actions can reveal much about them. Notice where people sit or stand, the pace at which they walk, and their nonverbal communication. People express social information, feelings, and attitudes through nonverbal communication, including gestures, facial expressions, and how they stand or sit (standing stiffly, sitting in a slouched position, etc.). People express relationships by where they position themselves in a group and through eye contact. You want to read social communication by quickly noticing who is standing close together, having a relaxed posture and making eye contact, and who is not. You also need to notice the context in which events occur: Who was present? Who just arrived or left the scene? Was the room hot and stuffy? Was it late or early in the day? Such details help you assign meaning and learn why events occur. If you fail to notice such details, they are lost. Also lost is a full understanding of the event.

224 Chapter 10 Lastly, you need to notice exactly what people say, their specific words and phrases. Also note how it is said— the loudness, tone of voice, accent, and so forth. Intense listening is tiring and difficult in a noisy room with many conservations or distracting sounds. If you are not part of a private conversation, listening to what people say is eavesdropping and considered very impolite. Yet a stranger who overhears a loud conversation is not being rude. Eavesdropping is intentionally listening to something meant to be private, but overhearing is unintentionally hearing when speakers show little concern for privacy. Field researchers engage in both with care and discretion. As you hear phrases, accents, and grammar, listen to what is said, how it is said, and what it implies. For example, people often use ambiguous phrases such as “you know” or “of course” or “et cetera.” A field researcher tries to discover meaning behind such phrases. If someone fails to complete a sentence but ends with “you know,” what does she mean? You hear a 14-year-old boy say, “We all went to the mall and hung out until 2, you know.” The phase “you know” may be a sloppy speaking habit. It could mean he expects you to be aware of what 14-year-old boys do with their friends hanging out at a shopping mall for two hours in the middle of the day. Or the phrase could be deflecting attention since he engaged in forbidden behaviors and does not want to tell you. USING SERENDIPITY Although field researchers plan

field observations and interviews, they are also prepared to take advantage of unplanned, unexpected events that can reveal much about a field site. An alert field researcher is always prepared to use casual eavesdropping or observing chance events not meant to be public as a means to learn about a field site. The importance of serendipity and ease-dropping is illustrated by Nolan (2011, pp. 54–55) who described a situation in a high school principal’s office while she was waiting to get an ID card to start conducting field research on policing inside an urban high school. While quietly sitting, several police officers brought in two boys in handcuffs. She said, “For me, the experience was new. I had not yet grown accustomed to the scene, and it seemed jarringly out of place in a school.” As she waited and watched, an aide in the office told her that her ID would not be ready until the next day, and she responded, “I’m watching this.” When told she did not have permission, she responded that the principal had given her permission. She continued to observe and noted, “One of the officers glanced at me, I felt very conspicuous. More angry words were exchanged between the boys and the officers as they waited for the police van. I pretended not to notice. Finally, my awkwardness got the better of me and I left the room.” This introduction to events occurring in her field site played a role in how her study developed.

Inexperienced field researchers complain that they observed in a field site but “nothing happened.” They express frustration with the amount of “wasted” time waiting for something to occur. Such responses indicate that they have not yet learned the importance of serendipity in field research. You do not know the relevance of what you observe until later. Keen observations are useful at all times, even when it appears that “nothing happens.” Although “nothing happened” from your perspective, what about from that of members in the field? As a field researcher, you need to operate on other people’s schedules and observe events as they occur within other people’s flow of time. You may be impatient to get in, get the research over, and get on with your “real life.” For people in the field site, this is their real life. You may need to subordinate your personal wants to the flow and demands of the field site. Field researchers are sensitive to the pace of events and flow of time in a field site. They appreciate wait time, or time in the field spent “hanging out” or a pause between significant activities. It can indicate a slow pace, a delay because of scheduling problems, a stalemate in a conflict, or power dynamics in which unimportant people display their respect for the powerful. Wait time is not always “wasted time” and can reveal the pace and rhythm of a setting. You can use wait time to reflect, observe details, develop social relationships, build rapport, and simply as a way to become a recognized, accepted part of the ongoing field setting. Wait time can communicate to others that you are committed and serious. Good field researchers know the value of perseverance, both for observing and collecting field data and for earning respect from field site members. Like the appearance of interest, it sends a signal that you are earnest and committed, and that you believe what occurs in the field site is valuable and important. SAMPLING Field research sampling differs from that of

survey research. Field researchers do not use random sampling. They sample by taking selective observations from all possible times, locations, people, situations, types of events, or contexts of interest. You might sample time by observing a setting at different times of the day. For example, in studying a bowling alley you observe at three times of the day, on weekdays and weekends, to get a sense of what remains the same and what changes. It is often best to overlap sampling times (e.g., 9:00 to 11:00 a.m., 10:30 a.m. to noon, 11:30 a.m. to 1:00 p.m.). You may want to sample locations because sitting or standing in different locations provides a better sense of the whole site. Let us say you are studying the emotion work of a waitress. You want to observe interactions in the front area (with customers), in a back area (with cooks, etc.), and in a back break room with coworkers. You would sample all shifts and meal times, both the slow and very busy times. Observations from

Observing People in Natural Settings 225

multiple locations and times will give you a much richer picture of the entire social setting. You can sample people by focusing attention on different kinds of people (old-timers and newcomers, old and young, males and females, leaders and followers). As you identify all the types of people, or people with various outlooks in a field site, you want to interact with and learn about all of them. You can also sample three types of field site events: routine, special, and unanticipated. • Routine. Events that occur over and over again the same way (e.g., opening up a store for business every day). Do not be mistaken and think they are unimportant simply because they are routine. • Special. Events announced and planned in advance (e.g., annual office party). These events focus member attention and can reveal aspects of social life not otherwise visible. • Unanticipated. Events that just happen while you are present (e.g., how unsupervised workers act when the manager gets sick and cannot oversee them for a day). In this case, you may see something unusual, unplanned, or rare. Such events might reveal new aspects of a setting, such as how much the workers really respect the manager and follow her rules even when she is not around to observe them. BECOMING A SKILLFUL NOTE-TAKER Field research

data include your memories of observations and experiences and your recorded field notes. Field notes are the permanent record of observations and experiences. Producing good field notes is essential for a high-quality ethnography or field research study. Do not plan to take notes while in the field site. After spending some time in the field site, plan to sit down in a quiet place and write from memory. Especially in the beginning, you should leave the site to write notes after only one or two hours of observation. Expect to spend nearly as much time writing as you did observing in the field site. If you observed for two hours, you may be writing for two hours. Any notes you take directly in the field site differ from full field notes. A common mistake of new field researchers is to try to take the full field notes while still in the field site. Only take jotted notes in the field site, not full field notes. If you take jotted notes, write them on a small scrap of paper inconspicuously or in private (such as in a bathroom). You use them to get down one or two critical key words or phrases that can stimulate your memory later. Immediately after you leave the field site, sit down in a quiet place to write the full notes. Writing field notes requires self-discipline and can be boring, tedious work. You need to make writing notes a compulsion. Your notes should contain extensive descriptive detail drawn from

short-term memory. New field researchers find they can recall more details with some effort. Generally, the quality and quantity of notes improve over time. Be sure to keep field notes neat and organized. You will return to them over and again. Always put the date and time of the observation at the top of the first page for a session in the field. Once written, the notes are private and valuable; treat them with care and protect their confidentiality. Sometimes hostile parties, blackmailers, or legal officials will want to read them. Some professional field researchers write their field notes in code. Your mood, state of mind, attention level, and conditions in the field can affect note-taking. Beyond detailed descriptions, your full field notes can contain maps, diagrams, photographs, interviews, audiotape recordings, videotapes, memos, objects from the field, and jotted notes taken in the field. For a field project of a few weeks, you might fill several notebooks or the equivalent in computer files. Some professional researchers have produced 40 single-spaced pages of notes from three hours of observation. With practice, even a new field researcher can soon produce four or five pages of notes for each hour in the field. MAKING IT PRACTICAL: RECOMMENDATIONS FOR TAKING FIELD NOTES

1. Record notes immediately after being in the field. Avoid talking with others until you record observations. 2. Begin the record of each field visit with a new page. Enter the date and starting and ending times. 3. Use the optional jotted notes as a temporary memory aid, with one or two key words or terms. 4. Use wide margins so that you can add to your notes at any time. Go back and add things you remember later. 5. Plan to type notes and keep the note levels (to be discussed later) separate so that you can easily return to them. 6. Record events in the order in which they occurred. Note their length (e.g., a 15-minute wait, a 1-hour ride). 7. Be as specific, concrete, complete, and comprehensible as possible. You can always throw away extra details. 8. Try to recall exact phrases. Use double quotes for exact phases and single quotes for paraphrasing. 9. Record small talk or routines that do not appear to be significant at the time; they may become important later. 10. “Let your feelings flow,” and write quickly without worrying about spelling or “wild ideas.” 11. Never substitute tape or video recordings completely for written field notes. 12. Include diagrams or maps of the setting. Outline your own movements along with those of other people during observation.

226 Chapter 10 13. Include your own words and behavior in the notes. Also, record your emotional feelings and private thoughts.

MAPS AND DIAGRAMS IN FIELD NOTES Many field

14. Avoid evaluative generalizing or summarizing words. For example, rather than writing “The sink looked disgusting,” it would be better to write, “The sink was ruststained and appeared as if it had gone without cleaning for weeks. Scattered bits of food lay about, and it had a pile dirty pans and dishes. A film of grey water lay across the bottom, emanating an odor of rotting garbage.”

researchers make maps and draw diagrams or pictures. Maps and diagrams help the researcher organize ideas and events in the field, and it helps when conveying life in a field site to other people. For example, you observe a café’s lunch counter with eight stools. You may draw and number eight circles to simplify recording (e.g., “Yosuke came in and sat on stool 8; Phoebe was already on stool 6”). Field researchers create three types of maps: spatial, social, and temporal.

15. Reread your notes periodically and record any ideas generated by the rereading.

• A spatial map orients the data in space or physical location.

16. Always make backup copies of your notes and store the copies in different places in case of fire or other disaster.

• A social map shows connections among people and follows the flow of interactions indicating power, influence, friendship, division of labor, and so on.

WRITING PROMPT

• A temporal map shows time starts, endings, and durations for people, goods, services, and communications.

Field Notes Describe at least three ways that you can practice and develop the self-discipline required to write 20–30 pages (single spaced) of field notes from a few hours observation? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

How do you take field notes? Compare Your Thoughts There are several ways to take field notes. The recommendations here are only suggestions. With experience, you may develop your own system. Full field notes have several levels. Keep all the notes for an observation period together, but

Figure 10.3 Example of Spatial Map in Field Notes This Map of field site, a small café with 21 seats (8 at the counter and the remainder at 5 tables) is an example of what a map may look like in field notes. Window

Window

Entrance

Restrooms

Observing People in Natural Settings 227

distinguish the levels of notes by starting each level on a new page or use a new color or font for each level. You will not produce the same amount of notes for each level. If you observe for six hours, you might have a tiny scrap paper of jotted notes, 40 pages of direct observation, 5 pages of researcher inference, and 2 pages in total for the methodological, theoretical, and personal level notes. Five Levels of Field Notes • Jotted notes. Short, temporary memory triggers such as words, phrases, or drawings taken inconspicuously, often scribbled on any convenient item (e.g., napkin, matchbook). You will incorporate them into direct observation notes. They are never substitutes for full notes. • Direct observation notes. These are the core of your field data and are written immediately after leaving the field site. Order them chronologically, with the date, time, and place on each entry. They are a detailed description of everything you heard and saw in very concrete, specific terms. To the extent possible, they are an exact recording of the particular words, phrases, or actions, never summaries or generalizations. • Researcher inference notes. You need to look and listen without inferring or imposing an interpretation. Observations without inferences go into direct observation notes. You record inferences in a separate section that is keyed to direct observations. Keep inferred meanings separate from direct observation because the meaning of actions is not always self-evident. For example, a couple registers at a motel as Mr. and Mrs. Smith. You record what actually happened in the direct observation level but put your inference that they are not married in the inference level. A great deal of social behavior is ambiguous and has multiple possible meanings. Your own feelings, interpretations, and reactions are part of data in the field and should be in the notes, but separated. • Analytic notes. Keep methodological ideas in analytic notes to record plans, tactics, ethical and procedural decisions, and selfcritiques. You may have educated hunches or emerging theoretical ideas during data collection, and you should put them in the notes. Analytic notes are a running account of your attempts to give meaning to field events. You can “think out loud” in your analytic notes by suggesting links between ideas, proposing conjectures, and developing new concepts. Your analytical notes can include ethical concerns that you encounter and strategies you invent. • Personal notes. Personal notes serve three functions: They are an outlet for you and a way to cope with stress, they are a source of data about personal reactions, and they are a way to evaluate direct observation or inference notes later when you reread notes. For example, if you were in a good mood during observations, your mood might color what you observed or how you felt about events.

MAKING IT PRACTICAL: THE LIMITATIONS OF RECORDING DEVICES IN THE FIELD The novice field

researcher could mistakenly think audio or video recorders make field note-taking unnecessary. Professional researchers sometimes use them as supplements in field research but rarely as complete substitutes. Recorders

rovide a close approximation to what occurred and can be p a permanent record to help you to recall events. However, recorders can never fully substitute for written field notes or your presence in the field and have many limitations. Carrying a small digital recorder on you (or using a cell phone to record) can prove useful, so long as it does not arouse suspicions in the field site or violate privacy protections. Limitations include the following: • You cannot introduce them into all field sites for practical reasons (large area with noise) or legal reasons. • People in the field will see them as a threat; recording devices frequently create a disruption and raise awareness of surveillance. You can only use them after you have developed rapport and trust. • Recorders frequently miss action or are out of range, they break down or fail, or they may require your attention and reduce direct personal involvement with what is happening in the field site. • Recorders rarely save time. You can expect to spend two to three times longer reviewing and transcribing recorded material than the time of the recoding. You may have three hours of recording, but it could take you an additional eight hours to review and transcribe what you recorded.

WRITING PROMPT People Watching How is field research observation different from casual people watching? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

10.5.2: Interviewing in Field Research Field researchers use unstructured, nondirective, in-depth interviews. These differ from formal survey research interviews in many ways. In a field research interview, you ask questions, listen, express interest, and record. The field interview is a joint production between you and a field site member or informant. People you interview are active participants in the discussion process, and their insights, feelings, and cooperation reveal their perspectives and subjective meanings. You interview in the field or a convenient location and are informal and nondirective. It is acceptable for you to share your background to build trust and encourage the informant to open up. Do not force answers or use leading questions. You want to encourage and guide in a process of mutual discovery.

228 Chapter 10 In field interviews, people express themselves in their habitual way of speaking, thinking, and organizing reality. You want to retain their expressions, jokes, and narrative stories in a natural form without repackaging them into a standardized format. You want to stay close to the field member’s experience. This means you ask questions in terms of concrete examples or situations—for example: It is better to ask “Could you tell me what happened that led to your quitting your job in June?” instead of “Why did you quit your job?”

Field interviews often occur in a series over time. You first build rapport. Avoid probing inner feelings until you established some comfort or intimacy. After several meetings, you may be able to probe into sensitive issues and seek clarification of less sensitive issues. In later interviews, you can return to topics and check past answers by restating them in a nonjudgmental tone and asking for verification—for example:

6. You maintain a professional tone and business-like focus; you suppress or ignore diversions. 7. Closed-ended questions are common, and probes are rare. 8. You as the interviewer alone control the pace and direction of the interview. 9. You ignore the social context in which the interview occurs. 10. You mold the communication pattern into a standard framework. Typical Field Interview 1. The beginning and end are not clear. The interview can stop and be resumed later. 2. You tailor the questions you ask and the order in which you ask them to specific people and situations. 3. You show interest in specific answers/responses and encourage elaboration.

“The last time we talked, you said that you started taking a few things from the store after they reduced your pay. Is that right?”

4. The typical field interview is somewhat like a friendly conversational exchange, but with more interviewer questions.

The field research interview is a “speech event” closer to a friendly conversation than to the stimulus/response model of a survey research interview. It differs from a friendly conversation in that it has an explicit purpose—to learn about the informant and setting. You may include explanations or requests that diverge from friendly conversations. You may say,

5. Field interviews can occur in a group setting or with others in the area.

“I’d like to ask you about,” or “Could you look at this and see if I’ve written it down right?”

The interview is less balanced than an ordinary conversation. You ask more of the questions and may express more ignorance and interest than an ordinary conversation. You do not have to ask every person you interviewed the same questions, but tailor questions to specific individuals and situations. Repetition is common. You may ask a field site member to elaborate on unclear abbreviations or slang. For example, in her study of Chicago’s middle-class black neighborhoods, Pattillo (1999) noted of a teen boy who got “plugged,” which she defines as Chicago slang for joining a gang. MAKING IT PRACTICAL: SURVEY INTERVIEWS VERSUS FIELD RESEARCH INTERVIEWS Typical Survey Interview

1. It has a clear beginning and end. 2. You ask the same standard questions of all respondents in the same sequence. 3. You appear neutral at all times. 4. You as the interviewer ask all questions, and the respondent only answers. 5. The interview is with one respondent alone.

6. The interview can be interspersed with jokes, asides, stories, diversions, and anecdotes, and you record them. 7. Open-ended questions are common, and probes are frequent. 8. You and the field site member jointly control the pace and direction of the interview. 9. You note the social context of the interview and treat it as important for interpreting the meaning of responses. 10. You adjust to the member’s norms and language usage. TYPES OF QUESTIONS IN FIELD INTERVIEWS Field researchers ask three types of field interview questions:

• descriptive • structural • contrast questions You can ask all concurrently, but each type is more frequent at a different stage in the research process. When you first enter the field site, you will primarily ask descriptive questions. You can gradually add structural questions until, in the middle stage after analysis has begun, most of the questions are structural. Contrast questions begin in the middle of a study and increase until, by the end, you ask more of them than any other type. You ask descriptive questions to learn about the setting and members. Descriptive questions can be about time and space—for example, “Where is the bathroom?” “When does the delivery truck arrive?” “What happened Monday night?” They can be about people and activities: “Who is

Observing People in Natural Settings 229

sitting by the window?” “What is your uncle like?” They can be about events or activities: “What happens during the initiation ceremony?” They can be about objects: “When do you use a saber saw?” Questions asking for examples or experiences are descriptive questions: “Could you give me an example of a great date?” ”What happens in a perfect game?” You can ask about hypothetical situations: To the new teacher—“If a student opened her book during the exam, how would you deal with it?” You ask a structural question after spending time in the field while beginning a preliminary analysis of your data, as you are organizing specific field events, situations, and conversations into preliminary conceptual categories. You ask structural questions for both clarification and confirmation. For example, you conduct field observations at a highway truck-stop restaurant. Over time, you learned that the employees informally classify customers patronizing the truck stop. Based on a preliminary analysis, you think employees classify the customers into five types. Using structural questions, you seek verification of the five types and their features. You might ask whether a category includes features beyond those you already identified—for example, “Are there any types of customers other than regulars, greasers, pit stoppers, road rangers, and long haulers?” You ask for confirmation: “Is a greaser a type of customer you serve early in the morning?” “Would you consider Johnny Jensen to be a long hauler?” After you have verified major categories, processes, or aspects of life in the field with structural questions, you can begin asking contrast questions. Contrast questions focus on the similarities or differences among the categories, processes, or aspects. You ask questions to verify similarities and differences that you believe to exist among categories. You may ask, “You seem to have a number of different kinds of customers come in here. Two types just stop to use the restroom without buying anything—entire families and a lone male driver. Do you call both of them pit stoppers?” INFORMANTS The ideal informant is a person currently

in the field site, completely familiar with its culture, and in a position to have witnessed and participated in its significant events. The informant should not be too busy to spend some time talking with you. The ideal informant is nonanalytic. He or she uses ordinary native folk theory or pragmatic commonsense thinking. People who are highly educated and try to analyze the field site by imposing ideas or categories borrowed from the media or academic work are not good informants. In a long field study, you want to interview various types of informants, such as rookies and oldtimers, people who recently changed status (e.g., through promotion), and those who are static. You seek out informants who are frustrated or needy as well as those who are happy and secure, people at the center of the action and those on the fringes, and the leaders as well as followers.

Expect mixed messages and contradictions when you interview a range of informants. This is not a problem; it only reflects the multiple points of view that often coexist in a field site and provides a more complete if complex picture. Most people only reveal highly intimate, confidential information in private settings. Likewise, the field interview varies by its context. You generally want to conduct field interviews in the field member’s home environment to increase the comfort level. Context and past interactions between you and an informant can shape what is said and how, so you want to note those things as well as what the informant actually says. You also want to note nonverbal communication in the interview that can add meaning, such as a shrug or a gesture.

10.6: How Do You Conclude? 10.6 Apply strategies for planning the exiting process for a professional field researcher Thus far, you have seen how to plan, initiate, develop and execute a field research study, but no study continues forever, and how you manage a study’s closing phase is also important. You want to anticipate and smoothly transition toward the ending of a field research study.

10.6.1: Leaving the Field A professional field researcher may be in a field site from a few weeks to a several years. At some point, work in the field ends. It may end naturally—when learning new things diminishes and theory building reaches a closure. Alternatively, it could go on without end, and you must decide to cut off relations and exit. At times, external factors force an ending. You should plan for and anticipate the disengaging and exiting process. Depending on the intensity of involvement and the length of time in the field, exiting can be disruptive or emotionally painful. You may feel guilty and depressed immediately before and after leaving. You may find it difficult to let go because of personal and emotional ties. Professionals with a long, intense involvement in a field site sometimes need a month or more of adjustment time after exiting before they feel at home with their original cultural surroundings. The exit process depends on specifics of the field setting and relationships developed. You have to decide length and form of disengagement. You can leave quickly (simply not return one day) or slowly reduce your involvement over several weeks. You also need to decide how to tell members and how much advance warning to give. If you spend a lot of time in a field site and have intense involvement with members, you should give some warning of the exit. Try to

230 Chapter 10 fulfill any bargains or commitments you built up in the field so that you can leave with a clean slate. Often a small ritual, such as a going-away party or thanking and shaking hands with everyone, helps to signal the social break. Anticipate the effect exiting may have on members. Some members may feel hurt or rejected because a close social relationship is ending. They may try to pull you back into the field. Over time, as warm social relations develop and many experiences are shared, members may have forgotten that you were an outsider there for research purposes. Bringing the research aspect of your relationship to the forefront may cause members to grow cold and distant or to become angry and resentful. Some researchers continue to maintain social relations with people they got to know in a field site after the study ends. For example, in her study of luxury hotels, Rachel Sherman (2006) reported that a few years after she left the field site, she continued to get together socially several times a year with workers who she got to know during the research.

10.6.2: Writing the Field ResearchReport Field researchers start to think about what will appear in a report while they are still gathering data. More than in other types of social research, the researcher may write in the first person in the study report and recount his or her personal observations and experiences. More than in other reports, a field research report depends on the researcher’s writing skill to convey a feeling of the field site, to describe individual people in the field, and to recount events in great depth. Unlike quantitative research reports, field research reports do not follow a fixed pattern. Many field research reports are book-length or are long, descriptive articles. Tables with numbers, graphs, or charts are very rare. In the report, the researcher provides supporting data in the form of photos and quotes or short selections of concrete situations taken from the observations in field notes. The researchers use the quotes both to document and illustrate the concepts or themes that are part of the analysis.

Tips for the Wise Consumer Field Research Reports You look for different things in a report on a field research study or ethnography than in a quantitative research study. Here are some things to look for: • Exactly who conducted the study? Was it one person or several? What are the backgrounds or other characteristics of the person who conducted the study? • What is the field site? Exactly when and where was the study conducted? • Who constitute the members in the field site?

• How did the researcher gain access to the field site? Was it easy or difficult to gain access? • How long did the researcher spend in the field conducting observations? • Did the researcher supplement field observations with other types of evidence or documentation? • What social role and what researcher role did the researcher use? • Did the researcher conduct interviews for this study? • How did the author use data to back up statements about themes, concepts, or processes in the field site?

10.7: How Can You Be an Ethical Field Researcher? 10.7 Describe how a researcher resolves the ethical issues that come up during field research Your direct, personal involvement in the social lives of others when conducting field research introduces several ethical issues. Often you are alone in the field and must make a quick ethical decision about situations that appear unexpectedly in the field. Privacy is the most common ethical issue. As you gain intimate knowledge in a field site and people give you information in confidence, you incur an ethical obligation to uphold the confidentiality of data. You must keep it confidential from other people in the field as well as from the public. You may want to disguise real names in field notes and in a report for the public. New field researchers often ask about deception. When do you not fully and honestly disclose your role as a researcher and true purpose for being at a site? Professional researchers debate the merits of using covert versus overt field research. Everyone agrees that covert research is not preferred. Some say it is never acceptable. Others see covert research as acceptable and necessary for entering into and gaining knowledge about certain areas of social life, such as a secret society or ring of illegal drug dealers. In general, you should be honest and openly disclose why you are in a site whenever possible. This is especially true if you are a beginning researcher. Covert research raises ethical and sometimes legal issues. It is more difficult to maintain a false front and to be in a constant anxiety over being caught. Professional researchers who conduct field research on people who engage in illegal behavior face additional ethical issues as well as personal risk. They may know of and are sometimes indirectly involved in illegal activity. Such knowledge is of interest both to law enforcement officials and to other criminals. The researcher faces an extra challenge in building trust and rapport but not becoming so involved as to violate personal moral standards or endanger other people. In such situations, professional researchers often make

Observing People in Natural Settings 231

an explicit arrangement with the members—such as, I will leave when certain serious illegal behavior occurs. Field research with criminals is for experienced researchers who have extra training and a knowledge of the risks involved.

10.8: What Are Focus Groups and How Do YouUse Them? 10.8 Analyze the advantages and limitations of focusgroups

What do you think about a male elementary school teacher? Compare Your Thoughts Most elementary school teachers are women. A traditional female gender role includes close physical contact with young children and nurturing emotional relationships that are necessary for successful elementary school teaching. By contrast, the traditional male gender role is to be emotionally remote, engage in coarse or rough behavior, and avoid intimate physical contact, such as hugging and touching. A male elementary school teacher may have his masculinity questioned. People may see him as being weak and not ambitious. They may suspect he is a dangerous pedophile. Beyond the gender role issue, males may avoid the job because of its low pay and low status. It is socially defined as “women’s work” not appropriate for a “real man.”

Many observers note that highly gender stratified elementary schools perpetuate and reinforce traditional gender roles. Young boys and girls learn a traditional adult male model, one that they also see daily in the mass media and among many adults. They learn that a male is supposed to be emotionally cool and aggressive. They are indirectly “taught” that adult men are not emotionally expressive and caring. Having male elementary teachers promotes social equality and gives young children a positive male role model. Young children learn gender roles in which it is acceptable for an adult male to express emotional intimacy and be nurturing. Having male elementary school teachers also enables males who enjoy working with active young children and who do not repress the caring, emotionally warm side of their self to have a rewarding professional career. The focus group is a special qualitative research technique that differs from traditional ethnography. It does not require an extended period of detailed observation in a field site. It is similar in that you acquire qualitative data from a small number of selected people participating in a naturalistic open group conversation. Focus group research has rapidly grown over the last 20 years. In it, you informally interview people in a groupdiscussion setting. It can take place in a natural field setting (e.g., restaurant, break room of work site) or a special setting (e.g., classroom, conference room). To create the group, you gather together 4 to 12 people in a room with a trained moderator to discuss a few issues. The group members should be hom*ogeneous but not include close friends or relatives. Most focus group sessions are 45 to 90 minutes long. The moderator must be nondirective and facilitate free, open discussion among all group members. The moderator offers open-ended questions and does not let one person dominate the discussion. In a typical study, you might create four to six separate focus groups. All would discuss the same questions or focus on the same topic. To understand why men become elementary-level teachers and how they feel about the reactions of others, Cushman (2005) conducted a focus group study with male elementary teachers. Cushman (2005) contacted 17 practicing elementary teachers to participate in 90-minute focus group discussions. Three or four teachers were together in each group with a moderator. The semistructured discussion centered on several questions, such as the following: • What aspects of elementary school teaching initially attracted you to this career? • What sort of reaction did you get when you told f amily and friends? • Did you go straight from school or try other career options first? • Was the salary or status of an elementary teacher a concern to you?

232 Chapter 10 • Do you face any challenges being part of a school staff in which males are a minority?

• They are limited to discussing one or a few topics in a session.

• To what extent does having physical contact with young children concern you?

• A moderator may unknowingly limit full, open, and free expression by all group members.

Cushman found that all the teachers overwhelmingly chose the career because of the intrinsic joy they got from working with children. They found the play aspect of teaching young children to be most rewarding. Many also liked the strong service aspect of teaching. They tended to value intrinsic over extrinsic rewards (e.g., job satisfaction vs. salary). Reactions by others to their decision varied but were often negative. Fathers were the most negative about the career choice. Positive reactions were most common when the men had teachers in their families. Many men came to elementary teaching after having tried other career options and only after developing a strong sense of self and their major life priorities. All were aware that hom*ophobic communities made false accusations against male teachers. They knew that false accusations could have tragic consequences for a teacher.

• Focus group participants tend to produce fewer ideas than in individual interviews.

WHAT ARE SOME OF THE ADVANTAGES AND LIMITATIONS OF FOCUS GROUPS? Researchers have exam-

ined many topics using the focus group technique. Topics include issue attitudes (e.g., race relations, workplace equality), personal behaviors and relations (e.g., how to live with a child who has disabilities), a new product (e.g., breakfast cereal), or a political candidate. The focus group procedure has its own strengths and weaknesses. Advantages • They are fast, easy to do, and inexpensive. • They occur in a more natural setting that helps to increase external validity. • They provide exploratory researchers with new insights and give survey researchers ideas for questions and answer categories. • They give quantitative researchers a window into how people naturally discuss topics and aid in the interpretation of quantitative results. • They allow research participants to query one another and explain their answers to each other. • They encourage open expression among members of marginalized social groups who may not otherwise speak. • They help people to feel empowered by a group setting, especially in action-oriented research projects. Limitations • You cannot generalize the discussion outcomes to a large, diverse population. • They create a “polarization effect” such that attitudes often become more extreme after group discussion.

• A large quantity of open-discussion results can be difficult to analyze. • Reports on focus groups rarely report all the details of study design/procedure. • Researchers find it difficult to reconcile differences that arise between responses given by an individual-only interview and those from a focus group. MAKING IT PRACTICAL: USING FOCUS GROUPS TO LEARN WHY STUDENTS PICKED A PRIVATE HIGH SCHOOL Several years ago, I used focus groups as part of

an applied study on why parents and students chose a private high school. I formed six focus groups. Each group had 8 to 10 student volunteers from the high school. A trained college-student moderator asked questions, elicited comments from group members, and made sure that no one person dominated the discussions. The six groups contained male and female members from either one grade level or two adjacent grades (e.g., freshmen and sophom*ores). For 45 minutes, students discussed their reasons for attending the high school. They were asked about the importance of specific factors in the decision. These included parent pressure, participating in school sports, academic reputation, cost of the school, being with their friends, siblings or parents who went to the same school, school size, and the school’s religious orientation. We tape-recorded the discussions and then analyzed the tapes to understand what the students saw as most important to their decisions. In addition, the data helped us to interpret survey data on the same topic. We used the results to design three survey questionnaires. One questionnaire was for all high school students, another for a sample of parents, and a third for students at junior high schools that often sent students to the high school. Results were part of a written report and presented at an open meeting of the parents and the school’s administration.

WRITING PROMPT Focus Groups Imagine that you are part of a focus group consisting of 10 people. As soon as the moderator introduces the topic, two people begin to dominate the discussion. What is your response to make sure that everyone has a chance to talk? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

Summary: What You Learned about Observing People inNatural Settings You learned about the field research process, including its logic, choosing a site and gaining access, developing relations in the field, observing and collecting data, and conducting a field interview. You saw how field research differs from quantitative research; field researchers begin data analysis and theorizing as they collect data. In field research, you as the researcher are directly involved with the people you study. You become immersed in the social life of a natural setting to learn about it. More than quantitative research, doing field research can have a large impact on your emotions, personal life, and sense of self. Field research is a way to study parts of the social world that you otherwise could not study. To do good field research requires a combination of skills. You need a strong sense of self, an excellent ability to listen and absorb details, tremendous patience, sensitivity and empathy for other people, superb social skills, a talent to think very quickly on your feet, the ability to see subtle interconnections among people/events, and an ability to express yourself in writing. Field research is especially valuable for studying micro-level social life and face-to-face interactions among small groups of people who interact with one another. It is less effective when the concern is macro-level processes and social structures, such as events that occurred in the distant past or that stretch across decades. Historical-comparative research, discussed in the next chapter, is better suited to investigating these types of concerns.

Quick Review What Is Field Research? 1. Field research is a distinct approach to social research that yields qualitative data from direct observations in the field and the researcher’s participation in a natural social setting. 2. The several kinds of field research include ethnography, participant-observation research, informal “depth” interviews, and focus groups. 3. One type of field research, ethnography, originated in cultural anthropology. Its goal is to construct a highly detailed description and up-close understanding from the standpoint of a field site’s natives/members/insiders.

4. An ethnographer notes how people construct social meanings as ongoing processes in specific settings by observing both explicit and tacit cultural knowledge. 5. Explicit cultural knowledge is what we easily see and directly know. Tacit cultural knowledge remains unseen or unstated. It includes unspoken cultural norms and aspects of culture that we rarely discuss or acknowledge directly.

How Do You Begin to Conduct FieldResearch? 1. Naturalism is a guiding principle of field research because the research occurs in “natural settings” over which a researcher has little control and in which the researcher is personally involved. 2. Flexibility is principle of field research and researchers recognize opportunities, “play it by ear,” and adjust to changing situations rather than follow a fix sequence of steps rigidly. A field researcher selects techniques and changes direction to pursue interesting new leads as they appear, but this makes becoming sidetracked or drifting without direction a danger. 3. Direct, personal involvement with other people means a researcher’s emotional make-up, personal biography, cultural experience and personal characteristics (including physical appearance, gender, age, racial-ethnic background) are relevant. Field researchers have a well- developed sense of self but are not self-absorbed. 4. Before starting a field research study, it is important to conduct the standard academic literature review and to engage in a wide-ranging background investigation with multiple sources. It is also important to practice taking detailed observational notes. 5. Selecting a field site is a major decision and researchers often try several sites before settling on one. Factors that influence choice of a field research site include containment, richness of data, unfamiliarity, and suitability. 6. Gaining access to a field site can be a complex process. Possible sites vary from open access public areas to closed, private, or semiprivate settings. Almost all field sites have a gatekeeper. The researcher must identify gatekeepers and negotiate access with them. 7. Entering a field site requires a flexible plan of action tailored to the specific site and field researcher characteristics. At the stage of entrée, a researcher considers ways of presenting the self in the setting, the amount of personal background and research details to disclose, and a social role to adopt in the field.

233

234 Chapter 10

What Should You Do in the FieldSite? 1. When first in the field site, a researcher must learn “the ropes” (i.e., acquire an understanding of micro-level norms, rules, and customs) as well as how to cope with stress. 2. Field researchers try to normalize the social research to reduce discomfort and suspicion by field site members. 3. With time in the field, a researcher slowly builds trust and establishes rapport through repeated positive interactions. This requires listening sympathetically to complaints, sharing experiences, and swapping stories with field site members. Self-presentation, field site role, and events in the field can facilitate, limit, or destroy trust. 4. Field researchers use several strategies to manage relationships in the field, including performing small favors, appearing interested, exercising selective inattention, being an earnest novice, avoiding conflict, and adopting an attitude of strangeness. 5. One field strategy, the appearance of interest, has a researcher say things and engage in behaviors that display an appearance of having an interest in, being engaged with, and valuing field site members, even when the researcher is not truly interested or engaged. 6. Field researchers adopt an attitude of strangeness as a way to make unspoken, tacit culture of a setting more visible. A field researcher who adopts both a stranger’s and an insider’s point of view can see events from an outsiderstranger’s perspective and as an insider-member does.

How Do You Collect Data in the Field? 1. The field researcher is an instrument for acquiring field data. This means a researcher uses all senses and is sensitive to all that occurs in the field. A researcher’s social relationships, personal feelings, and subjective experiences become valuable field data. 2. A researcher scrutinizes aspects of the physical setting to capture its atmosphere and record subtle, unconscious signals that might influence human behavior. By noticing and assembling hundreds of trivial and daily minutiae, a researcher can uncover important features of a field site. What many people often overlook in a setting is what a field researcher learns to notice and from which he or she can discover a great deal. 3. The researcher observes each field site member’s observable physical characteristics: age, sex, race, and stature, and records specific, descriptive details of the setting and events. Also recorded are aspects of physical appearance—such as makeup, neatness, clothing, or hairstyle— and their specific actions. In addition, the researcher notes and records the context in which events occur. Lastly, the researcher notices exactly what people say, and how they speak—their tone and specific words and phrases. 4. Field researchers are prepared to take advantage of unplanned, unexpected events that can reveal much

about a field site and are prepared to use casual eavesdropping or observing chance events not meant to be public as a way to learn about a field site. 5. Field researchers are sensitive to the flow of time in a field site and appreciate wait time. Noting time usage can indicate many details about the dynamics and rhythm of a field site. Wait time also communicates commitment to members of a field site and can earn respect from field site members. 6. Field research sampling differs from that of survey research. Field researchers sample by taking selective observations from all possible times, locations, people, situations, types of events, or contexts of interest. 7. Field research data include memories and field notes, the permanent record of observations and experiences. Field researchers rarely take notes while in the field site, except for short phrases as jotted notes. They usually leave the field to write notes after observation, and may divide the field notes into several types. 8. Researchers often supplement field notes with maps, photos, or recordings. Audio or video recordings are not substitutes for written field notes and have many limitations. 9. Field researchers use unstructured, nondirective, in-depth interviews that differ from formal survey research interviews. The field interview is a joint production between a researcher and an informant, a person with whom the researcher has a relationship. Often field research interviewing involves asking several types of questions, with the type varying by stage in the research process.

How Do You Conclude? 1. Researchers plan for and anticipate the disengaging and exiting process. The exit process depends on specifics of the field setting and relationships developed. It can be disruptive or emotionally painful, and researchers decide on length and form of disengagement. 2. A report on field research depends on the researcher’s writing skill to convey a feeling of the field site and recount events in great depth. Unlike quantitative research reports, the reports rarely follow a fixed pattern. Field research reports are booklength or are long, descriptive articles. Tables with numbers, graphs, or charts are infrequent. Often reports include photos, maps, and selections taken directly from field notes.

How Can You Be an Ethical Field Researcher? 1. Direct, personal involvement in the social lives of others in field research introduces several ethical issues that a researcher must resolve alone in the field. Privacy is the most common ethical issue because the researcher gains intimate knowledge of a field site and its members. To protect data confidentiality, researchers often disguise real names or places. 2. Researchers who conduct field research on people who engage in illegal behavior face additional ethical issues as well as personal risk. Police and others may be interested in their research or data, so they must take extra precautions.

Observing People in Natural Settings 235

SHARED WRITING: PARTICIPANT OBSERVATION Field research has the researcher being a participant in a social setting that he or she carefully observes. Critics charge that this close personal involvement, often over months or years, while logging in hundreds if not thousands of hours of conversations and observed interactions, creates a bias of familiarity. It yields unreliable qualitative data that other people cannot verify. Proponents argue that being in real social settings and getting to know people who are going through their actual daily lives provide the most authentic and richest form of social research data possible. Choose whether you are for or against participant observation. Defend it either as (1) contaminated by personal involvement or (2) a powerful form of social research. Next, read and respond to two classmates who took the opposite position in a constructive way by describing what

impact your classmates’ opposing view have upon how you approach your own view (e.g., reading this argument strengthened my view because or reading this argument made me reconsider my position because . . .). Be specific and cite key takeaways from the opposing arguments, follow up with questions, and/or offer alternatives. A minimum number of characters is required to post and earn points. After posting, your response can be viewed by your class and instructor, and you can participate in the class discussion.

Post

0 characters | 140 minimum

Chapter 11

Looking at the Past and across Cultures

Learning Objectives 11.1 Synthesize data to aid in explaining and

understanding micro to macro-level events for historical-comparative research 11.2 Identify the six stages of most historical-

comparative research studies 11.3 Compare the scope and goals of the

11.4 Explain how historical-comparative

research is significant in research related tothe study of different cultures 11.5 List the responsibilities of a historical-

comparative researcher in upholding ethical principles

historian and the historical-comparative social researcher Factors other than crime rates influence the rates at which countries send people to prison. After a century of stability, incarceration rates dramatically increased across the United States in the 1970s, until the United States rose to be the

236

world leader in locking away people in jails and prisons. With less than 5 percent of the world’s p opulation, today the United States has over 20 percent of the world’s prisoners. Moreover, the past 40 years has seen a “sea change in

Looking at the Past and across Cultures 237

penal regimes among modern Western societies, resulting in more punitive social policies in general and a trend toward higher incarceration rates in particular” (Sutton, 2013, p. 715). What accounts a new era of “mass incarceration” that is most evident in the United States but found in other advanced countries? Two recent h istorical-comparative studies examine this societal change by looking back over four decades and making comparisons, comparing across states within the United States and across nations. The following two studies look at the same phenomena over the same historical period to offer overlapping explanations for different units of analysis. They illustrate both the methodological variation and explanatory power of historical-comparative research. “The Transformation of America’s Penal Order” by Campbell and Schoenfeld After little variation for nearly a century, according to research conducted by Campbell and Schoenfeld: “Between 1977 and 2000, every state’s incarceration rate more than doubled, with an average increase of 285%.... By 2008, an astounding one in every 100 Americans was in jail or prison, including one in every 15 African-American adult men” (2013, p. 1375). The researchers examined qualitative data to explain incarceration growth by comparing states within the United States. They used historical records, legal documents, news reports, and interviews to track how talk about the “crime problem” and policy solutions for it shifted in each of eight states over time. From the eight state-level case studies as well as changes in national-

level narratives about crime and punishment, they identified three distinct periods within the 40-year era. In the initial, contested period, rising crime rates were an important factor in destabilizing the pre-1960 penal order and in shaping policy outcomes. In a second period after a political consensus had solidified around a new penal order, crime rates ceased to be a significant explanatory factor; State-specific factors—political traditions and institutions, partisan control, and early policy choices—explained state-level variation in incarceration. Political parties and special interest groups combined to push state-level lawmakers to redefine the crime problem as being due to too little punishment. State-level lawmakers responded with harsh new laws and policies to incarcerate many more people for longer lengths of time. The rising national political influence of Sun Belt states drove candidates for U.S. president to exploit for political gain popular southern cultural narratives that linked criminal behavior to a person’s racial background. America’s highly decentralized criminal justice and crime policy system enabled state-level political factors to reshape penal policy. The third phase was marked by the growing power of the Republican Party. The party abandoned its prior advocacy of strict fiscal conservatism and shifted to favor prison expansion and a harsh “law and order” stance. Yet, soon the political party in power ceased to matter because the Democratic Party changed course and it too embraced the same crime control policies as the Republicans. In sum, Campbell and Schoenfeld explained the transformation of America’s penal order as caused by succession penal policy changes that flowed back and forth between the statelevel and national-level, and that were primarily shaped by political factors at both levels (see Figure 11.1).

Figure 11.1 Increase in U.S. Imprisonment Rates over Time Source: Bureau of Justice Statistics Prisoners Series.

1,600,000

U.S. State and Federal Prison Population, 1925–2013 2013: 1,516,879

1,400,000

1,000,000 800,000 600,000 400,000 200,000

2013

0 1925 1928 1932 1936 1940 1944 1948 1952 1956 1960 1964 1968 1972 1976 1980 1984 1988 1992 1996 2000 2004 2008

Number of People

1,200,000

238 Chapter 11 “The Transformation ofPrison Regimes in Late Capitalist Societies” bySutton John Sutton (2013, p. 717) said he was, “less concerned with the genealogy of the contemporary penal regime than with the generalizability of the regime change argument and its implications for the comparative sociology of crime and punishment.” His study sought to explain a dramatic shift to higher incarceration rates in the United States, but also explain why incarceration increased in other advanced nations over the past 40 years. He focused on the hypothesis that economic forces, namely deindustrialization, combined with a spreading “neoliberal” political ideology, with its promarket reforms—d eregulation, fiscal austerity, and welfare retrenchment— caused more imprisonment. The neoliberal “reforms” occurred in many advanced capitalist nations after the 1970s and disrupted employment opportunities that had previously been available to many low-skilled males. Governments adopted new, harsher penal policies to contain and manage a growing underemployed population produced by the employment disruptions. Sutton analyzed quantitative data from 15 high-income capitalist democracies, 1960–2000, with the incarceration rate, i.e., the number of inmates per 100,000 population, as the dependent variable and statistically controlled for other variables, including crime rates. Sutton found that despite a general rise in incarceration rates, they varied widely by nation. This led him to reject a globalization-convergence explanation. It said the penal policies changed in the same way across all advanced nations. He found that the largest increase in incarceration rate occurred in nations that had the weakest labor unions and most meager national social-welfare programs to assist poor people. He concluded, “Put more simply, the increase in average incarceration rates was concentrated among countries with the least regulated labor markets” (p. 740). In short, nations with the most “free market” forms of capitalism, a decentralized government, and minimal social programs, are also the ones that showed the greatest instability in employment opportunities for less-skilled young males in the labor force. Nations in which declining opportunities for stable employment combined with minimal social-welfare protections are also those that saw the fastest growth in incarceration, with less-skilled young males most likely to be incarcerated. By contrast, nations with strong labor unions and extensive social-welfare protections had less labor market instability. Their incarceration rates only rose modestly if at all.

11.1: HistoricalComparative Research 11.1 Synthesize data to aid in explaining and understanding micro to macro-level events for historical-comparative research Historical-comparative (H-C) research is the most relevant research method for explaining and understanding macrolevel events—the rise of incarceration rates, growing social inequality, societal patterns of racism, large-scale immigration, violence based on religious hate, or urban decay. The major nineteenth-century founders of the social sciences, such as Emile Durkheim, Karl Marx, and Max Weber, used this method. Researchers used it to study issues in many macro-level topic areas—such as social change, political sociology, social movements, and social inequality—and in other areas—such as health care, criminology, gender relations, race relations, and family. H-C research has grown in popularity over the past three decades. H-C studies are ideal for addressing “big questions,” but some people question their practical value because they address basic rather than applied research questions. In addition, H-C studies can take a long time to complete, typically requiring several years, while many other social research studies can be finished within several months. On the other hand, there are often no alternative ways to answer some significant research questions. H-C studies also provide innovative ideas in fields, such as education, business, law enforcement, and medical care. Long forgotten practices or practices used in other cultures can stimulate exciting new ways to approach current issues and solve problems. In addition, the methodological issues of H-C research have wider implications for other research techniques. H-C research relies on a blend of research techniques. Some are like traditional history, some are like field research, and others extend quantitative data techniques such as surveys or existing statistics research. We will focus on H-C research that places historical time and/or crosscultural variation at the center. H-C research is most appropriate when a research question involves past events, and/ or two or more sociocultural contexts. H-C researchers look at how a specific mix of diverse factors have come together in time and place to generate a specific outcome (e.g., ethnic cleansing). They compare entire societies to see what they share or examine the same social process across several cultural or historical settings. Many people enjoy reading H-C studies because they learn about distant places or the past. However, H-C studies can be difficult to follow without sufficient background knowledge about history, geography, and other cultures.

Looking at the Past and across Cultures 239

H-C studies often assume that readers will have a minimal level historical-geographic literacy and cultural knowledge. Most H-C researchers use a variety of data to examine a central issue and situate it in a context. Some rely heavily on qualitative data, others quantitative data or mix data types. Those who gather qualitative data and focus on issues of culture are like field researchers. They want to see events through the eyes of the people they are studying. They examine specific individuals or groups and are sensitive to specific historical or cultural contexts. As with field research, and unlike quantitative data research, many H-C researchers only make limited generalizations.

Example Study Ethnic Cleansing

data are consistent with the patterns found in cross-national quantitative studies of political violence. Mann developed and outlined a theory of ethnic cleansing in the form of eight theses in the first three chapters (out of 17 in the book). The theory explains a range of outcomes of ethnic-racial relationships from peaceful multiculturalism to genocide. He applied the theory in each of the 13 case study chapters, noting the strength of empirical support for the theory, or qualifications for each thesis of the theory. The book’s core argument is that ethnic-cleansing and genocide is a modern invention found in countries organized along ethnonationalist lines and occurs when governments move toward a democratic form. Mann explained the apparent paradox that modern democracy contributes to the atrocities of ethnic cleansing by noting that democracy means rule by a majority of “the people.” When a nation defines “the people” to be members of a specific ethnic-racial group, the majority can be mobilized to tyrannize minorities under specific ideological, political-military, and economic conditions. When it occurred, murderous cleansing was rarely the initial intent; rather, the mass killing emerged through a sequence of increasingly aggressive stages, each in response to particular social-political conditions.

WRITING PROMPT Genocide Recall the study Mann conducted on genocide. What struck you most about how he went about the study? What are the possible problems with the data he used as evidence? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Mann (2005) conducted a study of ethnic cleansing, genocide, and related mass deaths worldwide from the eighteenth through twentieth century using qualitative historical sources. He considered “genocidal democracies” of the New World, including Mexico, the United States, and South West Africa, but the study concentrated on several better-known twentiethcentury genocides—the Armenians, Yugoslavia, Nazi Germany, and Rwanda. He also examined nonethnic mass deaths of millions caused by communist regimes—Stalin’s Soviet Union, Mao’s China, and the Khmer Rouge in Cambodia, and ethnic conflicts in India. The study is filled with a great many specific references and footnotes. For example, to document one paragraph about specific cities in which T urkish officials deported or killed large number of the Christian, Armenian minority in 1916– 1917, a footnote (note 8 on page 160) refers to pages in six book-length academic studies published between 1972 and 1998 and many selections from U.S. government documents.

Read More Mann meticulously anchored detailed information on particular locations, people and events that constitute the study’s evidence to mountains of specific documents, academic studies, government reports, and published speeches or official findings in English, French, or German. He noted that findings from his qualitative

Submit

11.1.1: How Are Field Research andH-C Research Alike? Let’s consider similarities between H-C research and field research. • They incorporate an individual researcher’s point of view as part of the research process. • They examine a great diversity of data types (diaries, maps, official statistics, newspapers, novels). • They focus on processes, time passage, and sequence. • They use grounded theory. • They make limited generalizations. An individual researcher’s characteristics, place in history, and geographic-cultural situation may influence the research process. The time, place, and culture in which the study was conducted affect data collection and interpretation. Both H-C and field research recognize this feature. It is the reason why in both, you may see specific biographic

240 Chapter 11 details of the person who conducted the research included in the study. In both types of research, you immerse yourself fully in a huge amount of qualitative data. The goal is to understand in depth the lives, language, and perspective of the people you are studying so that you can acquire an empathic understanding of them. You try to capture their subjective feelings and details of their daily lives. You narrow the focus to specific areas for analysis only after an immersion in the data. As you acquire an up-close understanding of the people studied, you “translate” their worldview for the readers of your research report. In field and H-C research, you devote attention to processes, time passage, and sequence. People or social life are in motion, not static or unchanging. You treat the passage of time (micro-level clock time as well as long-term historical time) as an essential aspect of the data. All researchers begin with concepts and ideas, but in H-C and field research new ideas often emerge during data collection and analysis in a process of grounded theory. As you gain greater knowledge about a particular place and time, it may be difficult to make broad generalizations that apply to all places or times. Typically, field and H-C research offer more limited generalization than quantitative research.

11.1.2: What Is Unique about H-CResearch? Despite its many similarities to field research, H-C research has several distinct features. Conducting research on the historical past and another culture differs from doing field research in the present in your home culture. The H-C researcher learns to do each of the following four things: 1. Work with limited evidence. 2. Interpret evidence with minimum distortion. 3. Integrate the micro and macro levels. 4. Use specific as well as transcultural, transhistorical concepts. Build on Limited and Indirect Evidence In all

studies, you construct an understanding of social life based on the empirical evidence you gathered. Historical evidence depends on data that have survived from the past. Even if you have many excellent historical documents, you are limited to what has not been destroyed and what has left a trace or other evidence behind. You cannot directly observe or be involved in the past. In comparative research, only a native member of two cultures can grasp all the similarities and differences. You may learn another language, study a new culture, and spend time in that culture; but unless you grew up in and fully absorbed the other culture, your understanding will always be limited to that of an outsider. To see and feel like a native in both cultures, you must be truly bicultural.

Interpret the Meaning of Events in Context

Data in H-C studies are rarely simple and unambiguous. The data usually contain multiple messages from which you extract meaning. Do not expect to get a full understanding based on a quick first glance of the evidence. Immerse yourself in it, absorb its complexity, and place it in context. Only after you reflect on the evidence and consider its many meanings can you interpret its possible significance. Perhaps you want to conduct a study on family relations of 120 years ago or in a distant country. To start, you need to learn the social context such as the nature of daily work, forms of communication, and transportation technology. You may want to study maps and events of that time and place. You want to be aware of local laws, the nature of health and medical care, types of foods eaten, daily household tasks, and common social practices. You also need to recognize kinship customs and obligations of the time. You may have data for a simple event, “the visit of a family member.” To put the event in context, you need to realize that the roads are made of dirt and mud, traveling is by foot, no one can call ahead of time, and there are few places to stop and rest. Without a full immersion in the evidence, you cannot grasp the meaning of “visit of a family member.” What are the three common types of distortion that you need to guard against? Compare Your Thoughts Supracontext Awareness. As a researcher today, you may be aware of events that people in the immediate setting that you are studying do not know about, such as events that occurred later in time or in other cultures. Knowing things that the people who are studying do not know could distort your understanding of their behavior and decisions. For example, you conduct a historical study people living in colonial America in 1760. As someone in the 21st century, you know that there will soon be a war of independence from Britain and the colonialists will win. However, the people you are studying in 1760 did not know that. Their decisions and actions may appear unwise or ill-advised based on your knowledge of what happened later in time. To understand people in the past, you must learn to see situations as they saw them. From their perspective, things may have looked different. Perhaps you conduct a comparative study, and become familiar with another culture in addition to your own, but the people you study know only their own culture. For example, you notice a food spoilage problem. The people you study spend lots of time storing food, putting it in jars in dark rooms, and often they get sick from spoiled food. You may ask: Why don’t they store food by wrapping it up as the people of my home culture do? It takes less time and the food rarely spoils. It would be inappropriate to evaluate how

Looking at the Past and across Cultures 241

people live in a different culture using knowledge that is not available to the local people. Coherence Imposition. Most of us like consistency and order. If you impose the expectation that people engage in predicable behavior and hold stable beliefs, you may introduce distortion. Prepare to study events or people who are contradictory or have “loose ends” that do not neatly come together. Try not to impose your own sense of order to make people’s beliefs or actions coherent, noncontradictory, and consistent more than they actually are. For example, you study people having a birthday party. You expect a clear beginning and ending, with the person having a birthday at the center of attention. You see people wandering in or out, one by one, without signaling that they are leaving, and you cannot even tell whose birthday it is. From an outsider perspective, you may want to organize and impose boundaries around the event, yet the people you are studying do not see or experience the situation in that way. A good H-C researcher does not impose an external sense of order but tries to describe how the people being studied see/experience the world, any order comes from their perspective. Capacity Overestimation. People learn, make decisions, change direction, and act on (or fail to act on) what they learn. However, people have a limited capacity to learn, make decisions, and modify the course of events. It is easy to overestimate the ability of people to act. Recognize that people may not quickly take actions. For example, you think, “Those parents could have met with the teacher about their child’s problem.” However, from the parents’ life situation and point of view, that may not be an easy or realistic option. Do not become frustrated and say, “They could have done X; why did they not act?” Try to grasp their point of view and recognize the constraints they feel.

Learning from History Conditions in Medieval WesternEurope

Often you need background information to read H-C studies and put them in context. Perhaps you read a study about Western Europe in the 1500s. Most travel was by foot or cart on mud roads, so travel difficulties meant few people had experience with the world beyond a couple of days journey away, the equivalent of 200 kilometers or 60 miles. You may already have learned that roughly one-third of the population (75 million people) died in the Black Death and about 95 percent of the population in that era could not read or write. However, did you know that using the number zero was considered the devil’s work, using Arabic numbers (the ones we use today) instead of Roman numerals was forbidden, and only a handful of math experts could do simple multiplication or division (what you learned in third or fourth grade). This lack of math skills profoundly affected business, accounting, and engineering. Basic banking—lending or counting money for buying and selling goods—was so primitive that it slowed the growth of trade. Most people considered loaning money for interest or making profits slightly immoral or possibly illegal. This example shows the importance of background knowledge. If you look at some past situation and ask, “Why didn’t people do things differently?” you often discover the answer in the many differences from the present. For example, you may find it unusual that people looked up to a local 29-year-old scribe as if he was a respected wise man when his scribe skills equaled those of a bright sixth-grader today. This makes sense in a social world where only 5 percent of adults had basic literacy skills (vs. over 95 percent today) and life expectancy was about 40 years old (vs. nearly 80 years old today).

WRITING PROMPT Transhistorical Concepts Give an example of a universal or “transhistorical” concept. The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit Integrate the Micro and Macro Levels H-C

researchers often examine and integrate data from both the micro (small-scale, face-to-face interaction) and macro (large-scale social structures) levels. For example, you read diaries or letters to get a feel for the everyday lives of individuals who lived in the distant past. You learn about the food they ate, their recreational pursuits, clothing, sicknesses, relations with friends, and so on. You link this microlevel view to macro-level, societal-wide processes (increased immigration, mechanization of production, tightened labor markets, and the like). Perhaps you want to compare schooling in two cultures. You visit and spend time in classrooms, talk to students and teachers, and devote hours to learning the daily routines and micro-culture of schools. In addition, you study the overall structure of education in each culture,

242 Chapter 11 such as requirements, graduation rates, numbers and types of schools, the books and tests used in schools, official curriculum guides, teacher training requirements, and so forth. You then integrate the micro-level or face-to-face life in classrooms with the macro-level structure of the national education system in each culture. Use Specific and Transcultural, Transhistorical Concepts We use many concepts to study

and think about the social world. Imagine them on a continuum. At one end are universal concepts. They apply across social settings, historical time, and cultures. They are transcultural or transhistorical. You can use the same idea to examine all times and all cultures. At the opposite end are concepts that apply only to particular social settings, cultures, or historical eras. Of course, many concepts fall between these extremes.

Summary Review Table 11.1 A Comparison of Approaches to Research Topic

Both Field and H-C Research

Quantitative Research

Researcher’s perspective

Include the researcher as an integral part of the research process.

Remove the researcher influence from the research process.

Approach to data

Become immersed in many details to acquire an empathetic understanding.

Precisely operationalize variables.

Theory and data

Use grounded theory, create a dialogue between data and concepts.

Compare deductive abstract theory with empirical data.

Present findings

Translate a meaning system to others.

Test specific hypotheses.

Action/structure

People construct meaning but do so within social structures.

Social forces shape people’s behavior whether or not they are aware of them.

Laws/generalization

Make limited generalizations that depend on context.

Discover universal, context-free general laws.

Summary Review Table 11.2 Features of a Distinct H-C Approach to Doing Research Topic

The Historical-Comparative Researcher

Evidence

Reconstructs from many fragments and incomplete evidence

Distortion

Guards against using own awareness of factors outside the social or historical context

Human role

Includes the consciousness of people in a context and uses their motives as causal factors

Causes

Sees cause as contingent on conditions, hidden beneath the surface, and due to a specific combination of factors

Micro/macro

Links the micro to macro levels or layers of social reality

Cross-contexts

Moves between concrete specifics in a context and across contexts for more abstract comparisons

A universal concept, such as fear, exists in all societies in all eras. Specific concepts may be found in one historical era or culture but in few if any others. Perhaps one culture has as an event marking a girl’s 15th birthday as a major signal of the end of childhood and her being ready for marriage and childbearing, such as the Quinceañera in many Latin America cultures, but in many other cultures such events are absent or very different. H-C researchers use both types of concepts. Sometime an event, activity, idea, or

social situation is unique to a time or place. The H-C researcher recognize and explain it in context with limited generalization. At other times, they can use universal concepts that permit comparisons across time and culture and build broader explanations. Quantitative studies often use transcultural, transhistorical concepts; however, this can create problems if the concepts are not first evaluated to determine whether the concepts are appropriate to apply in different cultures or historical eras.

Looking at the Past and across Cultures 243

11.2: How Do You Conduct a Historical-Comparative Research Study? 11.2 Identify the six stages of most historicalcomparative research studies Now we will discuss how to conduct qualitative H-C research. Like field research, you do not rigidly follow a fixed set of steps, but most H-C research studies proceed through the following six stages: • Acquire the necessary background. • Conceptualize and begin to focus. • Locate and evaluate the evidence. • Organize the evidence. • Synthesize and develop concepts. • Write the report.

11.2.1: Acquire the Necessary Background A preliminary step, unless you already have a great deal of knowledge about a setting, is to learn as much as you can about a study’s setting by engaging in an orientation reading. To start a serious study of a different culture, you will want to complete a few courses that focus on that country/culture and begin to study its language if you do not already speak and read it. Also read novels, biographies, travel guides, and general interest articles on the country/culture. If you have not visited the country, you should plan to spend some time exploring it. To start a serious study of a past historical era in your own country/culture, complete a few history courses, covering time periods during, before, and after the era of interest and on various historical topics like the country’s political history, social history, economic history, and world history of the era. Also ask professional historians for recommendations of general books to read as well as relevant historical biographies and novels. Visiting museums with exhibits on the historical era or topic of your study is often worthwhile. It is also useful to learn about the art, music, and popular entertainment of a cultural or historical setting.

11.2.2: Conceptualize and Begin toFocus Early in the process, think through the topic and develop clear definitions for the major ideas you will use. You may begin with a general topic (e.g., ethnic

cleansing) and a few ideas (meaning of an ethnic group, various forms of government), but as you acquire background knowledge and start to gather preliminary data, you want to think about how to focus on a specific issue. As an issue becomes more focused, it will direct you toward relevant evidence but you want to remain flexible. It is impossible to begin research without some assumptions, concepts, and theory—so you want to be aware of what they are. Allow concepts and evidence to interact and stimulate the direction you take. As in field research, you can change direction based on what you learn from the data. As your grasp of the details of a specific setting grows, continue to adjust or refine the concepts. Create new organizing concepts, subdivide the main issue, and develop lists of questions to investigate. Often you find that the data do not fit neatly with the original concepts, and you must revise them or seek additional data. For example, you plan to study a type of restaurant in the distant past or in another culture. You might begin with a few concepts, such as dining pleasure, consumer choice, price competition, type of customer, and so forth. Perhaps you discover that the restaurant does not always wash dishes in clean water, all customers are neighbors who live within a 15-minute walk of the restaurant, there is no written menu, everyone who visits the restaurant seems acquainted with one another, and everyone knows the price of each dish served. You may need to rethink what is happening, develop new ideas to make sense of the situation, and question whether the concept of “restaurant” is most appropriate.

11.2.3: Locate and Evaluate theEvidence You will need to do a lot of bibliographic work, especially for historical research. Historical research requires using indexes, catalogs, and special reference works that list what libraries or other sources contain. Professional researchers may spend many months searching for sources in libraries, travel to different specialized research libraries, and read dozens, if not hundreds, of books and articles. Taking detailed notes on what you read is essential. For comparative research, you must focus on one or more specific nations or areas. Comparative research often requires learning a foreign language and/or travel to another country and then establishing local contacts. Once you find evidence, you need to evaluate its accuracy by keeping two questions in mind: • How relevant is the evidence to emerging research questions and evolving concepts? • How accurate and strong is the evidence?

244 Chapter 11 Your research focus often shifts. As this happens, evidence that was once relevant becomes less relevant and previously ignored evidence may become highly relevant. Also, you want to constantly evaluate alternative interpretations of the evidence and look for silences. For example, you study a group of leading male merchants in the 1890s. You find a lot of evidence and documents about them and their business dealings. However, there is no evidence about their wives and many servants—who are invisible in the data. To assess the whole situation, you want to notice both what is clearly documented and what occurs but remains silent in the data.

11.2.4: Organize the Evidence As you gather and locate sources of information, you also organize the data. Obviously, it is unwise to take notes madly and let them pile up haphazardly; instead you sort, label, and categorize. Begin with a preliminary analysis by noting themes in the mass of details. As you sort and label, reflect and nurture insights that can stimulate new ways to organize data and new questions for your study. Let data and theory interact and influence one another. Evaluate the evidence based on emerging ideas or theory. Your thinking on an issue can advance as you re-examine and reorganize evidence. This occurs because you use newly created ideas to look at old data in new ways, and the emerging ideas will guide your search for additional data.

11.2.5: Synthesize and Develop Concepts Once most of the evidence is in, move toward creating an integrated explanation. You want to synthesize and pull the parts together into one story. As you read and re-read your notes and you sort and re-sort them based on various organizing schemes, look for new connections. Try to see the evidence in new ways. By looking for patterns, you can draw out similarities and differences. You might organize events into sequences and group them into a step-by-step process. You can synthesize by connecting a body of evidence with an abstract concept or causal mechanism. Many researchers find metaphors useful. For example, you note that relations between a foreman and workers are “like an emotional roller coaster drop” in which things seemed to be getting better and moving higher and higher, and then there is a sudden letdown after expectations have risen very fast. You can use metaphors as organizing and sensitizing devices.

11.2.6: Write the Report Assembling evidence, arguments, and conclusions into a written report is a crucial step in all research. If anything,

this step is more important for qualitative H-C than for quantitative studies. A carefully crafted, well-written report will often “make or break” the success of H-C research. You must distill mountains of evidence into clear exposition, document numerous sources with extensive footnotes, and weave the evidence and arguments together in a manner that communicates a coherent, convincing picture. You gathered mountains of specific details in the study but can rarely include more than a few critical examples of the raw evidence. These illustrate and give credence to your larger story. Achieving a good balance between generalization and documented specific details can be difficult. As you do this, you want tell readers a dramatic, compelling story.

11.3: How Do You Conduct Research on the Past? 11.3 Compare the scope and goals of the historian andthe historical-comparative social researcher The past began 5 minutes ago, but usually something is at least 10 years in the past before we call it history. After about 10 years, direct experience has faded and perspective shifts. The word history is confusing because of its three meanings: 1. Events that occurred in the past (e.g., it is history that the French withdrew troops from Vietnam); 2. A documented record of the past (e.g., I read a history of French involvement in Vietnam); 3. An academic field in which specialists study the past (e.g., a course in the department of history).

Specialists in the field of history, or professional historians, devote most of their time and efforts to locating, gathering, and analyzing original historical data. They often use specialized techniques. Nonhistorian social researchers rarely “do history,” but they too examine historical data—that is, evidence about actual past events, including original records. Compared to a professional historian, social researchers consider a wider scope of evidence and have different goals. Let us look at the contrast in goals and activities between the historian and the H-C social researcher. The typical historian sees collecting highly accurate historical evidence as a central goal in itself, interprets the data’s significance in light of other historical events, and is not overly concerned about developing a theory to explain social relations or processes.

Looking at the Past and across Cultures 245

By contrast, the H-C social researcher treats gathering carefully documented, accurate, and highly detailed descriptions of specific past events as an important but secondary concern; wants to extend or build a theory or apply social concepts to new situations; and uses historical evidence as a means to document, explain, and understand general social relations.

11.3.1: Types of Historical Evidence Social researchers and historians both draw on the following four types of historical evidence: • Primary sources • Running records • Recollections • Secondary sources The historian’s main goal is to locate, collect, validate, and analyze the first one on the list—primary sources. By contrast, a social researcher will look more at secondary sources or running records. Both use recollections. The letters, diaries, newspapers, magazines, speeches, movies, novels, articles of clothing, photographs, business records, and so forth from people in the past that have survived into the present are called primary sources. You can find them in official archives (a place where documents are stored), in private collections, in family closets, or in museums. A widely used primary source is a published or unpublished written document. Documents may be in their original form or in preserved form on microfiche or film. They are often the only surviving record we have of the words, thoughts, deeds, and feelings of people in the past. A classic primary source is a bundle of yellowed letters that a traveling businessman wrote to his wife and that a historian discovers in an attic 75 years after both the husband and wife have died. A limitation of written sources is that elites or people in official organizations write most of the documents. It is easy to overlook the views of the illiterate, the poor, or people outside official institutions. For example, during early nineteenth century, in the United States it was illegal for slaves to read or write. You will find it difficult to locate written sources on how a slave actually experienced slavery. By contrast, most slave owners could read and write. Written documents from the slave era tend to give a slave owner’s rather than the slave’s point of view. Making It Practical: Old Newspaper Articles as Sources A widely used and easily accessible pri-

mary source is the newspaper. A public source is the U.S. Library of Congress, “Chronicling America” is a database that allows you to search newspaper pages from 1836–1922.

For example, you are interested in articles on the topic of immigration between 1900 and 1910. You quickly find the term appeared on 22,551 front page articles, so you narrow the topic. Let us say you want to find out about immigration from one country, Japan. You discover 246 articles discussing the topic in the front pages of newspapers during those 10 years. As you narrow your search, you find an article in the New York Sun published January 2, 1908 that is of interest (see Figure 11.2). As you read, you will find and may notice old-style spelling and that Japan and the United States have a disagreement. The Japanese are called “coolies” entering the United States, specifically the West coast of the United States, in large numbers and Congress might block entry. Japan volunteered to restrict who left Japan to avoid, such as Tokio for Tokyo, action, and entry of Japanese to other U.S.-controlled territories—the Philippines, Hawaii, and the Panama Canal zone could continue. Japan issued special passports to allow entry into U.S. territories but not the continental United States. To understand everything, you need background knowledge. From a review of American history, you would learn of hostility toward Asians and rioting against Chinese in Pacific states in the 1880s. This turmoil led to the 1882 Chinese Exclusion Act, which ended most immigration from China. A look at the Statistical Abstract of the United States for 1900 (which is available online) shows you that immigration from all of Asia was very low—fewer than 4,400 entered in 1890—less than from the tiny country of the Netherlands in the same year. However, between 1890 and 1900, immigration from Asia increased to 17,000. From background knowledge, you might suspect that the increase around 1900 may be people from the Philippines. This is because the United States had made the Philippines a colony in 1899. The Statistical Abstract is not very useful. It divided immigration information on Asia into China and the rest of Asia. From it, you would see that the number coming from China dropped between 1890 and 1900. A look at more detailed U.S. Census records will show you that just 55 Japanese lived in the United States in 1870 (excluding Hawaii). This number rose to 2,000 by 1890 and 24,000 in 1900 and 72,000 in 1910. If you pursue the topic with secondary literature and books on the topic, you will find about 10 books written on pre1920s immigration from Japan. So by starting with the general topic of immigration and a 10-year time period, you located a primary source, a 1908 newspaper article. From the one newspaper article, you could develop a more focused search and move toward a research question. Why and how did Japan control which of its people immigrated to the United States? Did other countries try to control who came to the United States or was this unique to Japan?

246 Chapter 11

Figure 11.2 New York Sun Article on Japanese Immigration An example of a primary historical document http://chroniclingamerica.loc.gov/lccn/sn83030272/1908-01-02/ed-1/seq-1/#date1=1908&sort=relevance&rows=20&words=IMMIGRATION+JAPAN&searchType=b asic&sequence=1&index=7&state=&date2=1908&proxtext=jAPAN+iMMIGRATION&y=0&x=0&dateFilterType=yearRange&page=3

Looking at the Past and across Cultures 247

and may have to travel to archives or specialized libraries. Once you arrive, the sources, such as newspapers, diaries, letters, memos, and other records, often are stored in a dusty and rarely visited room filled with stacked boxes that contain fading documents. The documents may be incomplete, disorganized, and in various stages of decay. After locating relevant documents, photos, records, and other primary sources, you must evaluate them using external and internal criticism (see Figure 11.3).

WRITING PROMPT Primary Sources When looking at a “primary source” (i.e., very old magazine or newspaper, movie, radio, or television program that is about 50 years old), what about it did you notice as being most different or unusual? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

Running Records Luckily, many organizations

maintain files or running records for their own purposes that you can use. For example, a country church has records of every marriage, baptism, and funeral from 1880 to the present. By looking at the church’s running records in combination with other historical evidence, you can trace the social life of a village for over a century. Two limitations of running records are as follows:

Primary Sources Are Just a Small Part of Everything from the Past What survived rarely is

a representative sample of all that occurred or existed. A concern with reliance on primary sources is that only a fraction of what had existed in the past has survived into the present. Thus, researchers are forced to use a small nonrepresentative sample of all people, events, and relationships as they try to create an understanding of life in the past. As you read primary sources, you want to avoid the distortion of a supercontext historical awareness. A good H-C researcher learns to “bracket,” or hold back, knowledge of subsequent events, current technology, and values when looking at the past. For example, you read a source written in 1840 by a Southern slave owner. Moralizing against the owner based on the evils of slavery or faulting him for not seeing that slavery would soon end is not a good research strategy. It is better to withhold judgment and try to see things from the perspective of the person who lived in the past. Researchers who study primary sources try to avoid a specific form of supercontext awareness distortion, presentism, which has a parallel distortion in comparative studies, ethnocentricism. In both types of distortion a person treats his/her own culture or time in history as being “normal” or “the best” and uses his/her own frame of reference as a standard to evaluate other times or places, rather than learning to view events from another’s point of view Locating primary documents can be very time- consuming. You must search through specialized indexes

1. Organizations do not always maintain them. 2. Organizations do not record information consistently over time. Changing policy or other events may cause an organization to stop keeping records, or new administrators, clerks, policies, or events may cause changes in the way records are kept and what they record. Before the use of electronic telecommunications, computers, cell phones, and video technology, people usually communicated and kept records in writing. You can examine letters, memos, diaries, ledgers, or newspapers to learn about past communication or ideas. Many of today’s communication forms do not leave a permanent physical record (e.g., telephone conversations, e-mails, radio broadcasts) unless they are electronically archived. This could make the work of future historians and H-C researchers more difficult. People’s memories of the past and their past experiences are similar to primary sources in that they were created in the past, but they are not a primary source. Recollections, such as a written memoir or autobiography, have been created afterward from memory. A special type

Figure 11.3 Internal and External Criticism Evaluating Primary Historical Sources External Criticism When Written? Primary Document

Internal Criticism Eyewitness or Secondhand Account? Where Was It Written?

Why Was It Written? Primary Document

Why Did It Survive?

Authentic? Who Was the Real Author?

Meaning in Context?

Literal Meaning? Internal Consistency?

Connotations?

248 Chapter 11 of recollection, the oral history, is especially valuable for people who do not keep detailed written diaries and for the illiterate. Because memory is imperfect, recollections and oral histories might be a distorted picture of the past in ways that primary sources are not. While they may not be perfectly accurate, often they are our only window into people’s lives and experiences from the past. The strength of primary sources is their authenticity, but they have practical limitations. It can take a huge amount of time to locate, verify, and read primary sources. You may discover thousands of pages of primary source documents to sort and read. In addition, primary sources may cover a short period of time or one very specific location. Perhaps they may let you document events in one tiny village for a two-year period, but many research questions are about a longer time and a wider range of locations. H-C researchers rely on secondary sources when they want to address broader questions.

after the Klan was active, Blee had to be persistent and inventive. She identified Klan women by piecing together a few surviving rosters, locating old newspaper obituaries that identified women as Klan members, and looking for names by scrutinizing historical documents that were public notices or anti-Klan documents. She mailed a notice about her research to every local newspaper, church bulletin, advertising supplement, historical society, and public library in Indiana. Most of her informants were over age 80. They recalled the Klan as an important part of their lives. Blee verified parts of their memories through newspaper and other documentary evidence. Membership in the Klan remained controversial. In the interviews, Blee did not reveal her opinions about the Klan. She was tested, but Blee remained neutral and did not denounce the Klan. She stated, “My own background in Indiana (where I lived from primary school through college) and white skin led informants to assume—lacking spoken evidence to the c ontrary—that I shared their worldview” (p. 5). She did not find Klan women to be brutal, ignorant, and full of hatred. When she asked why the women had joined the Klan, most women were puzzled by the question. To them it needed no explanation—it was just “a way of growing up” and “to get together and enjoy.”

Example Study Women of the Ku Klux Klan Blee (1991) was interested in right-wing extremist groups and conducted a historical study on the Ku Klux Klan in the Midwest. Her six years of research illustrates great ingenuity. She focused on the state of Indiana, where 32 percent of the white Protestant population belonged to the Klan at its peak in the 1920s. In addition to reviewing studies on the Klan, she investigated newspapers, pamphlets, and unpublished reports. She conducted library research on primary and secondary materials at over half a dozen college, government, and historical libraries. She provides readers with historical photographs, sketches, and maps to create a feel for the topic and its context.

Read More Blee also conducted oral histories with women about being in the Ku Klux Klan. She noted that, prior to her research, no one had studied the estimated 500,000 women in the largest racist, right-wing movement in the United States. Many people had assumed that women were apolitical and passive. The Klan was a secret society, no membership lists survived, and finding information was difficult. To locate survivors 60 years

WRITING PROMPT The Women of the Klan Study Recall the study by Blee on women of the Klan, focusing on the livesof white Indiana women in the 1920s. What types of data sources did she use in the study? How did she combine information from the multiple data sources to get information on women’s lives in the 1920s? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

Secondary Sources Secondary sources are the dozens of books and articles written by specialist historians on specific people, places, or events. They can be valuable sources information about broad historical eras or topics. Historians have produced numerous books and articles on many historical topics that provide a maze of particular details and interpretations. The H-C researcher often must transform the numerous separate studies, often on very narrow issues, into an intelligible picture organized around broad research questions. It is important to evaluate the sources and exercise care when using them. Secondary sources have limitations. Despite producing many studies, historians have not studied every aspect of all topics from the past. There are holes or gaps in the historical record, and you may find few studies on your research question. Other limitations include the issue of inaccurate historical accounts and the historian’s interpretation of data. Even if there are many studies on your

Looking at the Past and across Cultures 249

research question and you locate and read them all, history books have more than theory-free, objective “facts.” Historians frame and organize their primary source data using concepts, ideas, and assumptions. The historian’s concepts originate in journalism, the language of past people, ideologies or philosophy, today’s language, and theory. Unfortunately, historians do not always define their concepts or apply them consistently. For example, you read a book in which a historian calls 10 families in a nineteenth-century town “upper class.” However, the historian never defines “upper class.” This makes it difficult for you to know what “upper class” means in the history book. Few historical studies reveal everything the historian found in primary sources. The historian might have read 10,000 pages of newspapers, letters, and diaries and then organized and reduced those data into a few hundred selected quotes in a 180-page history book. If you read the 180-page book, you must rely on the historian’s judgments about the primary source data. The historian’s judgments about what to include contain assumptions, selection criteria, and biases. You rarely know what they were or whether the historian omitted information that is relevant for your research question. Narrative History When organizing their data, many historians use a narrative history format. This organizes the evidence chronologically around a single coherent “story.” There is a “flow” that connects each part of the story to other parts by its place a time order of events. Together, all the parts form a unity or whole and the writer presents the narrative to make interesting reading. However, the historians often emphasize or downplay information based on how well it fits into the story. You might read a history book (such as Blee looking for information about the Klan in Indiana in the 1920s), but perhaps the historian had downplayed information most relevant to your research question (women’s activism in the Klan) because it was not central to the historian’s story (how financial corruption and infighting weakened the Klan). Also, historians add events in the narrative to enrich the background or to add color. They might include what a particular person has done or said but may not analyze unseen influencing factors. Just because a historian put a fact in the history narrative, it does not mean that it has theoretical significance. For example, a historian discusses family relations among a group of settlers in seventeenth-century America. The historian focuses on intra-family gender dynamics but downplays other factors such as problems with securing a stable food supply. This does not mean that lack of food was unimportant for your research question (decisions by settlers to give up and return home). It only means that the historian decided to downplay such factors in his or her narrative. You can learn a lot from reading many historical narratives, but may need to look deeper for

details that are most relevant to your own research question even if they are not central to the narrative. The writing of historians is influenced by the era in which they live and by their outlook or school of historical scholarship. In certain historical periods, various interpretations or themes are popular. You will find similar themes in most historical books written in that period. Historians tend to follow a school of historical scholarship when they conduct a study, such as diplomatic, demographic, ecological, psychological, Marxist, intellectual, and so forth. Each school has its own priorities about data and key questions. If you use secondary sources, you need to be aware of major themes popular when the historian wrote and the outlook he or she adopted. Making It Practical: A Life History Interview

The life history interview is a special kind of oral history in which you interview a middle-aged or elderly person and ask about his or her entire life. You ask many open-ended questions in a sympathetic, nonjudgmental manner to help the person open up and elaborate on specific details of his or her life. You usually begin by asking about events when the person was a child. Be flexible, and do not structure the interview too much. You want to encourage the person to relax and reveal many specific details. You can prepare a long list of possible topics to ask about like family, schooling, work, travel, friendships, and so forth but also rely on other things you already learned earlier in the interview for new questions. For example, you ask: “Do you remember any details about the year you were six years old? Was that the year your younger brother Jonathan got very sick and almost died?”

Anything the person recalls is relevant. You want to capture details, relationships, and events from the point of view of the person you interview. Many people never had an opportunity to sit down and reflect about all the past events in detail. Life history interviews can be therapeutic for some people. Approach the interview with care and consideration. People may block out certain events, and recalling some events may generate great sorrow, anger, or anxiety, even many years later. You can ask about events taking place around the person such as major political events, world events, and so forth, to get the person’s own perspective on them. Most researchers audio-record life history interviews. Depending on the person, you may have six or more hours of interviewing. People tire of a very long interview, so plan to break the interviewing into three or more two-hour interview sessions that are spaced across several days. This not only lets the interviewed person rest but also lets you review the early interviews and return to major turning points or life events to ask for additional details or clarification.

250 Chapter 11 WRITING PROMPT Oral History An oral history is data based on talking to older people about their experiences and events when they were much younger. When you listened to an elderly person, what did you learn about “the old days” from the stories? Do you think the person’s memory was 100 percent correct? If not 100 percent correct, about how much correct did you think it was? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

11.4: How Do You ConductResearch That Compares Cultures? 11.4 Explain how historical-comparative research is significant in research related to the study of different cultures Comparative research is as more an orientation toward research than a distinct research technique. The techniques used when doing a comparative study are not dramatically different from conducting other social research; however, a comparative study greatly magnifies many methodological issues of social research. In a sense, a comparative perspective exposes potential flaws that can occur in all forms of research. This awareness of possible methodological concerns can help improve the quality of all social research. Some comparative research takes place to demonstrate that basic social processes like psychological orientations or family relations hold across different countries; other research examines sources of differences. A common H-C research strategy is to examine similarities and differences among units of comparison. In comparative research, you see what is shared across units and what is specific to one unit alone. The comparative orientation improves measurement and conceptualization. One reason is that the concepts developed from comparative research are less likely to be restricted to a single, specific culture. It is possible to develop a concept for a familiar or home culture but be unaware of hidden biases, assumptions, and values until we try to apply it to different cultures. For example, you develop a concept of reverse discrimination based on the U.S. experience. You think it is a general idea, but until you try to apply the concept in many other cultural settings, you do not know whether it is applicable or limited to the U.S. context. By using a concept in a range of cultures, you see whether it applies in diverse settings. Comparative research also reveals important but unrecognized factors. If all the studies of a topic are conducted with one culture,

it is possible to overlook something in the culture that is not present elsewhere, as a cause. For example, gun ownership is about ten times greater in the United States than in other advanced nations. If your study only looks within the United States, you may not find gun ownership as a key causal factor, but if you compare the United States with other advanced nations, gun ownership may stand out as being a key factor.

11.4.1: Looking across Cultures toSee a Wider Range It is important to look across cultures to see a wider range or variation in a variable. For example, two researchers HsiPing and Abdul look at the relationship between the age at which a child is weaned and the onset of emotional problems. Hsi-Ping looks only at U.S. data. They show a range from 5 to 15 months at weaning and indicate that emotional problems increase steadily as age of weaning increases. She concludes that late weaning causes emotional problems. Abdul looks at data from 10 cultures. He discovers a range from 5 to 36 months at weaning. He finds that the rate of emotional problems rises with age of weaning until 18 months; it then peaks and falls to a lower level. Abdul has a more complete picture. Emotional problems are likely for weaning between the ages of 14 and 18 months. Weaning either earlier or later reduces the chances of emotional problems. Hsi-Ping reached false conclusions because of the narrow range of weaning age in the United States. Comparative research is more difficult, costly, and time-consuming than research that is not comparative. You may be limited in the types of data that you can collect and have problems with equivalence (see the following discussion). In comparative research, you cannot randomly sample cultures. Sufficient information is not available for all world cultures and is unavailable for a nonrandom subset (poor countries, nondemocratic countries, etc.). In addition, cultures or nations are not equal units. Some have over a billion people and others only 100,000.

Example Study Abortion Politics in the United States and Germany Ferree et al. (2002) conducted a comparative study of public debate and abortion politics in the United States and Germany. Abortion was made legal in both countries and became a heated and widely debated moral-political issue. The authors examined four types of data: (1) secondary historical evidence, (2) a content analysis, (3) a survey of leaders in organizations, and (4) open-ended interviews. The secondary evidence came from books about the history, organized movements, court rulings, and public debates

Looking at the Past and across Cultures 251

about abortion. The authors examined articles in the two major newspapers of both countries for the period 1970–1994, looking for all articles that mentioned the abortion issue in any way. They selected articles meeting certain criteria (such as longer than three paragraphs, not book reviews, no letters to the editor) and found 1,243 articles in the United States and 1,425 in Germany. They then content analyzed the articles. The authors also sent a questionnaire to 150 abortion-issue organizations. The questionnaire asked about organizational goals, activities, media relations, internal resources, and alliances with other organizations. The authors also interviewed leaders or media directors of 20 U.S. and 23 German abortion-issue organizations and several leading journalists in each country who regularly wrote on the abortion issue. The study produced many findings. They found that people discussed the same issue very differently in the two countries. In Germany, there is less open conflict, and people discussed abortion in terms of fetal rights and protecting women. In the United States, there was great conflict created by the activism of advocacy organizations outside government or political parties. The issue was framed in terms of individual rights or freedom from government interference, or as a religious-moral issue. How religious and women’s organizations related to the same issue varied by country. In the end, the authors identified specific features of the culture, advocacy organizations, mass media, and political institutions in each country that influenced the abortion debate. They found that a country’s government structure, media system, legal system, and so forth influenced how people thought about and debated the same public issue.

11.4.2: Can You Really Compare? For convenience, most comparative researchers use the nation-state as their unit of analysis. Most people see the globe as divided in terms of nation-state (or country), and this is how official statistics are collected. The nation-state is a socially and politically defined unit. In it, one government has sovereignty (i.e., military control, political authority) over a populated territory. People living in the territory may share a common language, culture, and customs, but the nation-state is not the only possible unit for comparative research. If you want to compare something that is clearly inside and uniform within nation-state boundaries, such as a national legal system, voting and government, or economic

relations, then it makes sense to compare nation-states. However, for comparative research, the relevant unit is often the culture, not the nation state. Culture is more difficult to define. It refers to a shared identity, social relations, beliefs, and technology. Cultural differences in language, custom, traditions, and norms often align along nation-state borders, but national borders do not always match cultural boundaries. A single culture may be shared across several nations, or one nation-state may be multicultural hosting several cultures. If you want to compare cultures, then the nation-state may not be the best unit. For example, the people of one cultural region may have a distinct ethnic background, language, customs, religion, social institutions, and identity, such as Quebec in Canada, Wales in the United Kingdom, and Flanders in Belgium. Over the centuries, wars and conquests carved new political units onto territory and in the process destroyed, rearranged, or diffused boundaries between cultures. In many world regions, Western empires imposed arbitrary boundaries on distinct cultural groups and made them into colonies. Later the colonies became independent nations. At other times, one nation-state expanded and absorbed the territory of people who had a distinct culture. For example, the U.S. government took over American Indian lands and the islands of Hawaii and Puerto Rico. Likewise, new immigrants or ethnic minorities did not always assimilate into the dominant culture. Such intra-national cultures can create tension or regional conflict, since ethnic and cultural identities are the basis for nationalism. In his Presidential speech to the Canadian political science association, McRoberts (2001) argued that Canada is more than multicultural—it is multinational. Multiple nations (British, French, aboriginal) exist within one political structure. This implies that we should treat Canada as three units, not one, for some purposes. You need to ask: What is the appropriate comparative unit for my research question—the nation, the culture, a region, or a subculture?

For example, your research question is: Are income level and divorce related? Your hypothesis is that higherincome people are less likely to divorce. Perhaps income and divorce are related among the people of one culture. However, elsewhere in the same nation-state where a different culture prevails, income and divorce are not related. If you use the nation-state as the unit and mix everyone together, your findings will be unclear and show a weak correlation. If you instead use two separate regions, you will find a strong correlation among the variables in one culture and no correlation in the other. Thus, it is best to ask what unit is most appropriate, rather than assume that every nation-state has only one culture. Some cultures do not fit into one geographic region and are scattered. Cultural comparison is more complex. For example, the Latino

252 Chapter 11 or Chinese American or African American are ethno-religious cultures scattered geographically across the U.S. territory. Should we study them as distinct cultures or as ethnic/racial groups within a larger U.S. national culture? Which units should you use? Compare Your Thoughts How do you decide on the appropriate units of analysis for comparative research? If you want to study two different national cultures, such as the United States and Kenya, is the nation or a smaller unit such as the state, region, or tribe best? The answer hinges on whether a nation-state has a single hom*ogenous culture. This suggests caution if a country has regional divisions in which people of different regions do not share the same religion, customs, or language. You must learn about the internal divisions and minority groups within a nation before conducting a study. In some nations, such as Belgium, formal political and geographic divisions separate the country by distinct cultures. In such situations, subnational units may be best. In other countries, such as the United States, the cultural dividing lines are blurred and are not confined to distinct geographic regions. The appropriate units to study vary by your research question. For example, you are interested in marriage customs and childrearing practices. If there is one hom*ogenous culture, you can study it anywhere in a nation. If the nation has many different cultures and marriage practices might vary by culture, it might be best to limit your study to one or two cultural groups within the nation rather than use the nation as a unit.

Example Study Immigrants in Two European Countries

Immigration is a contentious issue as millions of people from less developed, low-income countries increasingly migrate to developed, high-income countries. Calavita (2005) studied immigra-

tion in Spain and Italy, two countries that only in the past 20years became magnets for a large influx of immigrants, especially from North Africa and Eastern Europe. She used qualitative data that she had collected for over a decade. Between 1992 and 2003, she spent a year in Spain and visited Spain and Italy numerous times for several weeks to several months duration. Fluent in Spanish and Italian, her data included interviews with government officials, immigrant association leaders, labor union leaders, academics, and employers. She also examined published and unpublished Italian and Spanish government documents, news reports, press conference records, and academic sources. As often occurs in qualitative research, her initial focus failed to conform to the data. As she said (p. 8), “I relinquished my preconceived notions . . . my empirical focus shifted to . . . far more difficult, and ultimately more interesting, questions . . .” She provided extensive documentation for the study, with 49 pages of end notes and a bibliography containing 480 sources. By comparing the two countries, she found many similarities as well as a few key differences.

Read More A major similarity was the “racialization” of immigrants. Ordinary Spanish and Italian people as well as officials categorized all immigrants as belonging to a racial, cultural “Other” that was fundamentally different from them. Although the specific laws and administrative procedures differed by country, both countries only allowed immigrants to do the types of jobs that few local people wanted. Paid very low wages, immigrants became essential to the operation of several sectors of the local economy. For example, in some areas of Italy, immigrants were over one-third of factory workers, domestic workers, and agricultural workers. On top of poverty-level wages, immigrants had a highly uncertain legal status. Regulations and laws nearly guaranteed that very few immigrants could ever gain citizenship. In fact, a “Catch 22” system of rules and regulations ensured that immigrants would not only be social outsiders but would also remain legal outsiders with few rights or protections. The few immigrants who could obtain legal status soon discovered that the administrative-legal system made keeping that status nearly impossible. For example, all work permits for immigrants were temporary, yet a record continuous work was required to maintain one’s legal status and continuous legal status was necessary to apply to become a permanent resident. As she quoted a Spanish elected official (Calavita, 2005, p. 41), “you can’t get residence if you don’t have a work permit and you can’t get a work permit unless you have residence.” Another example Calavita described how despite formal, official guarantees that all people would receive health care, particular local requirements, administrative discretion, and bureaucratic hurdles made actual access to health care nearly impossible for the immigrant population. In her conclusion, Calavita described the sharp contrast in both countries between an economy heavily dependent on immigrant labor and formal policies to integrate immigrants, and a daily reality for immigrants who occupy a racialized and marginalized social position, are restricted to jobs with exploitative working conditions and pay, and have a highly precarious legal status with restricted access to social services.

Looking at the Past and across Cultures 253

11.4.3: Galton’s Problem

Edward B. Tylor

Francis Galton

Sir Edward B. Tylor (1832–1917), founder of British social anthropology, presented a paper, “On a Method of Investigating the Development of Institutions, Applied to Laws of Marriage and Descent,” at the Royal Anthropological Institute in 1889. Tylor had information on marriage and descent for 350 cultures and found correlations between marriage/descent forms and measures of social complexity. He interpreted the results in terms of an evolutionary sequence. Over time, as societies became more complex, they changed from a maternal to a paternal descent line. Francis Galton, a statistical genius also at the Institute, raised objections. He pointed out that the correlation could be a mirage because the cultures may not be totally distinct and separate. Marriage forms could diffuse, cultures could borrow forms, or a marriage form could have a common origin in a different culture that crossed into other cultures. Galton maintained that before we can say that evolving social complexity and marriage forms are correlated, we must first rule out diffusion across different cultures and a common origin. Tylor agreed with Galton, and this sparked an innovation in comparative research called Galton’s problem (see Figure 11.4). To make a valid comparison, the units being compared must be distinct and nonoverlapping. If two units are actually subparts of a single larger unit, patterns or relationships you find in both of them may have a common origin. Here is an extreme example to illustrate the idea. Example Let us say you are interested in the association of two traits: the language people speak and what they use as money. Your research question is: Does the language people use affect what they use as money? Your units are territories in five nations (prefectures in Japan, the states in the United States, Lander in Germany, provinces in Argentina, and governorates [muhafazat] in Egypt). You add 47 prefectures, 50states, 16 Lander, 23 provinces, and 26 governorates for a total of 162 units. You discover a perfect correlation between language and money for your 162 units. Where

people speak English, they use the dollar; when they speak Japanese, they use the yen; when they speak Spanish, they use the peso; when they speak German, they use the euro; and when they speak Egyptian Arabic, they use the pound as currency. The association between language and currency is not because language and currency are actually related to one another; rather, your units of analysis (i.e., the states, provinces, etc.) are actually subparts of a larger unit (i.e., nations). The larger unit is the common source of both traits. Before you examine the association of two variables/ traits, look closely at your units. Galton’s problem is important because the boundaries between cultures are unclear or changing. It may be hard to say where one culture ends and another begins. Galton’s problem is a special case of the spurious relationship.

Figure 11.4 Galton’s Problem Francis Galton, Units and Diffusion Galton’s problem occurs when a researcher observes the same social relationship (represented by X) in different settings or societies (represented as A, B, and C) and falsely concludes that the social relationship arose independently in these different places. The researcher may believe he or she has discovered a relationship in three separate cases. But the actual reason for the occurrence of the social relation may be a shared or common origin that has diffused from one setting to others. This is a problem because the researcher who finds a relationship (e.g., a marriage pattern) in distinct settings or units of analysis (e.g., societies) may believe it arose independently in different units. This belief suggests that the relationship is a human universal. The researcher may be unaware that in fact it exists because people have shared the relationship across units. A

WRITING PROMPT Boundaries of Cultures Provide examples of where cultural boundaries and national boundaries are not the same, i.e., one culture crosses national boundaries or multiple cultures exist within one nation’s borders. What are some ways that you could mark or measure the boundaries of a culture that may not exactly match national borders? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

254 Chapter 11

11.4.4: Gathering Comparative Data Comparative researchers use several types of data and combine types together in one study: • Comparative field research • Existing qualitative data • Cross-national survey data • Existing cross-national quantitative data Comparative Field Research Comparative

researchers use field research in cultures other than their own. The training of social-cultural anthropologists prepares them for this type of research. The overlap of research techniques between anthropological and field research suggests modest differences between doing a study in your home culture and in a different culture. Conducting field research in another culture is usually more difficult and places more requirements on a study than doing research in your home culture. You must first immerse yourself in and learn the other culture in depth. This takes time and requires many adjustments, but it can yield new insights and provide personal as well as professional gains. Existing Qualitative Data Comparative research-

ers use qualitative primary and secondary sources. For example, you want to conduct a comparative study of the Canadian, Chilean, and Chinese school systems. Ideally, you visit each country to learn about each first-hand. Before visiting, locate existing qualitative data, such as videos, photos, music, novels, and essays about the country’s schools, their teachers, and students. Read historical and field-research studies that describe the education systems in the three nations. Descriptive details you learned from these sources can be part of the evidence for a study. Cross-National Survey Research You can con-

duct a survey in many different countries, but survey research in multiple cultures is more complex and creates additional methodological issues. In principle, the issues also exist in a survey conducted in your home culture, but they differ in magnitude and severity. A survey in a different culture requires knowing the language, norms, practices, and customs. Without such knowledge, you can easily make serious errors in procedure and interpretation. Knowing a language is not enough. A deep understanding of cultural assumptions, expectations, and practices is also essential. It is best to work in close cooperation with people who are natives of the other culture. Substantive (e.g., theory, research question) and practical factors are relevant when choosing cultures for a crossnational survey. You must adjust each step of the survey process (question wording, data collection, sampling, interviewing, etc.) to the specific culture. People in all cultures do not approach survey research the same. In some cultures, people might see survey interviewing as a strange,

frightening experience, somewhat analogous to a police interrogation. You want to check local conditions and learn how people in the culture view surveys. Cultural context also influences sampling. In addition to concerns about an accurate sampling frame, there are issues such as records privacy, quality of mail or telephone service, and transportation to remote rural areas. You need to know how often people move, the types of dwellings in which people live, the number of people in a dwelling, the telephone coverage, and typical rates of refusal (i.e., refusal to participate in the survey). The difficulties of writing a good survey question for your home culture are greatly magnified when you study a different culture. The cultural context affects question wording, questionnaire length, introductions, and the topics you include. You need to learn the local norms and topics that are highly sensitive in a culture. Questions about political beliefs, alcohol use, religion, or sexuality that you might ask in your home culture might be taboo in a different culture. In addition to these cultural issues, translation and language equivalency often pose problems (see the later discussion of lexicon equivalence). It is best to use bilingual people, but it may be impossible to ask the exact same question in a different language/culture. Cross-cultural interviewing can be difficult. The selection and training of interviewers depend on the education, norms, and etiquette of the culture. An interview situation introduces issues such as privacy norms, ways to gain trust, beliefs about confidentiality, and differences in dialect. In some cultures, you must spend a day or more in informal socializing before achieving the degree of trust and rapport needed for an interview. In other places, an official, headman, or local elder must grant approval or a male cannot interview a female without her elder brother, husband, or father being present. Survey research across nations/cultures can be a challenge, but it can be highly rewarding and produce information not otherwise attainable. More than anything else, you need to be culturally aware at each step. You must evaluate whether you can do the same as in your home or need to make adjustments. This careful, step-by-step evaluation in itself raises your awareness of how to conduct good survey research. Existing Cross-National Quantitative Data

Several types of organizations gather and publish data on numerous features for many nations. A wide variety of data are available in major national data archives in a computerreadable format. This makes it easy to conduct secondary analysis on international existing statistics data. As with existing statistics, including survey data for secondary analysis, for a single nation, existing cross-national data have limitations. The limitations for cross-national data are of greater magnitude than is the case for existing statistics on a single nation. The definitions of variables and the reliability of data

Looking at the Past and across Cultures 255

collection can vary dramatically across nations. Perhaps you want to study statistics on crime in various nations. However, nations vary by what constitutes a crime and by legal system. What is criminal behavior in one nation might be acceptable and legal in another. What one nation treats as a breach of etiquette another nation may see as a serious crime that requires harsh punishment. The standards of legal evidence and law enforcement may also vary widely. Missing information is a common problem in cross-national existing statistics. Intentional misinformation in the official data from some governments is also a problem. Another limitation is

the nations for which data are collected. For example, during a 35-year period, new nations come into existence, others change their names or form of governments, and some have different territorial borders. In the cross-national study of work schedule control by Lyness et al. (2012), the researchers found that self-reported social class data were only available for 19 of the 21 countries in the study, and that the social class categories were not the same in all countries. As a result, the researchers were forced to make adjustments, curtail some analysis, and be more cautious and limited in their generalizations about social class.

Example Study Cross-National Study on Work Schedules tion, greater workplace commitment, and less family-work conflict. A key aspect of flexibility is schedule control, i.e., determining starting and stopping times and quantity of work, neither more nor fewer hours than desired. Lyness and colleagues examined two kinds of micro-level factors, at the individual level and the workplace, as well as macro-level or country factors that influence work schedule control in 21 countries using the secondary data from the International Social Survey Program (ISSP). In addition to learning whether the characteristics that predict schedule control and its outcomes held up across all the countries, they wanted to see whether country-level factors mattered (see Figure 11.5). Their cross-national study found that high-income, high-education workers consistently reported having greater work schedule control. However, very high income and self-employed people reported working more hours than desired, while low-education, low-income workers reported working fewer hours than desired. An employee’s ability to control hours of work and work schedule varies within and across countries. Past studies in a handful of countries found that flexible work arrangements, defined as “the degree to which workers, rather than employers, control their own schedules and realize their preferences” (Lyness et al., 2012, p. 1025) predict higher job satisfac-

Read More Using a measure of self-reported social class, the researchers found that lower and working class people consistently reported less work schedule control than middle- and upper-class people. A strong gender difference appeared. Women, but not men, con-

Figure 11.5 Antecedents and Consequences of Worker Control over Working Time

Workplace Schedule Control in 21 Countries Source: Adapted from Figure 1 Lyness al., 2012, p. 1092

Job characteristics and individual worker attributes Workers’ control over scheduling and hours worked Economic conditions and social policies of a country

Amount of job satisfaction and family strain a worker experiences

256 Chapter 11 sistently reported having less work schedule control and working more hours than they desired. Three country-level factors were associated with having more work schedule control, a higher GDP, high levels of unionization, and strong national paid leave policies. In particular, countries with the most generous paid leave policies (e.g., Sweden with 25 paid leave days per year) compared to countries with no national paid leave policy (e.g., the United States) showed the smallest gender gap in people working more hours than they desired. In general, more work schedule control predicted positive job satisfaction and workplace commitment across all countries. It did not predict a reduction in family-work conflict

consistently across all workers. Greater work schedule control strongly predicted reduced family-work conflict for women workers, but not for men, across the countries in the study. This quantitative cross-national study revealed significant consistency across 21 countries regarding key features of work schedule control. It also identified a few national-level characteristics or policies that increased work schedule control, which in turn improved job satisfaction, workplace commitment, and particularly for women, reduced family-work conflict. The study both demonstrated consistency across countries for the primary relationship, and revealed the importance of specific country-level factors.

equivalence. They are closely related, but we can subdivide them into four types:

Summary Review

Data in H-C Research

• Lexicon

Thus far, we have considered four types of historical evidence and four types of comparative data.

• Context

Table 11.3 Data in H-C Research Four Types of Historical Evidence

Four Types of Comparative Data

Primary evidence

Comparative field

Recollections

Existing qualitative

Running records

Cross-national survey

Secondary evidence

Existing cross-national quantitative

11.4.5: The Issue of Equivalence You have probably heard the expression: Compare apples to apples, not apples to oranges. The expression is not about fruit; it is about comparison. In other words, you cannot examine the similarities among things that are completely different; what you compare must be equivalent in some way. Equivalence is a critical concern in all types of social research. It is the issue of whether you can compare across divergent social contexts. You cannot easily compare family relations, crime rates, and business patterns for people or societies that are radically different—for example, an advanced urban industrial nation of 300 million with a tiny hunting and gathering tribal society of 1,000 people. The technology, material conditions, and social environment differ radically. The units must have some basic similarity in order to make a valid comparison. The issue of equivalence exists in all social research, but it is crucial in H-C research. Without basic equivalence, you cannot apply ideas or measures across different cultures or historical periods. You will misinterpret events in a different era or culture, or there may be no equivalent feature for you to compare. There are several forms of

• Conceptual • Measurement Lexicon Equivalence Lexicon equivalence is easi-

est to see across two languages. For example, in many languages you use one form of address and pronouns for private settings, intimates (close friends and family members), and subordinates (younger persons, lower-status people) and another for address in public settings, with strangers, or for persons of higher social status. Modern English usage lacks this feature, although it once had something similar (the pronouns thou, thee, thine, and thy). In cultures where age is an important status, many statusbased words exist that are absent in English. You cannot say, “my brother” without indicating whether you are talking about an older or younger brother because “my younger brother” or “my older brother” are different words. The meaning of words also varies across historical eras. For example, today the English word “weed” refers to unwanted plants or to marijuana. In Shakespeare’s era, “weeds” were clothing. Linguists study whether what you can express in one language (or the same language long ago) has a precise meaning in another language. If your study goes into the distant past or involves different languages or dialects, lexicon equivalence can be an issue. Do not assume that a word or expression used in a different cultural or historical context has the same or a simple and comparable meaning. When is lexicon equivalence relevant in H-C research? Compare Your Thoughts • When you are reading a document from the distant past or written in a different language • When you are writing a survey questionnaire or talking to people who use or think in a different language

Looking at the Past and across Cultures 257

• When you are describing events, activities, or social relations from a different culture or historical era that you have studied to an audience today, in your home culture • When you are theorizing and analyzing data or conditions across different cultures or historical eras Conceptual Equivalence Every day you use many concepts—such as friendship, loyalty, trust, family, parent, employee, and self-esteem—to discuss and examine social life. In a study, you use concepts from the research literature, personal experiences, and background knowledge to analyze and discuss the data or to test and build theory. The concepts you use are from a specific culture and historical era. The people who discuss concepts in the research literature live in specific cultural and historical settings, but they try to stretch beyond one time and one culture. The issue of conceptual equivalence is whether a concept developed and used in one historical era and in one culture also applies to a different time or place. An idea (sibling conflict, customer service, blood honor) that may be best for understanding and organizing data in one historical era or culture may be inappropriate and incomprehensible in a very different era or culture. Conceptual equivalence is part of the broader issue of using universal (transcultural, transhistorical) concepts versus culturally and historically specific ones. Let us say you want to discuss “income.” In premoney bartering societies in which most people grow their own food, make their own furniture and clothing, barter goods, and rarely use money, income means something very different from today’s advanced societies. The same concept of “income” may not fit both settings. Likewise, if you study people who believe that spirit forces cause certain diseases or health conditions, you may not have a concept that easily and fully captures the concept of “spirit forces” for comparison with a society that uses concepts such as bacteria, germs, DNA, and so forth. The concept of “attending college” has a different meaning today than in a historical context in which only the richest 1 percent of the population attended college, most colleges had fewer than 300 students, all were private all-male institutions, and the college curriculum consisted of learning classical languages and receiving moral training. The event or idea “attending college” has a different meaning from today. At one time we automatically treated the beliefs and ideas of other cultures or eras as being inferior or as “superstitions” and treated those of own era and culture as “better,” “real,” or “true.” Nowadays, most of us try to avoid ethnocentrism and presentism, but this does not eliminate the issues when conducting research. Cultural anthropologists and historians create new concepts that can capture what they study in a different

culture or past era. To compare and discuss similarities and differences, being sensitive to how well a concept fits a setting is essential. You must ask whether a concept really applies to both settings. If it does not, you may have to use different ones for each part of the comparison. Contextual Equivalence Our conversations, actions, events, and activities always take place in a social context. The same action (singing, telling a joke) may be acceptable in one social setting but not another. The context helps give meaning to the action. It might be acceptable to sing loudly at a lively party, but it is not acceptable for a physician to sing while carrying out a risky medical procedure on a patient. How a context shapes an action, event, or activity may vary across cultures and historical times. Perhaps in the historical past or in a different culture, singing during a medical procedure was not only acceptable but expected behavior. You need to evaluate not just what someone does or says, but how it fits into a context, and the influence of a context can vary by historical era or culture. Contextual equivalence means you should recognize how a specific context can shape the meaning of an event or activity. The specific context is within a broader culture or historical era. For example, you study a culture in which everyone has to pay a small amount of cash before a government official will perform a task, such as to stamp an official document and approve a driver’s license. In your home culture, this may be considered a bribe and clearly inappropriate or illegal. Perhaps in one culture if you want a driver ’s license application approved, you must slip the clerk a five-dollar bill with your application. If you do not include the money, the clerk will put your application in a huge pile that might be processed next year. In the local context, the payment is acceptable as a kind of small service fee. The larger culture shapes the local context. In the larger culture, small amounts are paid regularly to help out the extremely underpaid government officials who work with impossible-to-implement bureaucratic rules. In a different context, paying a fee for a government service (having your passport stamped at the local airport) may be unacceptable. In a local context, government officials want to appear “modern” and impress foreign visitors that one set of rules applies, but when they provide a service to local people, another set of rules applies. To understand what is occurring, you need to be sensitive to the local context and the larger culture. How do you compare this government service to that in a different culture where handing a cash payment to a government employee would be seen as an unacceptable bribe? The key point is, you should not assume that the same activity will have the same meaning across different contexts. Measurement Equivalence You may identify a concept, event, or activity for your study and adjust for the

258 Chapter 11 context and the broader cultural or historical setting. A question remains: Can you measure in the same way if you want to compare it to a different cultural or historical setting? Measurement equivalence recognizes that you cannot always use the same measure (such as survey questionnaire or official records such as birth certificates) in different historical or cultural settings. Let us say you are studying two cultures. In one culture, the local police keep a careful record of each household updated annually, including each person in a dwelling unit, their kinship or relations to each another, and where each person works or attends school. You want to compare this to another culture, where there is no police record of who lives where, except for the name of the legal owner of a building. There is no other information about the places where people live. You cannot use the same measure of household size—looking at official police records in both countries. In one culture, you have to conduct a survey and ask about household size. In the other culture, you use official police records. The issue is whether the two different methods give identical answers if you had used them in the other culture (i.e., use a survey in the culture where police keep records). Perhaps you adjusted the way you measure a concept, event, or activity for a culture or historical setting so that it differs from how you measure it in a different setting. Did the measurement method affect what you learned in each setting, and can you treat the information as the same? There is no simple solution to this issue. You may have to measure differently in various settings and then compare results across settings, and you cannot be certain whether the different measurement methods influenced what you found.

11.5: How Can You Be an Ethical H-C Researcher? 11.5 List the responsibilities of a historical-comparative researcher in upholding ethical principles Ethical concerns in H-C research are similar to those of nonreactive research techniques, especially when you use cross-national secondary sources or existing quantitative data. Primary historical sources can introduce special ethical issues. The ethical researcher needs to carefully document where and how primary sources were obtained, make explicit the selection criteria for including some sources and not others, and apply external criticism and internal criticism to the source materials. At times, protecting privacy may interfere with gathering evidence. A person’s descendants may want to destroy or hide private papers or evidence of scandalous

behavior. We know that political figures (e.g., U.S. presidents) have tried to destroy or hide embarrassing official records. Oral history studies require many of the same kinds of ethical concerns as interviewing in survey and field research. Comparative researchers want to be sensitive to issues of cross-cultural interaction. Ethical behavior means respecting the moral values and beliefs of research participants. You should learn what a culture considers to be acceptable or offensive and show respect for the traditions, customs, and meanings of privacy in the host culture. Actions that may be acceptable in your home culture may be unacceptable in a different culture, and vice versa. Be aware of and check your assumptions about what is appropriate behavior. Perhaps in your home culture, you can enter a religious building and take photos when there is no service in progress. In another culture, taking a photo of religious buildings, even from the street, may be offensive and unacceptable. If this is the case, is it ethical to take the photo in the other culture? Perhaps you study a culture in which a father invites a male researcher to have sexual relations with his teenage daughter as a friendly gift to a visitor after completing an interview. In your home country, such behavior might be considered highly unethical and perhaps illegal. What is ethical behavior in such situations?

If you visit another culture to do research, establish good relations with the host country’s government and do not take data out of the country without giving something (e.g., results) in return. At times, the military or political interests of the researcher’s home culture or the researcher ’s personal values may conflict with official policy in the host nation. This introduces complexity. People in the other culture may distrust you and suspect you of being a spy, or your home culture’s government may pressure you to gather covert information. In the past, these issues have created serious difficulties for social researchers. At times, a researcher’s presence or findings create diplomatic problems among governments. For example, a researcher examines health care practices in a different culture and then declares that official government policy is ignoring a serious illness and purposefully failing to provide medical care to a particular part of the population. This may create a major controversy. Likewise, a researcher may be sympathetic to the cause of people who oppose the government. Sometimes even talking to and working with such a group might cause the foreign government to threaten the researcher with imprisonment or deportation. Anyone who conducts comparative research needs to be fully aware of such issues and the potential consequences of their actions.

Summary: What You Learned about Looking at the Past andacross Cultures You learned about historical-comparative (H-C) research. The H-C approach is especially appropriate when you want to ask “big questions” about macro-level change, or when you want to understand social processes operating across historical eras or across cultures. H-C research involves a different orientation toward research that goes beyond applying specialized techniques. There are several ways to do H-C research, but a distinctly qualitative H-C approach is similar to field research in many respects. H-C research has some specialized techniques, such as the external criticism of primary documents. Nevertheless, the most vital feature is how you approach a question, probe data, and move toward explanations. H-C research is often more difficult to conduct than other types of social research. The same complexities or difficulties are often present to a lesser degree than in other research. For example, issues of equivalence are present to some degree in all social research. However, in H-C research the issues are at the forefront of how you do research and seek answers to research questions.

Quick Review Historical-Comparative Research 1. H-C studies are ideal for addressing “big questions,” they often take a long time to complete, and most H-C studies address basic rather than applied research questions.

How Do You Conduct a HistoricalComparative Research Study 1. H-C research does not follow a rigid set of steps, but most studies proceed through six stages. 2. The first stage is to acquire a basic background knowledge. It involves orientation reading and may include taking courses, travel, or learning a language depending on the topic and setting of research. 3. The second stage is to conceptualize major ideas and begin to focus on a research question. This is similar to the flexible development of a focus in field research. 4. The third stage is to locate and evaluate the evidence. It may require visiting specialized research libraries, reading dozens of books and articles, spending time in another country, or learning a foreign language. 5. The fourth stage is to organize large quantities of diverse evidence. The great variety of evidence and complexity of historical evidence or data from other cultures requires significant skill to organize. 6. The fifth stage occurs after the evidence is in. It requires reading and re-reading notes, synthesizing, developing new concepts, highlighting similarities and differences, and looking for patterns from which you can develop into a coherent “story” or explanation to tell others. 7. The last stage is to write the report. Writing clarity is more often important for qualitative H-C than for quantitative studies. The success of an H-C study often depends on a well-written report that is able to distill mountains of evidence into clear exposition while documenting numerous sources.

2. Conducting or reading H-C research requires an awareness of geography, history, and cross-cultural differences, and H-C studies may be difficult to follow without sufficient background knowledge. H-C researchers often use a variety of data. Some rely on qualitative data, others on quantitative data.

How Do You Conduct Research onthe Past?

3. Qualitative H-C research has many similarities with field research. Both incorporate an individual researcher’s point of view in the research process, examine a great diversity of data types, focus on processes and time passage, use grounded theory, and make limited generalizations.

1. Compared to traditional historical research, the H-C researcher views obtaining original documents and offering detailed descriptions of specific past events as secondary; instead he or she sees historical evidence as the means for extending or building theory to understand social relations.

4. H-C research has several important differences from field research in the present in your home culture. An H-C researcher must work with limited evidence, learn to interpret evidence with minimum distortion, integrate the micro and macro levels, and use cultural-historical specific as well as transcultural, transhistorical concepts.

2. An H-C researcher may use primary sources (i.e., original documents or records from the past) as a type of evidence. In doing so, he or she works to avoid presentism, or viewing the past through the viewpoint of the present, and subjects primary sources to internal and external criticism to verify authenticity and accuracy.

259

260 Chapter 11 3. H-C researchers also use running records when studying the past. These are systematic records kept over a long period of time by an organization. 4. Recollections are used in some H-C research. These could be a written memoir or autobiography, created by a person from memory, or an oral history based on an interview in the present. 5. H-C researchers tend to rely on secondary sources, or studies by other people about the past, more than traditional historians. Secondary sources have several limitations. Historians have not studied every aspect of all topics from the past, their studies include the interpretations of a specific historian with selective data, and a particular historian’s ideas and assumptions may influence primary source material in such studies.

How Can You Be an Ethical H-CResearcher? 1. Comparative research is as more an orientation toward research than a distinct research technique; however, issues in a comparative study magnify methodological issues of all social research. 2. Comparative research can reveal factors that would go unrecognized if you were restricted to one culture. By looking across cultures, it is possible to see a wider range or variation in a variable. 3. The issue of equivalence exists in all social research, but it is especially crucial in H-C research. It means some fundamental degree of similarity among units/ideas must exist in order to make comparisons. The equivalence issue takes several forms, lexicon, conceptual, context, and measurement. 4. To make a valid comparison, the units must be distinct and nonoverlapping. If two units are actually subparts of a single larger unit, relationships you find in both of them may have a common origin. This is called Galton’s Problem.

5. Data in comparative research can be qualitative, and involve field research in a different country that brings greater complexity than same-country field research, or existing qualitative data that require acquiring a good cultural background about a country to avoid mistakes. 6. Comparative research can also be quantitative such as cross-national survey data or existing statistics data. There are a few special concerns or issues that arise when a study is comparative, especially being culturally sensitive, ensuring equivalence and problems of missing data.

Shared Writing: Context Matters In 1870, the earliest data on record, about 2 percent of U.S. teens were completing a high school education. No one had telephones, radios, or autos. No homes had electricity or electric lights, and just 10 percent of them had running water. A “very fast train” went 40 miles per hour, and 1870 was the first year of the transcontinental railway, dropping the time to travel from San Francisco to New York from three months to seven days. To put it in perspective, today, about 3 percent of the U.S. population has a doctorate degree (MD, JD, or PhD) and 99 percent of homes have electricity and running water, and you can fly from San Francisco to New York in about six hours. If you were conducting a study about family or workplace relationships in the 1870s in the United States, what three conditions, beyond the ones listed above, would you want to learn about in order to adjust to the historical context? Explain why those three conditions are important. Next look at two classmate’s responses who include a condition different from the three you came up with and evaluate which would be most important and why. A minimum number of characters is required to post and earn points. After posting, your response can be viewed by your class and instructor, and you can participate in the class discussion.

Post

0 characters | 140 minimum

Chapter 12

Writing a Research Report

Learning Objectives 12.1 Describe the importance of the research report 12.2 Apply principles of the writing process to

create a good research report 12.3 Describe cause–effect relations in research

studies

12.5 Describe how to develop qualitative

research reports and how they differ from quantitative data research reports 12.6 Apply strategies to prepare an effective

research proposal

12.4 List the steps for writing a quantitative

research report Research is not complete until you share results with other people. Communicating how you conducted a study and its results with others is a critical stage of the research process. It is usually in the form of a written report. Whether a study you completed was in chemistry, criminology, education, engineering, marketing, nursing, psychology, public policy, or another field, you need to communicate how

you conducted it and its findings to others. Your communication can be oral but should also be in a written form. As with all written communication, you want to maximize readability. Readability indicates how accessible the writing is, and how effectively it reaches a given reading audience.

261

262 Chapter 12 An audience’s ability to read and understand what you wrote will depend on their prior knowledge and reading skills. These vary widely. Medical doctors differ from teens. Whatever your audience, try to make your writing both highly accurate and easy to understand. Do this by using precise wording, a clear organizational structure, and a logical flow of ideas. Hartley, Pennebaker, and Fox (2003) examined the readability of research reports. They noted that past studies suggested that men and women wrote differently, and that individuals writing alone differ from people who write together. To see whether this happened in research reports, they looked at reports in the Journal of Educational Psychology for 1997–2001. They found 21 articles that individual men wrote, 21 by individual women, and 19 written by pairs of men and 19 by pairs of women. They examined each article’s abstract, and most of its Introduction and Discussion sections. They measured readability with two computer-based writing style programs. They did not find gender differences or group versus individual author differences. However, they did find differences among article sections that replicated findings from past studies. The Introduction sections were harder to read than Discussion sections. In addition, Introduction sections had more specific quotations from other authors than the Discussion sections.

12.1: Why Write a Report? 12.1 Describe the importance of the research report You can learn a lot about writing a research report by reading many reports or taking a course in scientific and technical writing. A research report is organized in a specific way to communicate the methods and findings in a straightforward and serious manner. More than a quick summary, it is a full, complete, and detailed record of the entire research process. Do not wait until you finish all research activities to begin thinking about the report.

Begin to think ahead while early in the research process. The report is one reason to keep careful records while you are conducting research. In addition to the findings, the report includes the reasons for initiating a study, a description of the research procedures, a presentation of evidence, and a discussion of how the evidence relates to the research question. It has a reference section, also called works cited or bibliography, of the literature cited in the report. The references provide background on the research question and show how your study fits within past research. For some studies, especially those with historicalcomparative (H-C) research, the reference section includes documentary data sources.

The purpose of a report is to tell others what you did, how you did it, and what you discovered. People write research reports for many reasons: to fulfill a course, academic degree, or job assignment; to meet an obligation to an organization that paid for the research; to persuade a professional group about specific aspects of a problem; or to tell the general public about findings. However, telling the public what you learned is usually a second stage of disseminating findings. The first stage is to communicate with people who are knowledgeable about the study topic and the research process. They can seriously evaluate the research in the report; they are also in the best position to understand what you did in the study and why. Assume that people who will read the report are scientifically literate. Hurd (1998) suggested that a scientifically literate person can do the following four things: 1. Distinguish experts from uninformed people, theory from dogma, data from myth, empirical evidence from propaganda, facts from fiction, and knowledge from opinion 2. Understand that the research process is cumulative, tentative, and skeptical; as well as understand the limitations of scientific inquiry and causal explanations 3. Recognize the need to gather sufficient evidence to support or reject claims and the influence of society on research 4. Analyze and process data, and be aware that problems often have more than one accepted answer and that many issues are multidisciplinary with political, judicial, ethical, and moral dimensions

12.2: How Do You Proceed with the Writing Process? 12.2 Apply principles of the writing process to create a good research report There are many good books on how to write a research report. My favorite is by Howard S. Becker and is titled Writing for Social Scientists: How to Start and Finish Your Thesis, Book, or Article (2007). It has many tips and tricks, as well as a general philosophy of writing, social relationships, and doing research. Many other books focus more on step-by-step mechanics. In this section, we discuss the general consensus on writing research reports.

12.2.1: Know Your Audience Professional writers tell us, “Always know for whom you are writing.” This is because communication is most effective when you tailor it to a specific audience. You should write a research report differently depending on whether

Writing a Research Report 263

the primary audience will be an instructor, students, professional social scientists, practitioners, or the lay public. In all cases, the writing should be clear, accurate, and organized. This usually takes hard work and practice. Making It Practical: Tailoring a Report to the Audience Here are strategies for adjusting writing

about a study to tailor it to five different audiences. • Writing for instructors. An instructor assigns a research report for different reasons and may place requirements on how you are to prepare it. In general, instructors want to see good writing and an organization that reflects clear, logical thinking. They want to see a report that demonstrates a solid grasp of substantive and methodological concepts. They want to see technical terms used explicitly when appropriate, and not used excessively or incorrectly. Often instructors are concerned more with how the report demonstrates your thought process, specific details about research steps, and use of the correct format than with your findings. • Writing for other students. You want to define all technical terms and clearly label each part of a report. Your discussion section should proceed in a very logical, step-by-step manner. You should offer many specific examples. You will want to use less formal language to explain how and why you conducted the various steps of the research project. One strategy is to begin with the research question, then structure the rest of the report as an answer to that question. • Writing for expert professionals and scholars. You do not need to define technical terms in easy-to-understand language, or explain why you used research standard procedures (e.g., random sampling). Professionals are interested in how the research links to theory or past findings in the literature. They want to see a compact but detailed description of the research process. They will pay close attention to how you measured variables and gathered data. They like a condensed, tightly written, but extensive section on data analysis, with a precise and meticulous discussion of the results. • Writing for practitioners, managers, and policy makers. Provide a short summary of how you conducted the study, and present the main results in a few simple charts and graphs that are easy to read and understand. Practitioners like to see an outline of alternative paths of action implied by results with the practical outcomes of pursuing each path. Practitioners focus on the main findings, but you may have to caution practitioners not to overgeneralize from the results of one study. For practitioner reports, you want to place the details of research design and the long version of the results in an appendix.

• Writing for the public. You need to use simple language, provide concrete examples, and focus on the practical implications of findings for specific issues or problems. You do not have to include many details of research design or of results. You need to be very careful not to make unsupported claims when writing for the public, because people can easily misinterpret findings. Informing the public is an important service. It helps nonspecialists make better judgments about public issues.

WRITING PROMPT Audience Matters Consider the various audiences that you typically write for and whether you write the same way to a say, a close friend of your age and social position as you would to a prospective employer. If you do not write the same to everyone, what are some of the things that you adjust? If you do write the same to all audiences, what types of readers might find it too difficult or too simplistic? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

12.2.2: Pick a Style and Tone Write reports of research studies in a narrow range of styles with a distinct tone. They have one primary purpose—to communicate clearly the research method and findings. Style Style refers to the type of words, length and form of sentences, and pattern of paragraphs. In the report, avoid a poetic or flowery style with colorful adjectives and illusions. Also avoid a highly formal, dense, and turgid style that is common in government and bureaucratic documents. Such documents use passive voice and are full of long, complex sentences and unnecessary technical jargon. Research reports have a formal and succinct style, i.e., they say a lot in few words. Tone Tone is the attitude or relation toward the subject

matter that you express as a writer. For example, an informal, conversational way of writing (colloquial words, idioms, clichés, and incomplete sentences) with a personal tone is appropriate if you are writing a text, letter or e-mail to a close friend. You may use comedy or a whimsical tone to express humor. This is inappropriate for research reports. The tone in a research report expresses some distance from the subject matter. It is professional, semidetached, and serious. Field researchers sometimes use an informal style and a more personal tone, but this is the exception. Also, avoid moralizing and “preaching” a specific point of view. Your foremost goal is to inform, not to advocate a position or to entertain. You can persuade by

264 Chapter 12 systematically presenting empirical evidence, careful documentation, and fully explaining how you used accepted research techniques. Your research report should be accurate and clear. You must check and recheck details like page references in citations, and fully disclose exactly how you conducted the study. If readers detect carelessness in your report, they may question the research itself. The details of a research study can be complex. This complexity makes confusion possible. It also makes clear thinking and plain writing essential. To achieve these goals, think and rethink your research question and study design. You should explicitly define major terms, write in short declarative sentences, and limit the conclusions to what you can support with empirical evidence. Anticipate possible misunderstandings and review them with care.

12.2.3: Organize Your Thoughts Writing is serious, time-consuming work. It does not happen magically or simply flow automatically when you put pen to paper or fingers to keyboard. Think of writing as a process. It has a sequence of steps and activities that result in a final product. Writing a research report is not radically different from other types of writing. The steps may differ and the level of complexity may be greater, but what a good writer does to write a long letter, a poem, a set of instructions, or a short story also applies to writing a research report.

Figure 12.1 Outline Form The Logic of an Outline I. First major topic

High level of importance

A. Subtopic of topic I

Second level of importance

1. Subtopic of A

Third level of importance

a. Subtopic of 1

Fourth level of importance

b. Subtopic of 1

”

(1) Subtopic of b

Fifth level of importance

(2) Subtopic of b

”

(a) Subtopic of (2)

Sixth level of importance

(b) Subtopic of (2)

”

i. Subtopic of (b)

Seventh level of importance

ii. Subtopic of (b)

”

2. Subtopic of A

Third level of importance

B. Subtopic of topic I

Second level of importance

II. Second major topic

High level of importance

1. Have something about which to write. The “something” in the research report includes the topic, research question, design and measures, data collection techniques, results, and implications.

3. Avoid procrastination and writer’s block. Some people become afflicted with a strange ailment when they sit down to compose writing—the mind goes blank, the fingers freeze, and panic sets in. Writer’s block also occurs when procrastination becomes writing constipation! After many delays, you cannot get the writing process moving. You stall without ideas or motivation. Many people delay writing until the last possible moment and then, under great pressure or panic, force themselves to write. This is not a successful strategy for professional writing. The pressure of a deadline might be motivation, but frantic, rapid writing is only useful as a prewriting activity (discussed later), not as a final draft. Writers from beginners through experts occasionally experience writer’s block. If you experience it, calm down and work on overcoming it. There are many tricks to overcome it (e.g., taking a walk, getting a back rub, organizing files, listening to music). A common technique is to identify small parts of a larger task that you can easily accomplish and do them first. Set a reasonable schedule for such parts. After you finish a part, give yourself a small reward (e.g., have a special dessert). The best way to avoid writer’s block is by writing, at least a little, all the time. Set aside a small amount of time to write every day. If you are always writing something, even if you later throw it away, you are less likely to be blocked.

2. Get organized. When you have many parts to write about, organization is essential. The most basic tool for organizing writing is the outline. Outlines help you ensure that all ideas are included and that the relationship among them is clear. Outlines have topics (words or phrases) or sentences. Most of us are familiar with the basic form of an outline (see Tips for the Wise Researcher: Outlining and Figure 12.1).

4. Engage in prewriting activities. Writing is an ongoing process, not a one-time event. Sometimes it helps to build up momentum. Writing starts with prewriting activities—creating self-imposed deadlines and a schedule, arranging your notes, preparing lists of ideas, drafting rough outlines, making certain that bibliographic references are complete, and reviewing the data analysis. Of course, there is a danger that

Getting Started in Writing

Strategies for breaking the inertia and starting the momentum of writing are the same for writing a research report as for most other medium to long writing projects.

Writing a Research Report 265

you will get distracted and sidetracked by such activities. Self-discipline is central to the writing process.

ture after completing data collection and analysis. There are three reasons for doing this:

Tips for the Wise Researcher

1. Time has passed between the beginning and the end of a research project, and new studies may have been published.

Outlining Outlines can help you as a writer. However, they can also become a barrier if you use them improperly. An outline is a tool that helps you organize ideas. With it you do three things: 1. Put ideas in a sequence 2. Group related ideas together 3. Separate the more general, or higher-level, ideas from more specific ideas, and the specific ideas from very specific details Some students feel they must have a complete outline before they can start to write, and that once they prepare an outline, they cannot deviate from it. Few professional writers begin with a complete, detailed outline and stick to it rigidly. Your initial outline can be sketchy, because until you write everything down, it is impossible to put all ideas in a sequence, group them together, and separate the general from the specific. For most writers, new ideas develop or become clearer in the process of writing itself. Your beginning outline may differ from the final outline by more than its degree of completeness. The process of writing not only reveals and clarifies ideas, it also stimulates new ideas, new connections among ideas, a new sequence for the parts, or new relations between the general and the specific. In addition, the writing process may stimulate you to reanalyze or reexamine the literature or findings. This does not mean beginning all over again. Rather, it means you need to maintain an open mind to new insights and be candid about reporting how you conducted the research.

WRITING PROMPT Writer’s Block You must start a large writing project but you are experiencing serious “writer’s block.” Many people break the writing of a large paper or report into small parts that they later assemble. What are the possible pros and cons of this approach? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

12.2.4: Go Back to the Library Few researchers totally finish the literature review before they complete the study. You should be familiar with the literature before you start, but expect to return to the litera-

2. After completing a research project, you will know better what is or is not central to the study and may have new questions in mind as you reread studies in the literature. 3. As you write the report, you may find that your notes are not complete enough or a detail is missing in the citation of a reference source. The final visit to the library after data collection is less extensive and more selective or focused than that conducted at the beginning of research, but it is often necessary to fill in missing details. When writing a research report, you will probably discard some of the notes and sources that you gathered. This does not mean that your initial library work and literature review were a waste of time and effort. Researchers anticipate that some of the notes (25–30 percent) taken before completing the study will become irrelevant as the study gains greater focus. Do not include notes or references in a report that are no longer relevant. They will distract from the flow of ideas and reduce the report’s clarity. You return to the library to verify and to complete references. Going back to the library also helps you avoid plagiarism, a very serious form of cheating. Take careful notes and identify the exact source of phrases or ideas to avoid unintentional plagiarism. Cite the sources of what you directly quote and of ideas that you paraphrase. For direct quotes, always include the specific location of the quote with page numbers in the citation. Using another person’s words and failing to give credit is always clearly wrong. Researchers regularly paraphrase and cite the source. To paraphrase, you need a solid understanding of the core ideas you wish to cite and give credit to the source.

12.2.5: Engage in Prewriting Activities Many people find that getting started is difficult. Beginning writers often jump to the second step and go from there. This often results in poor-quality writing. Prewriting will help you spend thinking time to consider the form of the report and audience. Thinking time often occurs in spurts over time before the bulk of composing begins. Many writers begin to compose by freewriting. As you freewrite, do not stop to reread what you wrote, do not ponder the best word, do not worry about correct grammar, spelling, or punctuation—just get your ideas down as quickly as possible. You can later clean up what you wrote. Writing itself ignites the thinking process, which, in turn, feeds further writing.

266 Chapter 12 WRITING PROMPT Freewriting Have you found “freewriting” to be useful? Why or why not? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

12.2.6: Rewrite Your Report Perhaps one in a million writers is a creative genius who can produce a first draft that communicates with astounding accuracy and clarity. For everyone else, writing means that rewriting—and rewriting again—is necessary. For example, Ernest Hemingway is reported to have rewritten the end of A Farewell to Arms 39 times. Professional researchers may rewrite a report a dozen times. If rewriting sounds daunting to you, do not become discouraged. If anything, rewriting reduces the pressure; it means you can start writing and just produce a rough draft that you will polish later. Plan to rewrite a draft at least three or four times. A draft is a complete report, from beginning to end, not a few rough notes or an outline. Rewriting helps you to express yourself with a greater clarity, smoothness, precision, and economy of words. The focus is on clear communication. Do not use unnecessary pompous or complicated language. When rewriting, you slowly read through what you have written. Some people read what they wrote out loud to see whether it sounds right. It is always wise to share your writing with others. Professional writers have others read and criticize their writing. New writers quickly learn that friendly, constructive criticism is very valuable. Sharing your writing with others may be difficult at first. It means exposing your written thoughts and encouraging criticism. Yet the purpose of the criticism is to clarify the writing. A good critic is doing you a favor. Rewriting involves two processes: revising and editing. Revising gives you the opportunity for inserting, deleting or inserting new ideas, adding supporting evidence, clarify meaning, and strengthening transitions and links between ideas. Editing is cleaning and tightening up the mechanical aspects of writing. As you rewrite, go over a draft and revise it brutally to improve it. You will find this is easier to do if you allow some time to pass between writing a draft and rewriting. Phrases that seemed satisfactory in a draft may look fuzzy or poorly connected a week later. Even if you have not acquired great typing skills, it is a good idea to use a word processor to create at least one draft before the final draft. It is easier to see errors and organization problems in a clean, typed draft. Serious pro-

fessionals find that the time they invest in building keyboard skills and learning to use a word processor pays huge dividends later. Word processors make editing much easier. They also check spelling, offer synonyms, and check grammar. Do not rely on the computer program to do all the work; the computer can miss usage errors like if you have used their when you should have used there.

One last suggestion: Write the introduction and title after you finish a complete draft of the report. This ensures that they will accurately reflect what you have said. Make titles short and descriptive, so they communicate the topic and the major variables. Occasionally they describe the type of research but should not have unnecessary words or phrases. Making It Practical: Rewriting Good writing requires multiple drafts, and as you revise and rewrite each draft you should be attentive to several areas where you can improve the smoothness and clarity of communication.

1. Mechanics. Check grammar, spelling, punctuation, verb agreement, verb tense, and verb/subject separation with each rewrite. Remember that each time you add new text, new errors can creep in. Many mistakes are distracting, and they weaken the confidence readers place in the ideas you express. 2. Usage. Reexamine terms, especially key terms. When you rewrite, check to see whether you are using a word that best expresses your intended meaning. Do not use technical terms or long words unnecessarily. Use the plain word that best expresses a clear meaning. Get a thesaurus and use it. A thesaurus is an essential reference tool, like a dictionary. It has words of similar meaning and can help you locate the exact word for a meaning you want to express. Precise thinking and expression requires precise language. Do not say average if you intend to use mean. Do not say mankind or policeman when you intend people or police officer (i.e., strive for gender neutrality in your writing) (see National Council of Teachers of English, 2002; Weber, 2012). Do not use principal for principle, there for their, or than for then. Also avoid unnecessary qualifying language, such as seems to or appears to. 3. Voice. Some writers make the mistake of using the passive voice in a research report instead of the active voice. It may appear to be authoritative, but the passive voice obscures the actor or subject of action. Compare the voice in these two examples: PASSIVE: “The relationship between grade in school and more definite career plans was confirmed by the data.” ACTIVE: “The data confirmed the relationship between grade in school and more definite career plans.” PASSIVE:

Writing a Research Report 267

“Respondent attitude toward abortion was recorded by an interviewer.” ACTIVE: “An interviewer recorded respondent attitude toward abortion.” 4. Coherence. Make the sequence of ideas logically tight. Include transitions between ideas. Try reading the entire report one paragraph at a time. Does each paragraph contain a unified idea? Does it have a topic sentence? Did you include transitions between the topics and paragraphs within the report? 5. Repetition. Remove repeated ideas, wordiness, and unnecessary phrases. It is best to state ideas once, forcefully, rather than repeatedly in an unclear way. When revising, eliminate deadwood (words that add nothing) and circumlocution (the use of several words when one precise word will do). Directness is preferable to wordiness. Here is an example: WORDY: “To summarize the above, it is our conclusion in light of the data that x has a positive effect of considerable magnitude on the occurrence of y, notwithstanding the fact that y occurs only on rare occasions.” LESS WORDY: “In sum, the effect of x on y is large and positive but occurs infrequently.” 6. Structure. Make the organization of your research report transparent. Move sections around as necessary to fit the organization. It is wise to use headings and subheadings. A reader should be able to follow a report’s logical structure with great ease. 7. Abstraction. A good research report mixes abstract ideas with concrete examples. A long string of abstractions without specifics is difficult to read. Likewise, a mass of specific concrete details without a periodic generalization to summarize the main point can also lose readers. 8. Metaphors. Many writers use metaphors to express ideas. They use phrases such as “the cutting edge,” “the bottom line,” and “penetrating to the heart” to express their ideas. Metaphors can be an effective method of communication if you use them sparingly and with care. A few well-chosen, consistently used, fresh metaphors can communicate ideas quickly and effectively; however, excessively using metaphors, especially overused metaphors (e.g., the bottom line), is a sloppy, unimaginative method of expression.

WRITING PROMPT The Process of Revising When you write a report or paper that is five or more pages long, how many drafts do you write? When you revise a draft, what two or three things do you change most often during revision? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

Summary Review

Steps in the Writing Process The only way to learn to write is by writing. Writing takes time and effort and improves with practice. There is no single correct way to write, but some ways are better than others are. The writing process has three steps: 1. Prewriting. Prepare to write by arranging notes on the literature, making lists of ideas, outlining, completing bibliographic citations, and organizing comments on data analysis. 2. Composing. Get your ideas onto paper as a first draft by freewriting, drawing up the bibliography and footnotes, preparing data for presentation, and writing a draft introduction and conclusion. 3. Rewriting. Evaluate and polish the report by improving coherence, proofreading for mechanical errors, checking citations, and reviewing voice and usage.

12.3: How Do You Write about Cause-Effect Relations? 12.3 Describe cause-effect relations in research studies In many research studies, you want to demonstrate causeeffect relations by examining an independent variable (cause) and dependent variable (effect) in a simple hypothesis. Although most common in quantitative data research, causal relations appear in all types of studies. Experimental researchers can demonstrate cause-effect relations most clearly. Experiments best meet the three conditions required to show causality: • temporal order • association • control over alternative causes Hard-core experimenters tend to think that only an experiment can show cause-effect relations, and they question it in other types of research. You show cause-effect in an experiment by introducing the independent variable at a specific time, looking for an association in data on dependent variable measures, and creating an experimental design to control for factors that could influence the outcome (internal validity). Of course, the experimental approach has limitations, including a lack of random samples, difficulties with generalizing (external validity), an ability to examine only one or two variables at a time, and a micro-level focus. If you conduct an experiment, you can

268 Chapter 12 talk about causal relations in a report by showing a strong association between the independent and dependent variable in results and a high degree of internal validity in the experimental design. Talking about Cause-Effect Relations in Nonexperimental Studies Researchers who use sur-

vey or existing statistical methods also want to show causeeffect relations. Of the three conditions for causality, they can best demonstrate an association among variables. However, an association or correlation alone is not enough for causality. Demonstrating time order in survey or existing statistics data requires extra effort. For example, a respondent may answer all survey questions at one point in time. If some questions ask about past events (e.g., what did you do in high school, or how much education did you parents receive?) while others ask about present events or attitudes, you can establish time order by carefully looking at the time of a variable in a survey question or variable. This is how a survey or existing statistics researcher argues for temporal order. Survey and existing statistics researchers also control for alternative explanations. They cannot exercise physical control through experimental design as experimenters do. Instead, they measure any variables that might indicate an alternative explanation. These are control variables. Advanced statistical methods allow researchers to add control variables along with the main causal variables and check whether the main variables are still association, net of the control variables. If you conduct survey or existing statistics research, you can talk about causal relations by showing a strong association among variables, by noting the temporal ordering of variables, and by ruling out alternative explanations by using control variables. By using control variables and logically arguing that alternatives are unlikely, you can say that an association among variables supports a causal relationship and is not spurious. Many field researchers and H-C researchers also want to show cause-effect relations. They observe events over time, so meeting the temporal order condition is usually not a barrier. Showing an association is more difficult. To do this, they carefully note the co-occurrence of events in the qualitative data. For example, a researcher studies a small work setting and during a year of observation notices that employees become upset and complain about work conditions and superiors for a week on several occasions. The researcher notes when this occurred and what occurred before, during, and after the employee dissatisfaction peaks. Each peak of complaints occurred shortly after employees returned from a holiday or vacation. During weeks before a holiday or vacation, employees appeared happy and satisfied, but during the week or two following they complained. Listening to the informal talk among employees, the researcher learned that employees often spent vacation or holiday time with adult friends, family

members, and neighbors who worked in different types of jobs. In post-holiday informal conversations, the employees often talked about how their friends, family members, or neighbors had comfortable and rewarding jobs. The researcher has observed an association between: 1. expressions of dissatisfaction 2. the opportunity the employees had to discuss and compare their work situations with those of other people. Eliminating alternative explanations is the most difficult condition to satisfy. Qualitative researchers do this by becoming intimately familiar with a setting and gathering in-depth details about a particular context. They think about alternatives while conducting the research and try to look for evidence of possible alternative explanations. For example, is it having time away from work on a vacation or holiday and not comparing jobs with friends, family, and neighbors that increased dissatisfaction? If you conduct a field or H-C study, you can talk finding a causal relationship in a report if you show a sequence of events and can document events that occurred earlier or at the same time. To eliminate alternative explanations, be familiar with the literature on the topic and note whether anyone suggested or found evidence for alternative explanations. You can also suggest potential alternatives along with any evidence you found for them with your judgment of whether they might be present.

12.4: How Do You Write aQuantitative Research Report? 12.4 List the steps for writing a quantitative researchreport The principles of good writing apply to all types of reports, but the parts of a report differ depending on whether your study used quantitative or qualitative data. It is always wise to read many reports on the same kind of research for models and ideas. We begin by looking at the quantitative research report. The sections of the report roughly follow the sequence of steps of a research project.

12.4.1: Abstract or Executive Summary Quantitative research reports begin with an abstract. The size of an abstract varies; it can be as few as 50 words or as long as a full page. Most scholarly journal articles have abstracts on the first page of the article. The abstract has information on the topic, the research problem, the basic findings, and any unusual research design or data collection features.

Writing a Research Report 269

Applied research reports for practitioners substitute a long summary, the executive summary, for an abstract. The executive summary has more detail than an article abstract and is longer (often three to five pages). It has most major findings, the implications of findings, and major recommendations. The executive summary has more in it because many practitioners and policy makers only read the executive summary and then just skim the full report. Abstracts and executive summaries tell a reader what is in a report. They help the reader looking for specific information to screen many reports and decide whether to read the entire report. An abstract also gives serious readers who intend to read the full report a quick overview of it. This makes reading the report easier and faster. Although the abstract or executive summary is the first thing someone reads, you should write them last. Write them after you have finished the rest of report so that you know what the report contains.

WRITING PROMPT Article Abstracts Consider your strategy for reading a research report/article on quantitative research. Describe where you start, what you read next, what you read last or skip, etc. How does your reading strategy influence how you would write such a report? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

12.4.2: Presentation of the Problem The first section of a report is very important. In it you introduce readers to the main topic, research question, and overall tone of the report. Your goal is to present the research question and its significance, define major concepts, and begin to guide a reader through your report. The introductory section has several possible headings, such as “Introduction,” “Problem Definition,” “Literature Review,” “Hypotheses,” or “Background Assumptions.” Although the headings vary, there is always a statement of the research problem/question and a rationale for its importance. Explain the significance of your study by showing how different solutions to the problem lead to different applications or theoretical conclusions. This section is where you place the literature review and link the specific question addressed in your study to what others found in past studies. You also define key concepts and present the main hypotheses in conceptual terms.

12.4.3: Description of the Method The next section of the report describes how you designed the study and collected the data. It goes by several names

(e.g., “Methods,” “Research Design,” or “Data”) and may be subdivided into other parts (e.g., “Measures,” “Sampling,” or “Manipulations”). Professionals consider it to be the most important section for evaluating the quality of the study’s methodology. This section answers several questions for the reader: 1. What type of study (e.g., experiment, survey) did you conduct? 2. Exactly how did you collect the data (e.g., study design, type of survey, time and location of data collection, experimental design used)? 3. How did you measure/introduce each variable? Are the measures reliable and valid? 4. Did you sample? How many participants or units are there in the study? Exactly how did you select them? What are their characteristics (age, gender, race-ethnicity, geographic distribution and so forth)? 5. Did any ethical issues or specific design concerns arise, and if so, how did you address them?

12.4.4: Results and Tables or Charts After describing how you sampled, collected data, and measured variables, it is time to present the data. The goal is to present and reveal—not discuss in depth or fully interpret—the data. Researchers often combine the “Results” section with the next section, called “Discussion” or “Findings,” but many keep the two sections separate. You have a few choices in how to present the data. When you analyzed the data, you looked at dozens of univariate, bivariate, and multivariate tables and statistics to learn about the data. Do not put every statistic and table you looked at in a final report. Rather, select the minimum number of charts or tables that fully informs a reader. Charts or tables should summarize the data (e.g., means, standard deviations), show tests of hypotheses, and present statistical models and other statistics. Your goal is to give readers a complete picture of the data without overwhelming them—avoid excessive detail and irrelevant data. Readers can make their own interpretations by examining the data. Detailed summary statistics and very technical explanations belong in appendixes.

12.4.5: Discussion Section You separate the discussion section from the data results section to allow a reader to examine the data and arrive at his or her own interpretations. The discussion section gives the reader a concise, unambiguous interpretation of the data’s meaning. It is not a selective emphasis or partisan interpretation; rather, it is a candid and neutral discussion of data in the results section. Organize your discussion to make it easy for a reader to follow.

270 Chapter 12 Many beginning researchers struggle with a discussion section. Reading many research reports can help you see how to do it. One approach is to keep your research question in mind and organize the discussion according to hypotheses. You can discuss how the data relate to each hypothesis and describe how the findings specifically support, modify, or reject each. After reading the discussion section, a reader should have a clear picture of the major study findings. In addition to reporting how data relate to the hypotheses, discuss unanticipated findings that look interesting. Also, provide alternative explanations of results, including ones that challenge initial hypotheses. In addition, note the limitations of the study. Beginning researchers may find it unusual to point out the shortcomings or limitations of their study. You want to be fully open and candid about the study and data, never hide or defend a particular outcome. Readers will have more confidence in the report if you openly disclosed its limitations. Remember the goal in research is to seek truth without restraint, not to defend a particular position or pet idea.

12.4.6: Conclusion or Summary In the conclusion or summary section, you restate the research question and summarize the most important findings. This is a place to highlight major limitations of the study and possible implications or directions for future research. The only sections after the conclusion are notes, references, and appendixes (if any). Only include sources you referred to in the references section. If you have an appendix, place additional detailed information on methods of data collection (e.g., questionnaire wording) or specialized results in it. Use footnotes or endnotes sparingly to expand or elaborate on information in the text. Put secondary information (information that clarifies your in-text statements) in notes to avoid distracting readers from the flow of your text. Making It Practical: Examining a Quantitative Data Study Report The reader learns the fol-

lowing five things from the abstract of the quantitative data study by Weitzer and Tuch (2005): 1. The topic of this study is racial bias by police, specifically racial profiling in the United States. 2. The study is important because the extent of police racial bias and public perceptions of it are not well known. 3. Data for this study are a national survey on citizens’ views of and experiences with police bias. 4. Findings of the study are that three factors shape a person’s views on the issue: race, prior experiences with police discrimination, and exposure to news media. 5. The study finds support for a group-position theory of race relations.

The authors divided the introduction section into a short introductory paragraph that outlines the topic and research question, a section with the literature review, and the hypotheses they test. The researchers present three hypotheses; each has two parts. The authors also outline alternative explanations or theories that they will consider. The next section, Methods and Data, is divided into several parts: Sampling, Panel Representation (repeated measures of the same people over time), Independent, Dependent, and Control Variables. In the methods section, the authors also present readers with basic descriptive statistics for each variable in a table. The Results section combines results and discussion. There are many tables with numbers in this section. The authors organized the discussion of results by each hypothesis. The Conclusion section is three pages long. In it, the authors discuss the significance of their study, repeat their major findings, and point to directions for future research. After the conclusion are endnotes and references.

WRITING PROMPT Dissecting a Quantitative Journal Article When you encounter a quantitative scholarly journal article, what do you do first? For example, do you always read it from the beginning to the end, or do you jump around within the article? Have you looked at the bibliography or reference section to locate related sources on the same topic of an article? Describe your reading process and what aspects of this type of scholarly article cause you the most difficulties. The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

12.5: How Do You Write a Qualitative Research Report? 12.5 Describe how to develop qualitative research reports and how they differ from quantitative data research reports Compared to a quantitative data research report, most people find writing a report on qualitative research to be more difficult. It has fewer rules and is less structured. Nevertheless, all reports have the same purpose: to communicate how you conducted a study, what data you collected, and what you learned from a very complete examination of all the data. In a quantitative research report, you present hypotheses and evidence in a logically tight and condensed style. By contrast, qualitative study reports tend to be longer and less

Writing a Research Report 271

compact. In fact, book-length reports are common. Reports on qualitative research are longer for several reasons: • It is more difficult to condense data in the form of words, maps, pictures, or sentences than in the form of numbers. • Providing qualitative evidence often means offering readers specific quotes and extended examples with many details. • In a report on qualitative research you want to create a subjective sense of empathy and understanding of real people, events, and settings, and give highly detailed descriptions of settings and situations. • In qualitative data studies, you use less standardized research techniques for gathering data, creating analytic categories, and organizing evidence. You may create new the techniques for a particular setting. This means you must explain techniques in depth rather than refer to a well-known standard technique. • Many qualitative data studies construct new concepts or theories. It takes more words to develop new concepts and explain the relationships among them than to rely on existing concepts. Because theory flows from the evidence, you need to show readers how the concept is linked to the evidence. • Writers use a variety of writing styles in a qualitative data report. This freedom to depart from a precise, standard style tends to increase overall length.

12.5.1: Report on Field Research Reports on field research rarely follow a fixed format with standard sections, and often have a less objective and formal tone. You can write a report on field research in first person because you, as an individual person, were directly involved in the setting and interacted face to face with the people studied. Your decisions or indecisions, feelings, reactions, and personal experiences are legitimate parts of the field research data and process. In many qualitative data reports, researchers do not separate theoretical ideas and data into distinct sections; rather, they intertwine generalizations with empirical evidence. Evidence takes the form of detailed description with frequent quotes. You want to balance the presentation of data and their interpretation or analysis. Too much data without analysis and too much analysis with supporting data are both to be avoided. Data reduction is a major issue in field research. Field research data are in the form of a huge volume of field notes, transcripts, maps, photos, and documents. You can only share a very small percent of them with the readers. Field researchers organize reports in several ways. Two major ways are by chronological natural history and by themes. When you use natural history organization, you present the sequence of ideas and data in the same time order as you discovered or came across them. When you use a thematic organi-

zation, you present a theme and then provide data to illustrate or support it. You can choose from abstract analytic themes in the scholarly literature, ones you developed, or those used by the people you studied. Each has its advantages. The first makes it easier to connect your study with other studies or more abstract theory. The latter two sources of themes allow you to offer a theme that is closely tailored to a particular setting or to display the language, concepts, categories, and beliefs of the people studied. You can use a mix of types of themes. In a field research report, you discuss the methods used in the report, but its location and form vary. One technique is to interweave a description of the setting, the means of gaining access, the role of the researcher, and the subject–researcher relationship into the discussion of evidence and analysis. A chronological or theme-based organization allows you to put the data collection method near the beginning or the end. In book-length reports, most authors put a discussion of methodological issues in a separate appendix at the end. Field research reports can contain transcriptions, maps, photographs, and charts illustrating analytic categories. They supplement the discussion and are placed near the discussion they complement. The photographs give a visual inventory of the settings described in the text and help a reader see physical arrangements or conditions, as well as the people being studied. Direct, personal involvement in the intimate details of a social setting heightens ethical concerns. Researchers write in a manner that protects the privacy of people they study. They usually change the names of members and exact locations in field reports. Personal involvement in field research leads researchers to include a short autobiography. For example, in the appendix to Street Corner Society, William Foote Whyte (1955) gave a detailed account of the occupations of his father and grandfather, his hobbies and interests, the jobs he held, how he ended up going to graduate school, and how his research was affected by his getting married.

Learning from History Boys in White

272 Chapter 12 The book Boys in White (Becker et al., 1961) describes the pioneering study by Howard Becker and his colleagues into how medical students become doctors. It is a classic ethnographic study that details the lives of young men at the University of Kansas—their schedules, efforts to find out what professors wanted from them, “latent culture,” and assimilation of medical values. It also describes how they learned to negotiate a hospital or clinic in all its complexity as well as develop a perspective on their futures. The clarity and thoughtfulness of the method used and its write-up are equally important as the findings. The authors started with about 5,000 pages of single-spaced field notes. Of these, they put less than 5 percent in the book-length report as quotations. The remaining 95 percent were not wasted. The field notes are a rich, deep reservoir of empirical evidence from which the authors created the report. As you prepare a report, carefully select quotes and indirectly convey the rest of the data to readers.

WRITING PROMPT Field Research Studies Which organization for field research studies are you most comfortable with? Do you prefer reading (and writing) a natural history chronology, with themes periodically introduced, or a series of clearly explained themes with many examples to document the themes? Explain your answer. The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit Organizing Your Report Despite wide variation,

the following is a common way to organize the field research: 1. Offer an introduction to the main issues and setting(s) a. Discuss the most general aspects of the issue/situation with a reference to past studies b. Provide the main contours of the issue/situation and field setting c. Describe how you collected data/materials d. Provide details about the specific setting and the events/people in it e. Discuss how you organized the report 2. Describe the specific situation and place it in a broader context a. Introduce major concepts or analytic categories b. Contrast the specific situation/setting of this study with other situations/settings c. Describe the development of issues/situations over time and any tensions/conflicts 3. Discuss strategies various people in the field setting used 4. Provide a summary and implications

Making It Practical: Examining a Qualitative Data Study Report The reader learns the following

five things from the abstract of the qualitative data study by Rodriquez (2011): 1. The topic is how nursing home care workers use emotions at the workplace. 2. The study is based on data from 18 months of participant observation in two nursing homes. 3. Besides participant observation, the researcher interviewed 65 staff members. 4. A central finding of the study is that care workers use emotions both to create a sense of dignity in their work and to obtain compliance from residents. 5. The study has implications for how emotions operate in a workplace; not only do organizations generate and impose controls on how workers deploy emotions but the care workers also produce and use emotions for their own purposes. Rodriquez opens the report with a discussion of the nursing home industry and its trends, and also discusses issues of workplace dignity and emotion work. In a second section, he describes “care work,” especially care work that is paid. He says that paid care work “transcends a simple market-exchange.” He suggests it is beyond a purely economic transaction as with much paid labor, because taking care of another person requires using emotions. In nursing homes, care workers develop long-term relationships with the people for whom they provide intimate daily care. The author points out the apparent contraction between a nursing home industry that is increasingly rational, marketoriented, and bureaucratic and daily nonmarket, family-like emotional relationships that characterize performing care work in nursing homes. He also says that care workers face a “wage penalty.” Because the overwhelmingly female care workforce reaps emotional rewards from their work, they receive few financial rewards for their labor. The care workers operate within a bureaucratic organization that imposes many rules and regulations upon them. Thus, in the first two sections of this report, Rodriquez has introduced to the readers the major issues and concerns that he investigated in the study. In the third section, Rodriquez describes the field setting of two nursing homes in great detail. He also tells us when he gained access and how he conducted the field observation, including his arrival and departure times and the frequency of observation visits. He describes taking jotted notes and full field notes, and the techniques he used for analyzing the data. Lastly, he describes who he interviewed and how. In the next several sections, Rodriquez provides readers with data analysis interspersed with frequent quotes and excerpts from his field notes or interviews. He describes particular people and events at the field site as a way to

Writing a Research Report 273

document the central themes of the setting. In the final Discussion Section, he synthesizes and discusses the major themes or findings at a general, abstract level.

WRITING PROMPT A Close Reading of a Field Research Report Evaluate the Rodriquez article answering the following questions: Do you think the article gives too little or too much personal detail? What about the report on the specific setting or his observations of the people in it did you think was unnecessary? What did you want to see but was absent? Did you find his explanation of organizing themes or findings clear and compelling? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

12.5.2: Report on HistoricalComparative Research There is no single best way to write a report on H-C research. Most frequently, researchers “tell a story” and they provide many carefully documented details. They go beyond description alone to include some limited generalizations and “package” descriptive details into analytic categorizes. In these ways, a researcher connects abstract concepts with specific empirical details. H-C researchers rarely describe their methods in any detail in a report. The methods may have involved visiting many specialized libraries, obtaining dozens of obscure documents, and reading mountains of background material. However, beyond careful citations in footnotes and listing sources, other details of method are not necessary. You will rarely see an explicit section of a report or an appendix describing the methods used. Occasionally, a book-length report contains a bibliographic essay that discusses major sources used. More often, evidence resides in numerous footnotes or endnotes. While the typical 20-page report on quantitative or field research will contain 5 to 15 notes or sources, a report on H-C research of equal length may have over 60 notes or sources. H-C reports often include photographs, maps, diagrams, charts, or tables of existing statistics. The researcher places them throughout the report in sections that discuss related evidence and ideas. The charts, maps, and so forth are less the “make or break” evidence as occurs in a quantitative research report than one part of the gradual build-up of a large quantity of diverse types of evidence. The charts, maps, or photos provide the reader with a fuller picture of the places and people in the study. They help to convey a “feel” for the depth and richness of complex evidence. In conjunction with other evidence, the researcher builds a

web of meaning with many descriptive details. The organizational structure of presenting large amounts of qualitative evidence itself becomes means by which the researcher conveys interpretations and analysis to the reader. You can organize a report of H-C research in two major ways: by topic and chronologically. Most writers mix the two types. For example, you can organize information chronologically within topics, or organize by topic within chronological periods. If the report is truly comparative, you have additional options, such as making comparisons within topics. Some H-C researchers use quantitative research techniques, alone or with qualitative data, and they may mimic the form of a quantitative research report to describe findings. Many H-C researchers use a narrative style of report writing, in which they “tell a story.” If you use a narrative style, organize the data chronologically and try to “tell a story” around the actions, specific individuals, and flow of events over time.

12.6: How Do You Prepare a Research Proposal? 12.6 Apply strategies to prepare an effective research proposal You might need to write a research proposal for a supervised project that you submit to instructors as part of an educational degree or to request funding for a study you wish to conduct. You write the proposal to convince reviewers that the study is a worthy one, and that you, the researcher, are capable of successfully conducting the study. Reviewers will have confidence that you can successfully complete the study if the proposal is well-written, clearly organized, and demonstrates careful planning. A proposal is similar to a research report, but you write it before the research begins rather than after you completed a study. It describes the research problem, discusses its significance, and provides a detailed account of the methods that you will use and why they are most appropriate. See Appendix A for example proposals.

12.6.1: Proposals for Research Grants A research grant provides the resources needed to help complete a worthy research project. Researchers whose primary goal is to obtain funding for personal benefit or prestige, to escape from other activities, or to build an “empire” tend to be less successful. The strategies for writing excellent proposals and “winning” grants are related to, but distinct from, the skills required to complete a study. Colleges, private foundations, and government agencies have programs that award grant funds to enable people to complete research study (see Figure 12.2).

274 Chapter 12

Figure 12.2 Examples of Announcements to Fund Research What does an RFP look like? Spencer Foundation Eligibility: Principal Investigators applying for a Research Grant must be affiliated with a school district, a college or university, a research facility, or a cultural institution. The Foundation accepts proposals from institutions and/or researchers from the U.S. and internationally. Researchers must also have an earned doctorate in an academic discipline or professional field or appropriate experience in an education-related profession. Budget Restrictions: Indirect costs may not be charged to proposed budgets with less than $50,000 in direct costs. Proposals exceeding $500,000 in direct costs require particularly close scrutiny and are generally developed in close consultation with Spencer staff prior to submitting a proposal. Determine whether your project fits within one or more of the Foundation’s current areas of interest: • • • •

The Relation between Education and Social Opportunity; Organizational Learning in Schools, School Systems, and Higher Education Institutions; Teaching, Learning, and Instructional Resources; and Purposes and Values of Education.

In addition to proposals in these defined areas, the Foundation will continue to accept proposals that do not fit one of these areas. Deadlines: There are no deadlines for preliminary proposals. They are welcome at any time. If invited, full proposals in the Major Research Grants Program are considered four times per year. Neuman Family Foundation The foundation is currently considering grants in four major areas: • A lcoholabuse prevention — to provide support for alcohol abuse prevention programs and to provide services for victims who have been injured by drunk drivers. • K-12 global education — to foster increased awareness and encouragement for students to study less commonly taught foreign languages, travel abroad, develop cross-cultural awareness so they develop into global citizens. • Alzheimer research — to assist in research to determine cause of Alzheimer’s disease (and related dementia diseases) and develop effective therapies. • A ssistance to homeless people — to help with basic needs and support to improve opportunity for independence, health and self-esteem of individuals and families who are homeless. Proposals will be considered in the above areas or for special project areas established by the Board. Proposals will only be accepted for programs within the state of Iowa. Evaluation will be on the basis of the procedure outlined herein (Application Procedure). Multiple year proposals may be considered but ongoing annual support is not provided. The foundation’s desire to provide a match of between 50 to 80 percent of total funds for a project. There is no fixed minimum or maximum amounts but most grants are expected to be small ($5000-$15,000) reflecting the foundation’s limited budget.

You may use the funds to purchase equipment, pay salaries, hire helpers, purchase supplies, travel to collect data, or to help with the publication of results. Specific grant sources vary widely in what they allow for uses of funds. The degree of competition to obtain a grant also varies a great deal depending on the source. Some sources fund more than 3 out of 4 proposals they receive; others fund fewer than 1 in 20 proposals. Proposals for Quantitative versus Qualitative Data Studies Proposals differ depending on whether

they are primarily quantitative or qualitative data studies. A proposal for quantitative research has most of the parts of a research report: a title, an abstract, a problem statement, a literature review, a methods or design section, and references. It lacks results, discussion, and conclusion sections. It is a plan for data collection and analysis (e.g., types of statistics) with a schedule of the steps to be undertaken and an estimate of the time required for each step. A proposal for qualitative research is more difficult to write. This is because the research process itself is less structured and preplanned. You prepare a problem statement, literature review, and references. You can demonstrate an ability to complete a proposed qualitative project with a wellwritten document that has an extensive discussion of the literature and the significance of the research question, and a list of sources. This shows reviewers your familiarity with

qualitative research and the appropriateness of the method for studying the problem. Your proposal will be stronger if it describes a qualitative pilot study you conducted. This demonstrates motivation, familiarity with research techniques, and ability to complete a report about unstructured research. You need to investigate possible funding sources so that you can direct the proposal to the funding source that offers the best chance of success. As you investigate sources, ask five questions of each: 1. What types of projects are funded—applied versus basic research, specific issues/topics or only certain research techniques? 2. What are the deadlines and proposal format requirements (page length, font size)? 3. What kind of proposal is necessary (e.g., short letter, detailed plan, many forms)? Is the proposal accepted in an open competition, restricted to certain types of applicants, or only by an invitation from the source? 4. What has been the range of size and duration of previous grants? 5. What aspects (e.g., equipment, personnel, travel) are or are not funded by the source? There are many sources of information on funding sources. Librarians or officials responsible for research grants at a college are good resource people.

Writing a Research Report 275

Example RFP

Figure 12.3 Example of Request for Proposal Many funding agencies periodically issue requests for proposals (RFPs). These ask for proposals to conduct research on a specific issue.

National Science Foundation W here D iscoveries B egin Division of Social and Economic Sciences

law & social sciences (lss) LSS Program Clarifications

Special Announcements

• Please attend to the Law & Social Sciences Postdoctoral Fellows announcement in the solicitation. Please note that the fellowships are project-based, and that they require at least two PIs from different disciplines. PIs are to have a project upon which a postdoctoral fellow will work. • Please note the description of conference and workshop awards. Law and Social Sciences will make awards that promote interactions among scholars from multiple disciplines and that will include younger scholars and increase the participation of members of underrepresented groups, in keeping with NSF policy.

The National Science Foundation and the National Institute of Justice are pleased to announce signing of a Memorandum of Understanding that outlines a framework for cooperation and collaboration in the social, behavioral, and forensic sciences. For more information concerning the announcement of this partnership, see its Dear Colleague Letter.

CONTACTS Name Helena Silverstein - Pgm Director Jon B. Gould - Pgm Director Allison Smith - Program Specialist Fatima J. Touma - Science Assistant

Email [emailprotected] [emailprotected] [emailprotected] [emailprotected]

Phone Room (703) 292-7023 995 N (703) 292-7808 995 N (703) 292-7094 995 N (703) 292-7320 995 N

PROGRAM GUIDELINES Solicitation 15-514 Important Information for Proposers A revised version of the NSF Proposal & Award Policies & Procedures Guide (PAPPG) (NSF 16-1), is effective for proposals submitted, or due, on or after January 25, 2016. Please be advised that, depending on the specified due date, the guidelines contained in NSF 16-1 may apply to proposals submitted in response to this funding opportunity. DUE DATES Full Proposal Target Date: August 1, 2016 Standard and Collaborative Research and Interdisciplinary Postdoctoral Fellowships August 1, Annually Thereafter Full Proposal Target Date: January 17, 2017 Dissertation Research, Standard and Collaborative Research and Interdisciplinary Postdoctoral Fellowships January 15, Annually Thereafter CAREER Proposals CAREER proposals must conform to the annually announced NSF-wide CAREER proposal deadline. EAGER, RAPID, and Conference Proposals EAGER, RAPID, small workshop, and small conference proposals may be submitted at any time, with prior permission of the Program Officer. Proposals for Large Workshops and Conferences should be submitted in cycle with standard research proposals. REU Supplement Proposals REU supplement proposals may be submitted at any time, with prior permission of the Program Officer.

SYNOPSIS The Law & Social Sciences Program considers proposals that address social scientific studies of law and law-like systems of rules. The Program is inherently interdisciplinary and multimethodological. Successful proposals describe research that advances scientific theory and understanding of the connections between law or legal processes and human behavior. Social scientific studies of law often approach law as dynamic, made in multiple arenas, with the participation of multiple actors. Fields of study include many disciplines, and often address problems including though not limited to: 1. Crime, Violence and Punishment 2. Economic Issues

3. Governance 4. Legal Decision Making 5. Legal Mobilization and Conceptions of Justice 6. Litigation and the Legal Profession LSS provides the following modes of support: 1. Standard Research Grants and Grants for Collaborative Research 2. Doctoral Dissertation Research Improvement Grants 3. Interdisciplinary Postdoctoral Fellowships 4. Workshop and Conference Awards LSS also participates in a number of specialized funding opportunities through NSF’s crosscutting and crossdirectorate activities, including, for example: • Faculty Early Career Development (CAREER) Program • Research Experiences for Undergraduates (REU) • Research at Undergraduate Institutions (RUI) • Grants for Rapid Response Research (RAPID) • Early-concept Grants for Exploratory Research (EAGER) For information about these and other programs, please visit the Crosscutting and NSF-wide Active Funding Opportunities homepage.

EDUCATIONAL OPPORTUNITY This program provides educational opportunities for Undergraduate Students, Graduate Students, Postdoctoral Fellows. Individuals interested in applying for funding should see the program guidelines above.

REVISIONS AND UPDATES What Has Been Funded (Recent Awards Made Through This Program, with Abstracts) Map of Recent Awards Made Through This Program News

The National Science Foundation, 4201 Wilson Boulevard, Arlington, Virginia 22230, USA Tel: (703) 292-5111, FIRS: (800) 877-8339 | TDD: (800) 281-8749

276 Chapter 12 Neat and Professional Proposals The instructions for proposals may ask for a detailed plan for the use of time, services, and personnel. State them clearly and be realistic. Excessively high or low estimates, unnecessary add-ons, and omitted essentials will result in a less favorable evaluation. Creating a budget for a proposed research project can be complicated and requires technical assistance. There are official pay rates, fringe benefit rates, and so on. It is best to consult a grants officer at a college or an experienced proposal writer. In addition, endorsem*nts or clearances of regulations are often necessary (e.g., IRB approval). The proposal should also include specific plans for disseminating results (e.g., publications, presentations before professional groups) and a plan to evaluate how the project has met its objectives. The proposal is a type of contract between the researcher and the funding source. Funding agencies always require a final report. It includes details on how funds were spent, the findings, and an evaluation of whether the project met its objectives. A failure to spend funds properly, complete the project as described in the proposal, or file a final report has serious consequences. The funding agency may sue for a recovery of funds, or a researcher might be banned from receiving future funding. A serious misuse of funds may result fines, jail terms, or the penalties for the institution (school, hospital, research institute) where the researcher was employed. The process of reviewing proposals after they are submitted takes from a few weeks to almost a year. In most cases, reviewers will rank a large group of proposals, and only the top ranked proposals receive funding. Most government agencies or research centers use a peer review process. Private foundations may have a mix of nonresearcher lay people and professional researchers review proposals. Instructions on preparing a proposal indicate whether you are to write for professional specialists or for an educated general audience. As with writing a research report, it is always best to have others read it and give comments. You can then revise as necessary and improve it before submitting your proposal.

If your proposal is funded, celebrate, but only for a short time. If your proposal is rejected, do not despair. Given the competition, funding sources reject a majority of proposals the first or second time they are submitted. Many funding sources provide you with written reviewer evaluations of the proposal. You should always request evaluations if they are available. Sometimes a courteous talk on the telephone with a person at the funding source will reveal the reasons for rejection. You should strengthen and resubmit a proposal based on the reviewer comments. Most funding sources accept repeated resubmissions of revised proposals. Proposals that have been revised based on past reviews tend to be stronger in subsequent competitions. Making It Practical: A Successful Research Grant Proposal Before you submit a grant pro-

posal, be certain that the funding agency is looking for proposals on your topic or research question. If you are uncertain, check with the funding agency officials first. If you submit to an inappropriate source or past the deadline, it will be returned or automatically denied. Once you submit a proposal to an appropriate funding source, reviewers are more likely to rate it higher when the following are present: 1. You submitted the proposal well within the deadline. Funding agencies automatically reject late proposals and rarely accept proposals that are even one hour late. 2. You have followed all instructions in detail. For highly competitive grants, the failure to follow minor technical details (e.g., use correct font size, adhere to page length) may result in automatic denial. 3. The proposal is written clearly and has easy-to-follow objectives. Many proposals provide charts or diagrams that make it easy for a reader to follow all details. 4. For basic research: Your proposal addresses a significant research question, clearly builds on past studies, and represents a substantial advance in knowledge. For applied research: Your proposal carefully documents a major social issue, shows an awareness of several alternatives, offers possible solutions, and has a solid evaluation plan. 5. You completely described research procedures and used high standards of research methodology. The research techniques are the most appropriate ones for the specific research question. 6. You provide a realistic budget that only includes allowable items, and appears “cost-effective.” 7. You include specific plans for how you will disseminate the results and evaluate all project objectives. 8. You provided a study design that shows serious planning and realistic budgets and schedules.

Writing a Research Report 277

9. You give evidence that you have the experience or background required to complete the study successfully and/or include other researchers with the necessary background and experience as consultants. 10. You include letters of support from knowledgeable people or involved organizations.

WRITING PROMPT Research Proposals and Funding When seeking grant funds, you might need to rephrase or refocus your research proposal to fit the priorities of a funding agency. What strategies might allow you to go beyond simply repeating a funding agency’s priorities and possibility get funding for an important study? The response entered here will appear in the performance dashboard and can be viewed by your instructor.

Submit

Undergraduate Research

Many people engage in social research—high school students, undergraduates, graduate students working on master’s or doctoral degrees, and professionals in many fields. Learning to conduct research has long been an essential part of earning an advanced degree. The sooner a person starts to conduct or assists others with conducting research the better. Learning from an experienced mentor is very valuable because a much of skill and knowledge base needed to conduct top quality research can only be learned experientially. Learning research skills often begins while a student is an undergraduate. In the United States, the Council on Undergraduate Research (CUR), founded in 1978, promotes research by all undergraduates. The CUR has grown to include 900 colleges and universities. The CUR hosts an undergraduate research conference, produces a scholarly journal, and provides special institutes on undergraduate research for students and faculty. In addition to CUR, other national organizations and many colleges and universities actively promote undergraduate research as a highly engaging and effective form of learning and professional growth. They offer small grant programs or summer research opportunities to assist students in conducting research. Select undergraduate research sources beyond those of one university: Council on Undergraduate Research link to http:// www.cur.org National Conference on Undergraduate Research link to http://www.ncur.org National Science Foundation: Research Experiences for Undergraduates link to https://www.nsf.gov/funding/pgm_summ.jsp?pims_id=5517 Undergraduate Research Community for the Human Sciences EuroScholars link to http://www.kon.org/urc/ undergrad_research.html

Summary: What You Learned about Writing a Research Report You learned about the research report and the process of writing such a report. You saw how features of a research study, such as qualitative versus quantitative data, affect the organization and content of a report. You also learned about the process of preparing a research proposal and seeking funding for research, as well as the fast-growing area of programs and support for undergraduate research. Clearly communicating results is a vital part of the research process, as are the ethics and politics of social research. I urge you, as a consumer of social research or a new social researcher, to be self-aware. Be aware of the

place of research in society and of the societal context in which social research can thrive. Social researchers bring a unique perspective to the larger society.

Quick Review The Writing Process 1. A research report is organized in a particular way that differs from other types writing. It communicates the methods and findings in a straightforward and serious manner and a full, complete, and detailed record of the entire research process.

278 Chapter 12 2. The report should be written with a specific audience in mind for better communication; nonetheless, it should have a straightforward style and professional, semidetached tone. 3. Often the writer of a research report returns to the library to update the literature review, verify that the report reflects current research, and check references and quotes. 4. Some writers have a temporary writer’s block and need ways to avoid the block as well as pre-writing strategies such as freewriting. 5. Many writers use an outline to organize material into a report. While an outline helps to sequence ideas, group related material together, and separate general from specific material, the outline should be flexible and change during the writing process. 6. To improve clarity and accuracy, writers of research reports typically revise and edit the report multiple times, and they try to avoid using the passive voice.

How Do You Write about CauseEffect Relations? 1. Causal relations can appear in all types of studies, and to show them you need to think about how you will write about and explain it to readers. How you do this depends on the type of research. 2. Demonstrating a causal relationship is easiest with experiments. You need to clearly explain the experimental design and be aware of internal and external validity issues. 3. For most nonexperimental, quantitative studies, you will need to clearly specify independent, dependent, and control variables as you outline the main causal hypothesis of the study. 4. Many but not all field researchers and H-C researchers also want to show cause-effect relations. They observe events over time, and must carefully note the co-occurrence of events in the qualitative data to show association.

How Do You Write a Quantitative Research Report? 1. Nearly all quantitative research reports begin with an abstract that introduces the topic and most significant findings. For applied research, an executive summary, that is longer, is used instead. 2. The first section of the report introduces the topic, the research question, and links the current study to past research, often with a literature review. It is where you tell readers the significance and reason for conducting the study. 3. In the next section you outline, in great detail, the study design, method of collecting data, and measurement of each variable. If you used sampling, this is where you explain it to readers. 4. The Results section, sometimes combined with the Discussion section, is where you present the data in charts,

tables, and statistics. You want to show the minimum number of charts and tables to give readers a full picture without overwhelming them. 5. The Discussion section follows and is where you review and discuss findings, or what the data show. One strategy for this section is to relate the findings to each hypothesis. 6. The conclusion or summary section rarely presents new information. Its purpose is to review everything that has been said previously in a more compact way, while highlighting the most important findings and their implications. Directions for future research suggested by the study are placed in this section.

How Do You Write a Qualitative Research Report? 1. Compared to reports on quantitative research, reports on qualitative research tend to be more difficult to write and have more varied formats. They are also longer and less compact. 2. Field researchers often use less formal tone in their reports and write in the first person. They may interweave analysis and the presentation of descriptive data, rather than keeping them separate. 3. Field researchers share a very small fraction of their data in a report, but often include long quotes or experts directly from field observation notes or interviews. 4. H-C reports often mix analytic categories with documented evidence to “tell a story.” The evidence can include quoted passages, photos, or charts, and H-C reports often have a much larger number of footnotes and sources than those for other research. 5. In both field and H-C research reports, writers frequently use a chronological or theme-based organization, or mix the two forms of organization (e.g., repeated themes within chronology).

Shared Writing: Reports on Quantitative and Qualitative Research Which do you prefer—reading a qualitative field study or a quantitative study? Specifically, what is it about the type you prefer that makes you more comfortable with it? What do you find difficult or makes you uneasy about the type of study you did not choose? Select the responses of a classmate who makes the opposite choices between study types that you did and suggest ways to approach reading the study that would be more worthwhile. A minimum number of characters is required to post and earn points. After posting, your response can be viewed by your class and instructor, and you can participate in the class discussion.

Post

0 characters | 140 minimum

Appendix A

Sample Research Proposals Example Research Proposal, Qualitative Exploratory Study of Anime Fans in theUnited States

Introduction and ResearchTopic Anime, or Japanese-origin animation, has become widely popular among some teens and young adults. Enthusiasts watch many hours of the films, collect the films, read magazines about characters and films, attend fan conventions, create Internet sites with fan information, and dress up as their favorite characters. While the public might be familiar with a few box-office hits, such as Spirited Away or children’s cartoons, there is an entire world of fans in their teens through twenties who avidly follow anime. A few studies have been conducted on the cinematic form and industry of Japanese animation and its spill-over to Japanese popular culture products, but almost nothing is written about the anime fan subculture in America. It appears to have arisen in 1990s and greatly expanded during the early 2000s. Casual observation suggests it is about equally among popular both genders and all ethnic-racial groups. It primarily attracts young people from the preadolescence and early teens (11–15) through early adulthood (25–28). Apparently, children discover the Japanese-style cartoons, and some become attracted to more sophisticated animation forms as well as video game spin offs. The scant journalistic commentary on anime fans implies that many are “social misfits” or “geeks.” They do not cause trouble or break laws, but they do not fit in with mainstream peers. Many U.S. fans have self-adopted the Japanese term otaku (which translates as an obsessed misfit/geek and has negative connotations) as a badge of honor. Some apparently excel at academics, but few appear seriously involved in sports or other social activities common for their age group. There is speculation that these young people are bright and pulled into a fantasy world

that offers rapid action escape, adventure, morality tales, and intrigue. Somewhat socially separated from peers, they apparently seek out others with the same interest. While most appear socially adjusted and operate in daily life without serious difficulties, a few withdraw and devote more time in the fantasy world of animation than in reality. Reactions by parents and other adults who work with young people (teachers, librarians) are not known. There are many forms of anime. Most genres have an adventure-fantasy theme, but some offer elaborate alternative worlds, and others are very violent or graphically p*rnographic. With little formal adult or institutional support, the fans seek one another out to form clubs at schools, libraries, or community centers at which they watch and discuss their favorite characters and tales. They organize conventions at the state, regional, or national level. The role of the anime production and distribution industry in these is unknown. At the conventions, they discuss and analyze the animation stories and also engage in dress-up or “cosplay” (a Japanese term for costume play). It appears that many young people dress as their favorite characters then admire one another’s costumes and interact in ways that mimic the character. There are many products (posters, clothing, trinkets) sold to fans but little is known about the people who produce and sell these products.

Research Objective andProcedure 1. Research Objective This is an exploratory, qualitative study, in which we seek to describe the anime fan culture. Our goal is to gather preliminary information that can be used for a future study.

279

280 Appendix A

2. Research Participants The principal investigator and/or trained assistant will locate anime fans at clubs, anime conventions, and through referrals. The exact number is not clear since this is an exploratory and uncharted area. Additional fans will be located using purposive and snowball (referral) sampling. We hope to locate a least 30 fans for interviews. The age, race, gender make-up is unknown but will probably involve an equal mix of gender, all racial-ethnic groups, and persons aged 13 to 30 years. Persons under the age of 13 will be excluded.

3. Research Procedure The principal investigator and/or trained adult college student assistants will personally observe fans in public places using participant observation techniques, conduct informal small talk-conversations, and make arrangements to interview participants at a later time. We may take a few brief notes (such as a person’s name and address or phone number) in the field setting, but take extensive notes of the club activities and convention events after the observation. While at meetings and conventions, we will gather names for future interviews or conduct interviews after club meetings or during conventions. The interviews will be open-ended and tape-recorded. See Appendix for questions/topics in the interviews. Prior to interviewing or tape-recording, we will explain the purpose of the study to participants and tell them that their involvement is voluntary. We will collect the names of participants but hold them in confidence. Personal identifiers (age, gender, etc.) will be released to the public in a way that protects the identity of participants. Because some anime fans may be under the age of 18, we will obtain parental permission prior to interviews. We expect interviews to vary in length (ten minutes to an hour) and may take place on more than one occasion in a semi-private location (e.g., room alcove, table or booth in a restaurant). Interview questions will not be fixed prior to interviewing but will follow a general list of topics (see below). We may ask different participants different questions based on their early responses. We will listen to tape recordings and take notes, but not transcribe the interviews. We may take photos, with permission, of conventions and participants, and will collect artifacts (e.g., announcements, brochures on conventions, etc.). We will document the types of products (shirts, posters, etc.) sold to fans at the conventions to identify patterns and trends.

4. Anticipated Results As an exploratory study into a relatively unknown area, we can only speculate about possible results. We will describe anime fan activities (clubs, conventions, etc.) and

characteristics of the fans we interview. We will identity repeated themes and patterns within the clubs/conventions and examine themes in anime fan conversations, and about fan social activities. We will use the results to develop a more systematic study into the anime subculture.

5. Schedule and Budget Schedule Months 1–2 Locate fan clubs and conventions and scan Internet sites. Months 3–4 Visit clubs and conventions, interview fans. Months 5–6 Assemble and organize collected materials and field notes to analyze them. Budget Supplies notebooks tape recorder and supplies (batteries) Travel To clubs and conventions

Appendix Questions in the interviews will include the following topics: • What is your age? If in college, what is your major and GPA? • Do you have a part-time job, what is it? • What are your career goals and aspirations for the future? • What is your favorite anime film/character? Has this changed over time? • At what age did you first develop an interest in anime? Explain, please. • What got you interested in anime? • Have your interests/favorites changed over time? • About how many hours per week do you watch anime? • How many anime films do you personally own? • How often do you get together with friends to discuss anime? • Of all those who you consider close friends, how many follow anime? • Do you engage in cosplay or other anime social activities? • What other hobbies or interests do you have besides anime? • If you have social activities, what proportion are centered on anime? • About how much money do you spend on anime per month?

Appendix A 281

• Do you have any other Japan-related interests other than anime?

• Do you encourage people younger than yourself to learn about anime?

• Do you have friends who were once interested in anime but dropped out?

• How would you describe your relations with your parents?

• What do you and your anime friends talk about?

• Does anime relate to your sexuality and interest in sex in any way?

• What interests you in anime, how does it make you feel? • Do you ever get very angry or upset watching anime? • What types of anime do you like? Dislike? • What types do you feel excited by or bored by? • What do you think about people who do not like anime? • Do you watch/play video games related to anime? • In general, what type of student are you? • In what ways, if any, does anime relate to your school work? • Do you have any other hobbies or interests based on your anime interest? • Do you think you will always love anime? Why or why not?

Question Topics for Adult Participants (18 and older) Only • Do you currently consume alcohol or recreational drugs while viewing anime? • Did you consume alcohol or recreational drugs while viewing anime when you were younger, under 18? • Of your sexual partners, do many share your interest in anime? • Do you get sexually aroused when you watch anime?

Example Research Proposal A quantitative study of Anime Enthusiasts in the United States Introduction and Theory We all “consume” many popular cultural products, such as media forms (e.g., video or music), food dishes, electronic devices, and so forth. Most achieve a mass distribution, but some are specialized. Specialized products can attract a small number of people who become enthusiasts. At times, the enthusiasts develop social relations with one another, exchange information, and discuss product details. More than being casual consumers, their interest includes studying, collecting, and becoming experts on the products. By communicating and interacting, they might develop a distinct subculture around the products. A cultural product subculture is likely to develop when the cultural product is unusual or obscure, requires special knowledge, or has devotees in a geographic region or specific age group. Early in subculture formation, the devotees may meet to exchange information, create publications, or form clubs. Fans set themselves apart from “outsiders” unfamiliar or not yet entranced by the product. Their skill and expertise with the product helps them develop self-esteem and gain respect from their like-minded peers. Even more than other cultural products, media forms change “fashion” very

quickly. Young people tend to be more interested in new media that appeared in the late 20th and early 21st centuries. With fewer family or job responsibilities and a disposable income, young people are the primary consumers of popular media and may develop a “fan” following around specific products, artists, musicians, or a genre. With globalization, some cultural productions, and particularly new media, are shared across international borders and marketed to people in many countries. The cultural products are part of a developing international youth culture.

Research question This study examines the U.S. fans of the cultural media product Japanese animation, or anime. Artists, writers, and producers in Japan create anime. It has a “different” foreign or exotic look compared to most American-created animated media. Compared to traditional U.S. animation, Japanese anime is more diverse, has more complicated plots and developed characters, and appeals to a wider age range. Anime takes some themes from Japan and builds on Japanese settings or situations that are not widely known

282 Appendix A in the United States. This study looks at several questions about anime: 1. Does anime with its “foreignness” attract people who feel somewhat outside the U.S. cultural mainstream? 2. Are males and females attracted to different aspects or themes in anime that relate to emerging gender issues? 3. Over time do anime fans develop an interest in the country or culture of its origin, Japan? Each research question has broader implications. With globalization, more products from other countries for people are available. It could be that some people that are not part of a somewhat hom*ogenized mainstream culture find foreign products more attractive and a way to express their feelings of difference or individuality. One aspect of a foreign cultural product is that they might offer an alternative set of social relations or cultural viewpoint, even it if is not realistic or easy to act upon. When a cultural product of foreign origin includes some elements of the foreign culture or country, avid consumers of the product may develop an interest in the foreign country or culture as an ancillary effect of their devotion and interest in the cultural product.

Literature Review Several studies of anime productions have documented the themes and situations borrowed from Japanese culture in the stories, characters, and situations (see references). Anime films fall into a set of categories (fantasy, adventure, and so forth) and is somewhat differentiated by age and gender. Both male and female characters are often shown with superpowers, and changing or ambiguous gender of characters is present in several anime series (reference). Other studies of the anime industry emphasize its rapid growth and connection to other media, Japanese-style comics and video games (reference). A few theorists emphasized anime as part of a transnational youth culture (reference). Only two studies (see reference) have looked at anime fans. One unpublished study found that college student anime fans were first attracted to anime but knew very little about Japan or had little interest in Japan. Another study that was a doctoral dissertation found that while many anime fans fit a “geek” or “nerd” stereotype, this was not universal. All did share a strong interest in media (watching film) and related media product (video games), and most began at a young age.

Hypotheses HY 1: Persons with fewer “mainstream” hobbies or interests, and with fewer “mainstream” close friends are likely to be stronger anime fans.

HY 2: Female anime fans identify more with androgynous/ ambiguous gendered characters than male fans. HY 3: Intense and committed anime fans are the most likely to want to learn about Japanese culture, learn the Japanese language, and/or wish to visit the country of Japan.

Method 1. Sample The population is self-identified anime fans between the ages of 13 and 26. A fan is defined as someone who attends club meetings or a convention that is focused on anime, or who describes his or her main hobby as watching anime films, dressing and acting as anime characters, and/or talking with other anime enthusiasts about anime films. We will draw a stratified random sample of 200 fans from university, school, and community anime clubs in the a three state geographic area and conventions. First, we will identify ten anime clubs or conventions and attend multiple meetings to obtain lists of members or attendees. Next, we will create a sampling frame that has the names, ages, addresses, phone numbers, and e-mails of the members, attendees, and selfdescribed fans at the clubs or conventions. Next, we will divide the sampling frame into two age groups: (1) fans aged 13–18, and (2) those aged 19–26. Finally, we will draw a random sample of 100 names from each age group. We will contact each sampled person to set up an interview. For persons under 18 years of age, we will use a twostage process. First, we will contact the sampled person and request the name, address, and phone number of a parent or legal guardian. Before scheduling an interview, we will mail the parent or legal guardian an informed consent form that explains the study and asks permission to interview the legal minor along with a stamped return envelope. We will telephone parents who fail to respond in seven days to make an oral request and offer a second informed consent form. If we cannot contact or obtain permission to interview a sampled name after six tries using phone, regular mail, and e-mail, we will randomly draw a replacement name from the same age-stratified sampling frame until we have 200 people who agree to be interviewed.

2. Data Collection Procedure We will conduct face-to-face or telephone interviews with each respondent. We estimate that about one-half will be face-to-face and one-half by telephone, depending on logistics and scheduling. We will ask permission to tape record all telephone interviews. After obtaining permission to record, we will read an informed consent statement prior to interviewing. For face-to-face interviews, we will ask each respondent to sign an informed consent form.

Appendix A 283

Informed consent will be obtained for persons under 18 in addition to a parent or guardian consent form. We will conduct the face-to-face interviews in any public place (school grounds, shopping mall, or restaurant) but without another person participating and no one listening in. After completing a questionnaire, we will offer to send a copy of the report to a respondent and provide contact information should he or she has further questions. We will number and store the questionnaires and begin to enter data from each questionnaire into a statistical computer program after the first twenty are completed. We anticipate spending 10 minutes per interview locating and arranging for the interview, and the interview itself to take about 15 minutes to complete. We estimate about 20 minutes to travel to and meet with people for each face-to-face interview and 20 minutes to transcribe

Variable Name

Questionnaire Item

each of the telephone interviews. To complete the 200 interviews it will take about 33 hours for locating and scheduling, 50 hours for actual interviewing, and 67 hours for traveling and transcribing.

3. Variable Measurement We will create three general measures by combining multiple survey questions to measure: (1) being a “nerd” or “geek” or being outside the mainstream of U.S. culture, (2) being a committed anime fan based on years of watching anime, spending time with anime and other fans, and expressing a strong interest in anime, and (3) having an interest in learning about Japan and Japanese culture. We anticipate the questionnaire will have about 40 items. Examples of some preliminary questions to be in the final questionnaire are listed below.

1. Gender

Are you, Male Female LGBT

2. Age

How old are you now?

3. School

Do you now attend school? If No Next, If Yes, what school/grade

4. Start

At what age did you first start watching anime regularly?

5. Games

Do you play video games? If No skip to #8, If Yes, how often? daily 2–3 times a week weekly less often

6. Games 1

What are your favorite two games (1)___________________________________

7. Games 2

(2)

8. Own

How many anime DVDs do you personally own?

9. Often

How often do you watch anime? every day 3–5 times a week about once a week several times a month less than once a month

10. Where

Where do you usually watch anime films?

11. Friends

Of your five best friends, how many are anime fans?

12. Alone

Think back to the last 10 times you watched an anime film. How many of those 10 times were you watching it alone?

13. Favorite1

Name your four favorite anime films of all time, (1)

14. Favorite2

(2)

15. Favorite3

(3)

16. Favorite4

(4)

17. Character1

Name your two favorite anime characters of all time, (1)

18. Character2

(2)

19. Products

Do you own any anime-related products, such as posters, clothing, stuffed animals, etc.? If No skip to #21,

20. Type

If Yes, what products do you own, type and number?

21. Cosplay

Do you ever dress up in a costume as an anime character? Yes No

22. Japan1

Have you ever read a book about Japanese history or society?

23. Japan2

Have you ever traveled to Japan? If Yes #25, If No,

24. Japan3

How interested are you in traveling to Japan? Extremely Very Somewhat A little Not at All

25. Club

Do you belong to an Anime Club, No Yes

284 Appendix A 26. Conven1

Have you ever attended an anime convention? If No skip to #28, If Yes,

27. Conven2

How many conventions have you attended in the past three years?

28. FriendG

Of your five closest friends, how many are your same gender?

29. Internet

How often do you go to anime related sites on the Internet? Never once a month several x a month Weekly Daily or more

30. Magazine

Do you subscribe to an anime magazine? Yes No

Time Schedule

Budget Estimate

Month 1 Obtain IRB approval, continue literature review, prepare draft of complete questionnaire, develop list of anime clubs and conventions Months 2–3 Visit anime clubs and conventions to collect names and create sampling frame, pilot test questionnaire Month 4 Draw a random sample of names and contact under 18 sample for parental permissions, revise questionnaire, begin to schedule interviews Contact and arrange for all interviews, Months 5–7 begin interviewing Month 8 Complete last interviews, and start to code data into computer program Month 9 Finish coding, and analyze data using statistics program Month 10 Write up results as a report and present findings

Supply and Service Expenses Printing and postage Tape recorder and supplies Telephone Travel Expenses To go to anime clubs and conventions To go to interviews To go to professional meeting to present final report Labor Expenses General clerical help Interviewing help Tape transcription help Data entry help Statistical analysis

Appendix B

Data and Literature Research Sources of Online Statistics It is important that research be supported by up-to-date information. There are many online databases that will provide the latest statistics on many topics of interest. Some examples are as follows: Albany.edu/sourcebook: The Sourcebook of Criminal Justice Statistics has over 600 tables, covering diverse aspects of criminal justice in the United States. CDC.gov: The Centers for Disease Control and Prevention provides health statistics, and is an excellent source for birth and mortality data. Census.gov: The U.S. Census Bureau provides extensive population statistics, including online access to the Statistical Abstract. FedStats.gov: This site provides data and information from over 100 federal government agencies. FedStats also has links to many other helpful resources. Gallup.com: The Gallup Poll is an excellent source of public opinion statistics. Some of this information is free online, and some polls require a subscription. ICPSR.umich.edu: The Inter-University Consortium for Political and Social Research is the world’s largest archive of social science statistical data. The site houses data sets that can be analyzed with statistical software, but also provides publications based on the data collected. NCES.ed.gov: The National Center for Education Statistics analyzes and provides data collected from U.S. schools. Thearda.com: The American Religion Data Archive has demographic, church membership, and religious practice information for the United States.

Literature Search Databases Literature databases can help you navigate the seemingly limitless amounts of professional journals and research articles that are available electronically. College libraries have subscriptions to some databases, allowing students to download full-text articles, in many cases. Examples of literature databases include the following: EBSCO: College libraries have free access to EBSCOhost, a database that allows users to search an extensive collection of popular and scholarly publications.

ERIC: Free via the Internet, the ERIC database provides abstracts to over one million unpublished documents and published articles on educational research and practice. Ingentaconnect: The Ingenta database allows the user to search an extensive collection of abstracts by most major journal publishers. Ingenta is an excellent source for browsing journals by academic discipline. JSTOR: JSTOR has digitally preserved hundreds of scholarly journals, including historical documents. Through this database, researchers can access back runs of journal articles, spanning over 40 disciplines. Medline: MEDLINE has over 10 million searchable records of life science and biomedical information. PsycARTICLES: PsycARTICLES is a database of full-text articles from the more than 50 journals published by the American Psychological Association.

Selected Peer Reviewed Journals, by Discipline Anthropology and Archaeology American Anthropologist American Antiquity American Ethnologist American Journal of Archaeology Annual Review of Anthropology Ethos Folklore Man

Criminology and Legal Studies Aggressive and Violent Behavior Crime and Delinquency Criminal Justice and Behavior Criminology Journal of Criminal Justice Journal of Criminal Law and Criminology Journal of Interpersonal Violence Journal of Quantitative Criminology Journal of Research in Crime and Delinquency Justice Quarterly

285

286 Appendix B Law and Society Review Victimology

Education American Educational Research Journal Computers and Education Contemporary Educational Psychology Early Childhood Research Quarterly Educational Leadership Educational Psychology Educational Researcher International Journal of Educational Research Journal of Education Policy Journal of Special Education Learning and Instruction Literacy Review of Educational Research Review of Research in Education Teaching and Teacher Education Teaching Exceptional Children

General Business and Consumer Behavior Administrative Science Quarterly Advances in Consumer Research Business and Society Review Consumption, Markets and Culture Harvard Business Review Journal of Consumer Behavior Journal of Consumer Culture Journal of Consumer Marketing Journal of Consumer Research Journal of Marketing

Human Development, Family Studies, Gerontology Child Development Developmental Psychology Families in Society Family Relations The Gerontologist Human Development Infant Behavior and Development Journal of Adolescence Journal of Gerontology Journal of Human Development Journal of Marriage and Family The Journal of Sex Research

Human Resource Management Academy of Management Journal Academy of Management Review Human Resource Management Industrial and Labor Relations Review International Journal of Human Resource Management Journal of Human Resources Journal of Organizational Behavior Organizational Behavior and Human Decision Processes Personnel Psychology Work and Occupations

Political Science American Journal of Political Science The American Political Science Review Comparative Politics International Studies Quarterly The Journal of Conflict Resolution The Journal of Politics Political Behavior Political Science Quarterly Politics and Society The Public Opinion Quarterly Social Science Quarterly World Politics

Psychology American Journal of Psychology American Psychologist Journal of Abnormal Psychology Journal of Applied Social Psychology Journal of Personality and Social Psychology Psychological Bulletin Psychological Methods Psychological Review Psychological Science Social Psychology Quarterly

Sociology American Journal of Sociology American Sociological Review Annual Review of Sociology Comparative Studies in Society and History Current Sociology Demography Gender and Society Journal of Health and Human Behavior Journal of Marriage and the Family Social Forces

Appendix B 287

Social Problems Sociological Quarterly

Social Work Affilia: Journal of Women and Social Work Child Abuse and Neglect Child Welfare

Health and Social Work Journal of Social Service Research Journal of Social Work Education Research on Social Work Practice Social Service Review Social Work Social Work Research

References Chapter 1 References Anderson, Elijah. 2012. The Cosmopolitan Canopy: Race and Civility in Everyday Life. New York: W.W. Norton. Bernhardt, Amy M., Cara Wilking, Anna M. Adachi-Mejia, Elaina Bergamini, Jill Marijnissen, and James D. Sargent. 2013. “How television fast food marketing aimed at children compares with adult advertisem*nts.” PLoS ONE 8(8): e72479. http://www.plosone. org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0072479 downloaded November 9, 2014. Brown, Joel H., and Ita G. G. Kreft. 1998. “Zero effects of drug prevention programs: Issues and solutions.” Evaluation Review 22:3–14. Elbel, B., J. Gyamfi and R. Kersh. 2011. “Child and adolescent fastfood choice and the influence of calorie labeling: A natural experiment.” International Journal of Obesity 35:493–500. Gantz, Walter, Nancy Schwartz, James R. Angelini, and Victoria Rideout. 2007. Food for Thought, Television Food Advertising to Children in the United States. Menlo Park, CA: Henry J. Kaiser Family Foundation. http://kff.org/other/food-for-thought-televisionfood-advertising-to/ (accessed December 1, 2014). Gusrang, Jamie L. and Simon Cheng. 2010. “Government influence and community involvement on abstinence-only programs in 1999 and 2003.” American Journal of Sexuality Education 5:240–267. Guttmacher Institute. 2006. “Facts on sex education.” In Brief (online publication). December 2006. http://www.guttmacher.org/pubs/ fb_sexEd2006.html (accessed December 1, 2014). Harris Interactive. 2009. “More Americans believe in the devil, hell and angels than in Darwin’s Theory of Evolution.” http://www. harrisinteractive.com/vault/Harris-Interactive-Poll-Research- Religious-Beliefs-2008-12.pdf (accessed Nov 9, 2014). Heneghan, Carl, et al. 2012. “The evidence underpinning sports performance products: a systematic assessment.” BMJ Open 2012 http://bmjopen.bmj.com/content/2/4/e001702.full. pdf+html?sid=cc4b5c8d-2fda-4654-9667-aa3884f32bb6 (accessed November 9, 2014). King, Laura A., Chad Burton, Joshua Hicks, and Stephen Drigotas. 2007. “Ghosts, UFOs, and magic: Positive affect and the experiential system.” Journal of Personality & Social Psychology 92(5):905–919.

Pew Research Center Report. 2007. “Public knowledge of current affairs little changed by news and information revolutions” http:// www.people-press.org/2007/04/15/public-knowledge-of-currentaffairs-little-changed-by-news-and-information-revolutions/ (accessed November 9, 2014). Pew Research Center Report. 2009. “Public praises science; scientists fault public, media” http://www.people-press.org/2009/07/09/ public-praises-science-scientists-fault-public-media/ (accessed November 9, 2014). Robinson, Thomas N., Dina Borzekowski, Donna Mutheson, and Helena Kraemer. 2007. “Effects of fast food branding on young children’s taste preference.” Archives of Pediatrics and Adolescent Medicine 161:792–796. Roksa, Josipal and Tania Levey. 2010. “What can you do with that degree? College major and occupational status of college graduates over time.” Social Forces 89(2):389–415. Rozenblit, Leonid and Frank Keil. 2002. “The misunderstood limits of folk science: An illusion of explanatory depth.” Cognitive Science 26(5): 521–562. Scheitle, Christopher and Bryanna Hahn. 2011. “From the pews to policy: Specifying evangelical Protestantism’s influence on states’ sexual orientation policies.” Social Forces 89(3):913–933. Sherkat, Darren, Kylan Mattias de Vries, and Stacia Creek. 2010. “Race, religion, and opposition to same-sex marriage.” Social Science Quarterly 91(1):80–98. Stutts, Mary Ann, Gail Zank, Karen Smith, and Sally Williams. 2011. “Nutrition information and children’s fast food menu choices.” Journal of Consumer Affairs 45(1):52–86. Tobler, Nancy and Howard H. Stratton. 1997. “Effectiveness of school-based drug prevention programs: A meta-analysis of the research.” Journal of Primary Prevention 18: 71–128. Tandon, Pooja S., et al. 2011. “The impact of menu labeling on fast-food purchases for children and parents.” American Journal of Preventative Medicine 41(4):434–438. Whyte, William. 1984. Learning from the Field. Beverly Hills, CA: Sage. Wysong, E., et al. 1994. “Truth and DARE: Tracking drug education to graduation and as symbolic politics.” Social Problems 41:448–472.

Larsen, Alli Orr. 2014 (forthcoming). “The Trouble with Amicus Facts.” Virginia Law Review. William & Mary Law School Research Paper No. 09-273. http://papers.ssrn.com/sol3/papers. cfm?abstract_id=2409071

Zagumny, M. J. and M. K. Thompson. 1997. “Does D.A.R.E. work? An evaluation in rural Tennessee.” Journal of Alcohol and Drug Education 42:32–41.

Liptak, Adam. 2014. “Seeking Facts, Justices Settle for What Briefs Tell Them.” New York Times, September 1, 2014.

Chapter 2 References

Mathematica Policy Research, Inc. 2007. Impacts of Four Title V, Section 510 Abstinence Education Programs, Final Report. Princeton, NJ: Mathematica Policy Research, Inc. http://www.mathematica-mpr. com/~/media/publications/PDFs/impactabstinence.pdf (accessed December 1, 2014). Mills, Candice M. and Frank C. Keil. 2004. “Knowing the limits of one’s understanding: The development of an awareness of an illusion of explanatory depth.” Journal of Experimental Child Psychology. 87(1):1–32. Otten, Jennifer J., et al., 2012. “Food marketing to children through toys: Response of restaurants to the first U.S. toy ordinance.” American Journal of Preventive Medicine 42(1):56–60.

288

Atkinson. Michael. 2003. Tattooed: The Sociogenesis of a Body Art. Toronto: University of Toronto Press. Atkinson, Michael. 2004. “Tattooing and civilizing processes: Body modification as self-control.” Canadian Review of Sociology & Anthropology 41(2):125–146. Caplan, Jane (editor). 2000. Written on the Body. Princeton NJ: Princeton University Press. DeMello, Margo. 2000. Bodies of Inscription. Durham NC: Duke University Press. Dickson, Lynda, Richard Duke, Hilary Smith, and Noel Strapko. 2014. “Stigma of ink: Tattoo attitudes among college students.” Social Science Journal 51:268-276.

References 289 Dukes, Richard L., and Judith A. Stein. 2011. “Ink and holes: Correlates and predictive associations of body modification among adolescents.” Youth & Society 43(4):1547–1569. Fisher, Jill A. 2002. Tattooing the body, marking culture. Body & Society 8(4):91–107. Harris Interactive. 2012. “One in Five U.S. Adults Now Has a Tattoo” http://webcache.googleusercontent.com/search?q= cache:a90tIB8uP9wJ:www.harrisinteractive.com/NewsRoom/ HarrisPolls/tabid/447/mid/1508/articleId/970/ctl/ ReadCustom%2520Default/Default.aspx+&cd=1&hl=en&ct= clnk&gl=us (accessed October 20, 2014). Hawkes, Daina, Charlene Seen, and Chantal Thorn. 2004. Factors that influence attitudes toward women with tattoos. Sex Roles 50 (9/10):593–604. Heywood, Wendy, Kent Patrick, Anthony M.A. Smith, Judy M. Simpson, and Marian K. Juliet Richters. 2012. “Who gets tattoos? Demographic and behavioral correlates of ever being tattooed in a representative sample of men and women.” Annals of Epidemiology 22(1):51–56. Horne, Jenn, David Knox, Jane Zusman, and Marty Zusman, 2007. “Tattoos and piercings: Attitudes, behaviors, and interpretations of college students.” College Student Journal 41 (4):1011–1020. Kang, Miliann and Katherine Jones. 2007. “Why do people get tattoos.” Contexts 6(1):42–47. Martin, Jamin A. and Chris S. Dula. 2010. “More than skin deep: Perceptions of, and stigma against, tattoos.” College Student Journal 44(1):200–206. Resenhoeft, Annette, Julie Villa, and David Wiseman. 2008. “Tattoos can harm perceptions: A study and suggestions.” Journal of American College Health 56(5):593–596. Silver, Eric, Stacy Rogers Silver, Sonja Siennick, and George Farkas. 2011. “Bodily signs of academic success: An empirical examination of tattoos and grooming.” Social Problems 58(4):538–564.

Chapter 3 References Albergotti, Reed and Elizabeth Dwoskin. 2014. “Facebook study sparks soul-searching and ethical questions.” Wall Street Journal, June 30, 2014. Bostwick, Michael. 2012. “Blurred Boundaries: The therapeutics and politics of medical marijuana.” Mayo Clinic Proceedings 87(2): 172–186. Crawford, Chris. 2014. “AAFP calls for careful study of medical marijuana use, gun violence.” American Association of Family Physicians News, October 1, 2014. http://www.aafp.org/news/ health-of-the-public/20141001policyrndup.html (accessed December 1, 2014).

http://www.apa.org/science/about/psa/2013/02/gun-violence. aspx (accessed May 11, 2016). Science Agenda, February 2013. http://www.apa.org/science/about/ psa/2013/02/gun-violence.aspx (accessed December 1, 2014). Kirt, David and Andrew Papachristos. 2011. “Cultural mechanisms and the persistence of neighborhood violence.” American Journal of Sociology 116(4):1190–1233. Klitzman, Robert. 2014. “Did Facebook’s experiment violate ethics?” CNN.com, July 2, 2014 http://edition.cnn.com/2014/07/02/opinion/ klitzman-facebook-experiment/ (accessed December 1, 2014). Klitzman, Robert and Paul S. Appelbaum. 2014. “Facebook’s Emotion Experiment: Implications for Research Ethics.” Bioethics Forum, July 21, 2014. http://www.thehastingscenter.org/Bioethicsforum/ Post.aspx?id=7002&blogid=140 (accessed December 1, 2014). Kramera, Adam, Jamie Gillory, and Jeffrey Hanco*ck. 2014. “Experimental evidence of massive-scale emotional contagion through social networks.” PNAS, Proceedings of the National Academy of Sciences of the United States of America 111(24):8788–8790. Meyer, Robinson. 2014. “Everything we know about facebook’s secret mood manipulation experiment, it was probably legal. But was it ethical?” The Atlantic. Jun 28, 2014. Milgram, Stanley. 1963. “Behavioral study of obedience.” Journal of Abnormal and Social Psychology 6:371–378. Milgram, Stanley 1965. “Some conditions of obedience and disobedience to authority.” Human Relations 18:57–76. Milgram, Stanley. 1974. Obedience to Authority. New York: Harper and Row. Mooney, Chris. 2005. The Republican War on Science. New York: Perseus Books. Oreskes, Naomi and Erik M. Conway. 2010. Merchants of Doubt: How a Handful of Scientists Obscured the Truth on Issues from Tobacco Smoke to Global Warming. New York: Bloomsbury Publishing. Piliavin, Irving, J. Rodin, and Jane Piliavin. 1969. “Good Samaritanism: An Underground Phenomenon?” Journal of Personality and Social Psychology 13:289–299. Savelsberg, Joachim, Lara Cleveland, and Ryan King. 2004. “Institutional environments and scholarly work: American criminology, 1951–1993.” Social Forces 82:1275–1302. Savelsberg, Joachim, Ryan King, and Lara Cleveland. 2002. “Politicized Scholarship? Science on Crime and the State.” Social Problems 49:327–349. Scarce, Rik. 1994. “(No) Trial (But) Tribulations: When courts and ethnography conflict.” Journal of Contemporary Ethnography 23:123–149.

Draus, Paul J., Harvey Siegal, Robert Carlson, Russel Falck, and Jichuan Wang. 2005. “Cracking the Cornfields.” Sociological Quarterly 46:165–189.

Scarce, Rik. 1999. “Good Faith, Bad Ethics: When scholars go the distance and scholarly associations do not.” Law & Social Inquiry 24:977–986.

Goel, Vindu. 2014. “After Uproar, European Regulators Question Facebook on Psychological Testing.” New York Times, July 2, 2014.

Taylor, Steven. 1987. “Observing abuse.” Qualitative Sociology 10:288–302.

Gorski, David. 2014. “Did Facebook and PNAS violate human research protections in an unethical experiment?” Science Based Medicine, June 30, 2014. http://www.sciencebasedmedicine.org/ did-facebook-and-pnas-violate-human-research-protections-in-anunethical-experiment/(accessed December 1, 2014).

Underwood, Emily. 2013. “Gun control agenda is a call to duty for Scientists.” Science 339(6118):381–382.

Gutting, Gary. 2012. “The N.R.A.’s Blockade on Science.” New York Times, December 20, 2012. Harris, Gardiner. 2010. “Researchers find study of medical marijuana discouraged.” New York Times, January 18, 2010. Humphreys, Laud. 1973. Tearoom Trade. Chicago: Aldine. Jamieson, Christine. 2013. “Gun violence research: History of the federal funding freeze.” Psychological Science Agenda, February 2013,

Van Maanen, John. 1982. “Fieldwork on the beat.” Varieties of Qualitative Research. Edited by J. Van Mannen, J. Dabbs, Jr., and R.Raulkner (pp. 102–151). Beverly Hills, CA: Sage. Vidich, Arthur and Joseph Bensman. 1968. Small Town in Mass Society, rev. ed. Princeton, NJ: Princeton University Press. Yong, Ed, Heidi Ledford, and Richard Van Noorden. 2013. “Research ethics: 3 ways to blow the whistle: Reporting suspicions of scientific fraud is rarely easy, but some paths are more effective than others.” Nature 503(7477):454–457. Zimbardo, Philip. 1972. “The Pathology of Imprisonment.” Society 9:4–6.

290 References Chapter 4 References Browning, Sarah E. 2012. “Neighborhood, school, and family effects on the frequency of alcohol use among Toronto youth.” Substance Use & Misuse 47:31–43, 2012. Draus, Paul J., Harvey Siegal, Robert Carlson, Russell Falck, and Jichuan Wang. 2005. “Cracking in the heartland.” Sociological Quarterly 46:165–189. Hemenway, David, Mary Vriniotis, Renee M. Johnson, Matthew Miller, and Deborah Azrael. 2011. “Gun carrying by high school students in Boston, MA: Does overestimation of peer gun carrying matter?” Journal of Adolescence 34:997–1003. Kirk, David S. and Andrew V. Papachristos. 2011. “Cultural mechanisms and the persistence of neighborhood violence.” American Journal of Sociology 116(4):1190–1233.

upport for Conciliatory Measures in the Context of Violent InterS group.” Journal of Research in Crime & Delinquency 51(5):585–619. Rosenberg, Morris. 1965. Society and the adolescent self-image. Princeton, NJ: Princeton University Press. Shirley, Carla D. 2010. “‘You might be a redneck if . . .’ Boundary work among rural, Southern Whites.” Social Forces 89(1):35–51. Van Laar, Colette, Shana Levin, Stacey Sinclair, and Jim Sidanius. 2005. “The effect of university roommate contact on ethnic attitudes and behavior.” Journal of Experimental Social Psychology 41:329–345. Yu Xie and Xiang Zhou. 2012. PNSA (Proceedings of the National Academy of Sciences) 109(29):11646–11651.

Chapter 6 References

Newman, Katherine S. 2008. “Ties That Bind: Cultural Interpretations of Delayed Adulthood in Western Europe and Japan.” Sociological Forum 23(4):645–466.

Davis, Darren and Brian Silver. 2003. “Stereotype Threat and Race of Interview Effects in a Survey on Political Knowledge.” American Journal of Political Science 47:33–45.

Paul, Anju Mary. 2011. “Stepwise international migration: a multistage migration pattern for the aspiring migrant.” American Journal of Sociology 116(6):1842–1886.

Foddy, William. 1993. Constructing questions for interviews and questionnaires. New York: Cambridge University Press.

Chapter 5 References Ayers, John W., C. Richard Hofstetter, Keith Schnakenberg, and Bohdan Kolody. 2009. “Is immigration a racial issue? Anglo attitudes on immigration policies in a border county.” Social Science Quarterly 90(3):593–610. Babbie, Earl. 2004. The Practice of Social Research, 10 ed. Belmont CA: Wadsworth. Collier, Kate L., Henry M. W. Bos, Henny M. W., and Theo G. M. Sandfort. 2012. “Intergroup contact, attitudes toward hom*osexuality, and the role of acceptance of gender non-conformity in young adolescents.” Journal of Adolescence 35(4):899–907.

Hurwitz, Jon and Mark Peffley. 2005 “Playing the Race Card in the PostWillie Horton Era: The Impact of Racialized Code Words on Support for Punitive Crime Policy.” Public Opinion Quarterly 69(1):99–112. Presser, Stanley. 1990. “Measurement Issues in the Study of Social Change.” Social Forces 68:856–868. Schuman, Howard, and Stanley Presser. 1981. Questions and answers in attitude surveys: Experiments on question form, wording and content. New York: Academic Press. Smith, Tom W. 1987. “That which we call welfare by any other name would smell sweeter.” Public Opinion Quarterly 51(1):75–83. Wilson, David C. 2010. “Perceptions about the amount of interracial prejudice depend on racial group membership and question order.” Public Opinion Quarterly 74(2):344–356.

Desmond, Matthew. 2012a. “Eviction and the reproduction of urban poverty.” American Journal of Sociology 118(1):88–133.

Chapter 7 References

Desmond, Matthew. 2012b. “Disposable ties and the urban poor.” American Journal of Sociology 117(5):1295–1335.

Elbel, B., J. Gyamfi, and R. Kersh. 2011. “Child and adolescent fastfood choice and the influence of calorie labeling: A natural experiment.” International Journal of Obesity 35:493–500.

Duncan, Otis Dudley and Beverly Duncan. 1955. “A methodological analysis of segregation indexes.” American Sociological Review 20:210–217. Ellison, Christopher G., Heeju Shin, and David L. Leal. 2011. “The Contact Hypothesis and Attitudes Toward Latinos in the United States.” Social Science Quarterly 92(4):938–958. Gee, Gilbert C., Michael S. Spencer, Juan Chen, and David Takeuchi. 2007. “A nationwide study of discrimination and chronic health conditions among Asian Americans.” American Journal of Public Health 97:1275–1282. Heise, David R. 1970. “The Semantic Differential and Attitude Research.” In Attitude Measurement, edited by Gene F. Summers. Chicago: Rand McNally, pp. 235–253. Jon, Jae-Eun. 2013. “Realizing Internationalization at Home in Korean Higher Education: Promoting Domestic Students’ Interaction with International Students and Intercultural Competence.” Journal of Studies in International Education 17(4):455–470. Kleg, Milton and Kaoru Yamamoto. 1998. “As the world turns.” Social Science Journal 35:183–190. Marinellia, Nicola, et al. 2014. “Generation Y, wine and alcohol. Asemantic differential approach to consumption analysis in Tuscany.” Appetite 75(1):117–127. Parrillo, Vincent N. and Christopher Donoghue. 2013. “The national social distance study: Ten years later.” Sociological Forum 28(1):597–614. Pickett, Justin T., Thomas Baker, Christi Metcalfe, Marc Gertz, and Rose Bellandi. 2014. “Contact and Compromise: Explaining

Locke, Kenneth D. Locke, Traci Craig, Kyoung-DeokBaik, and KrutikaGohil. 2012. “Binds and bounds of communion: Effects of interpersonal values on assumed similarity of self and others.” Journal of Personality and Social Psychology 103 (5):879–897. Palmgreen, Philip, Lewis Donohew, Elizabeth PugzlesLorch, Rick Hoyle, and Michael Stephenson. 2001. “Television campaigns and adolescent marijuana use.” American Journal of Public Health 91:293–296. Payne, B. Keith. 2001. “Prejudice and Perception.” Journal of Personality and Social Psychology 81:181–192. LaMarre, Heather L., Silvia Knobloch-Westerwick, and Gregory J. Hoplamazian. 2012. “Does the music matter? Examining differential effects of music genre on support for ethnic groups.” Journal of Broadcasting & Electronic Media 56(1):150–167. Van Boven, Leaf, Charles M. Judd, and David K. Sherman. 2012. “Political polarization projection: social projection of partisan attitude extremity and attitudinal processes.” Journal of Personality and Social Psychology 103(1):84–100. Zhong, Chen-Bo and Katie Liljenquist. 2006. “Washing Away Your Sins.” Science [September 8, 2006] 313(5792):1451–1452.

Chapter 8 References Brown, Hana. 2012. “Race, legality, and the social policy consequences of anti-immigrant mobilization.” American Sociological Review 78(2):290–314.

References 291 Chavez, Leo. 2001. Covering Immigration. Berkeley: University of California Press.

Rodriquez, Jason. 2011. “‘It’s a dignity thing’: Nursing home care workers’ use of emotions.” Sociological Forum 26(2):265–286.

Brady, David and Ryan Finnigan. 2014. “Does immigration undermine public support for social policy?” American Sociological Review 79(1):17–42.

Perry, Pamela. 2001. “White means never having to say you’re ethnic: White youth and the construction of ‘cultureless’ identities.” Journal of Contemporary Ethnography 30(1):56–91.

Foster, Gary, Richard Hummel, and Donald Adamchak. 1998. “Patterns of conception, natality and morality from Midwestern cemeteries.” Sociological Quarterly 39:473–490.

Sherman, Rachel. 2006. Classic Acts: Service and Inequality in Luxury Hotels. Berkeley: University of California Press.

Frendreis, John and Tatalovich Raymond. 2013. “Secularization, modernization, or population change: Explaining the decline of prohibition in the United States.” Social Science Quarterly 94(2):379–394.

Van Maanen, John. 1991. “The Smile Factory: Work at Disneyland.” Pp. 58–76 in P. Frost, L. Moore, M. Luis, C. Lundberg, and J. Martin (eds.), Reframing Organizational Culture. Thousand Oaks, CA: Sage.

Maghelal, Praveen K. and Cara Jean Capp. 2011. “Walkability: A review of existing pedestrian indices.” Journal of the Urban & Regional Information Systems Association 23(2):5–19.

Chapter 11 References

Quinn, Malcolm. 1994. The Swastika: Constructing the Symbol. New York: Routledge.

Calavita, Kitty. 2005. Immigrants at the margins: Law, race and exclusion. New York: Cambridge University Press.

Rashotte, Lisa Slattery. 2002. “What does that smile mean? The meaning of nonverbal behaviors in social interaction.” Social Psychology Quarterly 65(1): 92–102. Rathje, William and Cullen Murphy. 1992. Rubbish: The Archaeology of Garbage. New York: Vintage.

Blee, Kathleen. 1991. Women of the Klan. Berkeley: University of California Press.

Campbell, Michael C., and Heather Schoenfeld. 2013. “The transformation of America’s penal order: A historicized political sociology of punishment.” American Journal of Sociology 118(5): 1375–1423.

Stevenson, Richard W. 1996. “U.S. to revise its estimate of layoffs.” New York Times, Oct 16, 1996.

Ferree, Myra Marx, William Gamson, Jurgen Gerhards, and Dieter Rucht. 2002. Shaping Abortion Discourse. New York: Cambridge University Press.

Vanderbilt, Tom. 2012. “What’s Your Walk Score?” Slate. http:// www.slate.com/articles/life/walking/2012/04/walking_in_ america_how_walk_score_puts_a_number_on_walkability_.html (downloaded February 5, 2015).

Kriesi, Hanspeter and Dominique Wisler. 1999. “The Impact of Social Movements on Political Institutions” Pp. 42–65 in How Social Movements Matter, (Ed.) M. Giugni, D. McAdam, and C. Tilly. Minneapolis MN: University of Minnesota Press.

Chapter 9 References Batalova, Jeanne and Philip N. Cohen. 2002. “Premarital cohabitation and housework: Couples in cross-national perspective.” Journal of Marriage and Family 64:743–755. Copen, Casey E., Kimberly Daniels, and William Mosher. 2013. “First premarital cohabitation in the United States, 2006–2010 National Survey of Family Growth.” National Health Statistics Reports no. 64, April 4, 2013 (Center for Disease Control and Prevention, National Center for Health Statistics). Kuperberg, Arielle. 2014. “Age at co-residence, premarital cohabitation, and marriage dissolution: 1985–2009.” Journal of Marriage and Family 76(2):352–369.

Lyness, Karen S., Janet C. Gornick, Pamela Stone, and Angela R. Grotto. 2012. “It’s all about control: Worker control over schedule and hours in cross-national context.” American Sociological Review 77(6):1023–1049. Mann, Michael. 2005. The Dark Side of Democracy: Explaining Ethnic Cleansing. New York: Cambridge University Press. McRoberts, Kenneth. 2001. “Canada and the Multinational State.” Canadian Journal of Political Science / Revue Canadienne de Science Politique 34:683–713. Rusche, Georg and Otto Kirchheimer. 1939. Punishment and Social Structure. New York: Russell and Russell. Sutton, John. 2013. “The transformation of prison regimes in latecapitalist societies.” American Journal of Sociology 119 (3):715–746.

Chapter 10 References

Chapter 12 References

Crawley, Elaine M. 2004. “Emotion and performance: Prison officers and the presentation of self in prisons.” Punishment and Society 6(4):411–427.

Becker, Howard S. et al. 1961. Boys in White: Student Culture in Medical School. Chicago: University of Chicago Press.

Cushman, Penni. 2005. “It’s Just Not a Real Bloke’s Job: Male teachers in primary schools.” Asia-Pacific Journal of Teacher Education 33:321–338.

Hartley, James, James Pennebaker, and Claire Fox. 2003. “Using New Technology to Assess the Academic Writing Styles of Male and Female Pairs and Individuals.” Journal of Technical Writing and Communication 33(3):243–261.

Davis, Fred. 1959. “The cabdriver and his fare: facets of a fleeting relationship.” American Journal of Sociology 65:158–165. Fine, Gary Alan. 1987. With the Boys: Little League Baseball and Preadolescent Culture. Chicago: University of Chicago Press.

Hurd, P. D. 1998. “Scientific Literacy: New Minds for a Changing World.” Science Education 82:407–416.

Gurney, Joan Neff. 1985. “Not one of the guys: The female researcher in a male-dominated setting.” Qualitative Sociology 8(1):42–62.

National Council of Teachers of English. 2002. “Guidelines for Gender-Fair Use of Language,” http://www.ncte.org/positions/ statements/genderfairuseoflang (accessed March 24, 2015).

Hochschild, Arlie. 1983. The Managed Heart: Commercialization of Human Feeling. Berkeley: University of California Press.

Rodriquez, Jason. 2011. “‘It’s a dignity thing’: Nursing home care workers’ use of emotions.” Sociological Forum 26(2):265–286.

Lopez, Steven Henry. 2007. “Efficiency and the fix revisited: Informal relations and mock routinization in a nonprofit nursing home.” Qualitative Sociology 30:225–247.

Weber, Jean Hollis. 2012. “Gender-neutral Technical Writing,” http:// techwhirl.com/gender-neutral-technical-writing/ (accessed March 24, 2015).

Nolan, Kathleen. 2011. Police in the Hallways: Discipline in an Urban High School. Minneapolis MN: University of Minnesota Press.

Weitzer, Ronald and Steven Tuch. 2005. “Racially Biased Policing Determinants of Citizen Perceptions.” Social Forces 83(3):1009–1030.

Pattillo, Mary. 1999. Black Picket Fences Privilege and Peril among the Black Middle Class. Chicago: University of Chicago Press.

Whyte, William Foote. 1955. Street Corner Society: The Social Structure of an Italian Slum. Chicago: University of Chicago Press.

Credits Photographs: Cover Image Shutterstock Chapter 1 Chapter opener p. 1 Alinute/Fotolia; p. 4 Angie D’Amico/Shutterstock; p. 4 Venus Angel/Shutterstock; p. 7 Stock Photo/123RF; p. 9 Joy Stein/Shutterstock; p. 12 Tudor Photography/ Pearson Education Ltd.; p. 13 bikeriderlondon/Shutterstock; p. 14 Rachel Epstein/PhotoEdit; p. 17 Robert Brenner/PhotoEdit. Chapter 2 Chapter opener p. 21 Marie Killen/Moment/Getty Images; p. 23 bonciutoma/Fotolia; p. 25 James Hardy/PhotoAlto Agency RF Collections/Getty Images; p. 26 Bill Aron/PhotoEdit; p. 43 Scott L. Williams/Shutterstock. Chapter 3 Chapter opener p. 50 National Archives; p. 54 Scherl/ Sueddeutsche Zeitung Photo/The Image Works; p. 55 Duke Downey/ San Francisco Chronicle/Corbis; p. 56 The Drs. Nicholas and Dorothy Cummings Center for the History of Psychology, The University of Akron; p. 57 Henryk Sadura/Shutterstock; p. 61 Fotolia. Chapter 4 Chapter opener p. 70 Fotolia; p. 73 Popperfoto/Getty Images; p. 74 JOSEPH EID/AFP/Getty Images; p. 79 Dmitry Sosenushkin/Fotolia; p. 83 tupungato/123RF; p. 83 Bill Aron/PhotoEdit; p. 83 David Bacon/The Image Works. Chapter 5 Chapter opener p. 92 Mike Kemp/Getty Images; p. 93 Cleve Bryant/PhotoEdit; p. 99 John Moore/Getty Images; p. 100 S ashkin/ Fotolia; p. 102 Ed Kashi/Corbis; p. 105; p. 111 Africa Studio/Fotolia. Chapter 6 Chapter opener p. 116 Lisa F. Young/Fotolia; p. 131 Kenishirotie/Shutterstock; p. 134 Bonnie Kamin/PhotoEdit; p. 135 Michael Jung/Fotolia. Chapter 7 Chapter opener p. 138 Nicky Larson/Fotolia; p. 142 Thinkstock Images/Stockbyte/Getty Images; p. 142 Fotolia; p. 146 Jacobs Stock Photography/Getty Images; p. 151 Alex Milan Tracy/ Demotix/Corbis; p. 155 Fotolia. Chapter 8 Chapter opener p. 160 Fotolia; p. 162 Agencja Fotograficzna Caro/Alamy Stock Photo; p. 166 Ingo Jezierski/Photodisc/ Getty Images; p. 166 Donna Beeler/Shutterstock; p. 167 Fotolia; p. 171 Fotolia. Chapter 9 Chapter opener p. 182 Andrey Popov/123RF; p. 187 SW Productions/Corbis; p. 198 Monty Rakusen/Cultura/Getty Images; p. 205 Stokkete/Shutterstock. Chapter 10 Chapter opener p. 210 Alamy Stock Photo; p. 213 Jupiterimages/Stockbyte/Getty Images; p. 213 Ariel Skelley/Blend Images/Alamy Stock Photo; p. 214 Robert Kneschke/Shutterstock; p. 218 123RF; p. 231 Christina Kennedy/PhotoEdit. Chapter 11 Chapter opener p. 236 Dennis MacDonald/PhotoEdit; p. 239 World History Archive/Alamy Stock Photo; p. 241 Bettmann/AS400 DB/Corbis; p. 248 Bettmann/AS400 DB/Corbis; p. 251 Alex Wong/Getty Images; p. 253 Bettmann/AS400 DB/Corbis; p. 253 Bettmann/AS400 DB/Corbis; p. 252 Gari Wyn Williams/Alamy Live News; p. 255 Jim Hughes/Shutterstock; p. 259 Jim Hughes/ Shutterstock. Chapter 12 Chapter opener p. 261 Wavebreakmedia/Shutterstock; p. 264 WavebreakmediaMicro/Fotolia; p. 271 akg-images/ Newscom; p. 276 adrianhancu/123RF; p. 277 Monkey Business Images/Shutterstock.

292

Text Credits

Chapter 1 Figure 1-1 on p. 2: Lawrence W. Neuman; Figure 1-2 on p. 2: Lawrence W. Neuman; Excerpt on p. 4: From “The evidence underpinning sports performance products: a systematic assessment” by Carl Heneghan, Jeremy Howick, and Braden O’Neill, © 2012 BMJ Publishing Group; Figure 1-3 on p. 14: Youth illicit drug use prevention: Dare long-term evaluations and federal efforts to identify effective programs, January 15, 2003 United States General Accounting Office; Figure 1-4 on p. 17: Lawrence W. Neuman; Figure 1-5 on p. 18: Based on Neuman, W. Lawrence 2007. Basics of Social Research, 2nd ed. Allyn & Bacon; Excerpt on p. 18: From “Seeking Facts, Justices Settle for What Briefs Tell Them” by Adam Liptak, © September 1, 2014 The New York Times. Chapter 2 Excerpt on p. 23: Source: Isaac Newton; Figure 2-1 on p. 24: Fotolia; Figure 2-2 on p. 28: Lawrence W. Neuman; Excerpt on p. 29: From “Time for the U.S. to Reskill? What The Survey of Adult Skills Says” © 2013 © Organisation for Economic Co-operation and Development (OECD); Figure 2-4 on p. 31: Lawrence W. Neuman; Excerpt on p. 34: Lawrence W. Neuman; Figure 2-5 on p. 33: © EBSCO Publishing, Inc. 2015 All rights reserved; Excerpt on p. 35: Lawrence W. Neuman. Chapter 3 Excerpt on p. 56: From “Fieldwork on the Beat.” Varieties of Qualitative Research. Edited by J. Van Mannen, J. Dabbs, Jr., and R. Raulkner, © 1982 Sage Publication; Excerpt on p. 53: Lawrence W. Neuman; Excerpt on p. 56: From the Journal of Abnormal And Social Psychology, Vol. 67, No. 4, 1963, Copyright, 1963, American Psychological Association, Inc.; Excerpt on p. 56: From Experiencing Fieldwork by John Van Maane, © 1991, Sage Publications, Inc.; Table 3-1 on p. 60: Lawrence W. Neuman; Excerpt on p. 60: Lawrence W. Neuman; Excerpt on p. 68: Lawrence W. Neuman; Excerpt on p. 61: From “Cracking the Corn fields.” Sociological Quarterly by Draus, Paul J., Harvey Siegal, Rober Carlson, Russel Falck, and Jichuan Wang, Jichuan, © 2005 John Wiley and Sons, Inc.; Excerpt on p. 56: “Learner Demands to be Shocked” from Obedience to Authority: An Experimental View by Stanley Milgram, © 1974 HarperCollins Publishers and Pinter & Martin, Ltd.; Excerpt on p. 63: From American Association of Public Opinion Research of Code of Professional Ethics. Copyright © 2005. Reprinted by permission of American Association of Public Opinion Research; Excerpt on p. 66: From Medical marijuana 2010: it’s time to fix the regulatory vacuum. J Law Med Ethics; 38(3):654–666, © 2010 American Society of Law, Medicine, & Ethics; Excerpt on p. 66: From Blurred Boundaries: The Therapeutics and Politics of Medical Marijuana, Mayo Clin Proc. 87(2): 172–186, 2012 Feb, © Elsevier Inc.; Excerpt on p. 65: From “Compromised Police Legitimacy as a Predictor of Violent Crime in Structurally Disadvantaged Communities.” by Kane, Robert J, © copyright 2005 John Wiley and Sons; Excerpt on p. 68: From “Good Faith, Bad Ethics: When Scholars Go the Distance and Scholarly Associations Do Not” Law & Social Inquiry 24:977–986 by Rik Scarce, © 1999 John Wiley and Sons Inc. Chapter 4 Figure 4-4 on p. 79: Dmitry Sosenushkin/Fotolia. Chapter 5 Figure 5-3 on p. 101: Lawrence W. Neuman; Excerpt on p. 97: From Public Attitudes towards hom*osexuality by Tom W. Smith © 2011 NORC University of Chicago; Excerpt on p. 102: From The Contact Hypothesis and Attitudes Toward Latinos in the United States by Christopher G. Ellison, Heeju Shin, and David L. Leal © Southwestern Social Science Association 92(4):938–58; Figure 5-3a on

Credits 293 p. 105: Shutterstock; Figure 5-3b on p. 105: Ratan Mani Banerjee/Pearson India Education Services Pvt. Ltd.; Figure 5-3c on p. 105: Arvind Singh Negi/Red Reef Design Studio/Pearson India Education Services Pvt. Ltd.; Figure 5-3d on p. 105:123RF; Figure 5-3e on p. 105: HighlyAnimated Studios/Shutterstock; Figure 5-3f on p. 105: Axelberg/ Shutterstock; Figure 5-3g on p. 105: Georgios Kollidas/Shutterstock; Figure 5-3h on p. 105:Mallinka1/Shutterstock; Figure 5-3i on p. 105: Lovely Kukreja/Pearson India Education Services Pvt. Ltd.; Figure 5-3j on p. 105 :Danish Zaidi/ Pearson India Education Services Pvt. Ltd.; Figure 5-3k on p. 105:Lorelyn Medina/Fotolia; Figure 5-3l on p. 105: Ayupovaguzel/Fotolia; Figure 5-3m on p. 105: Steinar/Fotolia; Figure 5-4 on p. 107: From New Racial Segregation Measures for Large Metropolitan Areas: Analysis of the 1990–2010 Decennial Censuses © University of Michigan, Population Studies Center. Chapter 6 Excerpt on p. 128: From The ANES Guide to Public Opinion and Electoral Behavior, Table 6A.2. Copyright © by American National Election Studies. Reprinted by permission of The American National Election Studies; Figure 6-1 on p. 118: From “Political Survey” Pew Research Center for the People & the Press Final Topline, © May 2013 Pew Research Center; Figure 6-2 on p. 118: Data from Pew Research Center; Table 6-1 on p. 119: Lawrence W. Neuman; Table 6-3 on p. 124: From How unclear terms affect survey data, Volume 56, Issue 2 (Public Opinion Quarterly) © 1992 Oxford University Press; Excerpt on p. 125: Lawrence W. Neuman; Excerpt on p. 126: From Assessing the Representativeness of Public Opinion Surveys, © May 15, 2009 Pew Research Center; Table 6-6 on p. 127: Based on Schuman and Presser (1981:116–125). Standard format is from Fall 1978; quasi- and full-filter are from February 1977; Excerpt on p. 129: From The Right-to-Die Debate and the Tenth Anniversary of Oregon’s Death with Dignity Act © 2007 Pew Research Center; Figure 6-1 on p. 118: From General Social Surveys, © 1972–2006 National Opinion Research Center; Figure 6-4 on p. 131: Lawrence W. Neuman. Chapter 7 Figure 7-7 on p. 148: From “Binds and bounds of communion: Effects of interpersonal values on assumed similarity of self and others” in Journal of Personality and Social Psychology 103 (5):879– 897, © by American Psychological Association (APA); Excerpt on p. 156: From Child and adolescent fast-food choice and the influence of calorie labeling: a natural experiment by B Elbel, J Gyamfi and R Kersh 35, 493–500 © 2011 Nature Publishing Group. Chapter 8 Figure 8-2 on p. 170: Lawrence W. Neuman; Excerpt on p. 164: Lawrence W. Neuman; Excerpt on p. 166: Lawrence W. Neuman; Figure 8-2 on p. 169: Lawrence W. Neuman; Excerpt on p. 171: Lawrence W. Neuman; Excerpt on p. 174: Lawrence W. Neuman; Excerpt on p. 179: American Sociological Association 2013. Chapter 9 Figure 9-1 on p. 183: “First premarital cohabitation in the United States, 2006-2010 National Survey of Family Growth" by Casey E. Copen, Kimberly Daniels and William D. Mosherin in National Health Statistics Reports no. 64, April 4, 2013. Centers for Disease Control and Prevention (CDC); Figure 9-4 on p. 187: From ISSP 2012 Family and Changing Gender Roles IV: Questionnaire Development in GESIS-Technical Reports by Evi Scholz, Regina Jutz, Jonas Edlund, Ida Öun and Michael Braun, © by GESIS LeibnizInstitute for the Social Sciences; Figure 9-7 on p. 192: “Births: Final data for 2013” in National Vital Statistics Reports by Joyce A. Martin, Brady E. Hamilton, Michelle J.K. Osterman, Sally C. Curtin, and T.J. Mathews, Volume 64, Number 1, January 15, 2015, Centers for Disease Control

and Prevention (CDC); Figure 9-8a on p. 193: Lawrence W. Neuman; Figure 9-8b on p. 193: Lawrence W. Neuman; Table 9-2 on p. 196: From OECD - Social Policy Division - Directorate of Employment, Labour and Social Affairs, OECD Family Database © OECD; Excerpt on p. 203: From Age at Coresidence, Premarital Cohabitation, and Marriage Dissolution: 1985–2009 in Journal of Marriage and Family 76(2), 352–369 by Arielle Kuperberg, © National Council on Family R elations & Wiley; Table 9-12 on p. 204: Lawrence W. Neuman; Figure 9-9a on p. 195: Lawrence W. Neuman; Figure 9-9b on p. 195: Lawrence W. Neuman; Figure 9-10a & b on p. 195: Lawrence W. Neuman. Chapter 10 Excerpt on p. 213: Lawrence W. Neuman; Excerpt on p. 215: © Springer International Publishing AG; Excerpt on p. 217: Lawrence W. Neuman; Excerpt on p. 217: Lawrence W. Neuman; Excerpt on p. 219: From Kitchens: The Culture of Restaurant Work by Gary Alan Fine, © 1996 University of California Press; Excerpt on p. 223: From “Classic Acts: Service and Inequality in Luxury Hotels. Berkeley” by Rachel Sherman, pp. 86–97, © 2006 University of California Press; Excerpt on p. 224: From The Urban Ethnography Reader by Mitchell Duneier, Philip Kasinitz, Alexandra Murphy, © 2014 Oxford University Press. Chapter 11 Excerpt on p. 237: From “The Transformation of America’s Penal Order: A Historicized Political Sociology of Punishment” by Michael Campbell, Heather Schoenfeld in American Journal of Sociology, Vol. 118, No. 5 (March 2013), pp. 1375–1423, © 2013 University of Chicago Press; Figure 11-1 on p. 237: From The Sentencing Project. Reprinted by permission of The Sentencing Project; Excerpt on p. 238: From “The Transformation of Prison Regimes in Late Capitalist Societies” by John R. Sutton in American Journal of Sociology, Vol. 119, No. 3 (November 2013), pp. 715–746, © 2013 University of Chicago Press; Excerpt on p. 238: From “The Transformation of Prison Regimes in Late Capitalist Societies” by John R. Sutton in American Journal of Sociology, Vol. 119, No. 3 (November 2013), pp. 715–746, © 2013 University of Chicago Press; Figure 11-2 on p. 246: The sun (New York), 02 Jan. 1908. Chronicling America: Historic American Newspapers. Library of Congress; Excerpt on p. 248: From Women of the Klan: Racism and Gender in the 1920s by Kathleen M. Blee, © 2008 University of California Press; Excerpt on p. 250: Lawrence W. Neuman; Excerpt on p. 252 Efficiency and the fix revisited: Informal relations and mock routinization in a nonprofit nursing home” by Steven Lopez in Qualitative Sociology 30:225–247, © 2007 Springer International Publishing AG; Excerpt on p. 252: From Immigrants at the Margins: Law, Race, and Exclusion in Southern Europe by Kitty Calavita, © 2005 Cambridge University Press; Excerpt on p. 252: From Immigrants at the Margins: Law, Race, and Exclusion in Southern Europe by Kitty Calavita, © 2005 Cambridge University Press; Excerpt on p. 253: Sir Edward B. Tylor; Excerpt on p. 255: Lawrence W. Neuman; Excerpt on p. 255: From “It’s All about Control: Worker Control over Schedule and Hours in Cross-National Context” by Karen S. Lyness, Janet C. Gornick, Pamela Stone, Angela Grottoa in American Sociological Review, © 2012 American Sociological Association; Figure 11-4 on p. 255: Adapted from Figure 1 Lyness, et al., 2012, p. 1092. Chapter 12 Excerpt on p. 274: From Spencer Foundation website “The New Civics Small Grant Guidelines”. Copyright © 2015. Reprinted by permission of Spencer Foundation; Figure 12-2 on p. 274: Lawrence W. Neuman; Figure 12-3 on p. 275: Law & Social Sciences (LSS), The National Science Foundation.

Index A Abstracts, 30–31, 268 Accidental sampling. See Convenience sampling Adachi-Mejia, A. M., 3 Adamchak, D., 162 Albergotti, R., 58 Alternative hypotheses, 42 American Association for Public Opinion Research, ethics code of, 62–64 Analysis content. See Content analysis level of, 45–46 multiple regression, 203–204 units of, 45, 96, 164 Analytic notes, 227 Anderson, E., 12 Angelini, J. R., 1 Anonymity, 60 Appelbaum, P. S., 58 Applied research, 16–17 results, decision makers and, 17–18 Article search tools, 30 number of articles found with, 31 Association, variables measure of, 200–201 Atkinson, M., 22 Attitude of strangeness, 221–222 Authenticity, 103 Ayers, J. W., 110 Azrael, D., 70 B Babbie, E., 101 Baik, Kyoung-Deok, 147 Baker, T., 103 Basic social research, 15–16 Batalova, J., 194, 195 Becker, H. S., 272 Bellandi, R., 103 Bensman, J., 60 Bergamini, E., 3 Bernhardt, A. M., 3 Bivariate statistics, 186, 194–204 cross-tabulation, 197–200 measure of association, 200–201 scattergrams, 194–197 Blee, K., 248 Blind review, 26 Books finding studies in, 31 as research report source, 28 Borgadus social distance scale, 110 Borzekowski, D., 2 Bos, H. M. W., 103 Bostwick, M., 66 Brady, D., 179 Brown, H., 164 Brown, J. H., 14 Browning, S. E., 82 Burton, C., 5 C Calavita, K., 252 Campbell, M. C., 237

294

Caplan, J., 22 Carlson, R., 86 Cases, qualitative data, 43 CATI. See Computer-assisted telephone interviewing (CATI) Causal explanation, 40, 44 Cause-effect relationships experiments and, 139–140 writing, 267–268 Census, 71 Chavez, L., 166 Chen, J., 96, 100 Cheng, S., 17 Citations, 35, 36 Classical experimental design, 144 Cleveland, L., 67 Closed-ended response format, 124–125 Cluster sampling, 83–85 Codebook, 184–186 Code of ethics, 62 Coding system, content analysis, 164 latent, 165 manifest, 165 Coercion, 59 Cohen, P. N., 194, 195 Collier, K. L., 103 Comparative field research, 253 Computer-assisted telephone interviewing (CATI), 132, 135–136 Conceptual equivalence, 256–257 Conceptual hypothesis, 97 Conceptualization, 95–99 qualitative, 98–99 quantitative, 97–98 Confidence interval, 88–89 Confidentiality, 60–61, 133 Contact hypothesis, 102–103 Content analysis limitations of, 169–170 measure/coding in, 163–165 as quantitative data collection technique, 10 research steps, 167–169 with visual material, 166 Content files, creation of, 32 Content validity, 102 Contexts, qualitative data, 43 Contextual equivalence, 257 Contingency question, 128 Continuous variable, 103 Control group, 143 Control variables, 46, 119, 201–202, 268 Convenience sampling, 71–72 Conway, E. M., 67 Copen, C. E., 182, 183 Correlational, 119 Covert field research, 59 Craig, T., 147 Crawford, C., 66 Crawley, E. M., 211 Creek, S., 13 Criterion validity, 102

Critical thinking features of, 8–9 meaning of, 8 Cross-national surveys, 254 Cross-tabulation, 197–200 Cultures, and historical-comparative research, 250–257 Cushman, P., 231 D Daniels, K., 182, 183 Data comparative, 253–254 field, collection of, 222–229 qualitative. See Qualitative data quantitative. See Quantitative data raw, organizing into machine-readable format, 183–184 standardization of, 176–177 for statistical analysis, 183–186 systematic methods to collect, 95 Data coding, 183 Data records, 183–184 Davis, D., 135 Davis, Fred, 219 Debrief, 155 Deception, 58–59 Decision making, research and, 3–6 Deductive approach, 37 DeMello, M., 22, 36 Dependent variable, 142, 267 Descriptive research, 12 Descriptive statistics, 186 bivariate statistics. See Bivariate statistics measures of central tendency in, 187 univariate statistics. See Univariate statistics Design notation, 151 Desmond, M., 98 de Vries, K. M., 13 Diagrams, in field notes, 226–227 Direct observation notes, 227 Discrete variable, 103 Discrimination, defined, 95, 96 Dissertations, as research report source, 28 Donoghue, C., 96 Donohew, L., 151 Double-barreled question, 122 Double-blind experiment, 153 Draus, P. J., 86 Drigotas, S., 5 Dukes, R. A., 22, 36 Dula, C. S., 22 Duncan, B., 107 Duncan, O. D., 107 Dwoskin, E., 58 E Ecological fallacy, 175–176 Editing, and rewriting reports, 266 Elbel, B., 3, 155 Ellison, C. G., 102 Empirical evidence, 9–10

Index 295 Empirical hypothesis, 97 Empirical social research, 6–7 Equivalence conceptual, 256–257 contextual, 257 lexicon, 256 measurement, 257 Equivalent time series, 149 Ethics, 51–52 and experimental research, 157 in field research, 230–231 in historical-comparative research, 258 and humans as research participants, 53–64 in nonreactive research, 180 sponsors and, 64–66 and survey research, 136 Ethnocentricism, 247 Ethnographic field research, as qualitative data collection technique, 11 Ethnography, 211–214 Evaluation research, 13–15 Evidence empirical, 9–10 historical, 245–249 physical. See Physical evidence Executive summary, 269 Exhaustive variable, 106 Existing qualitative data, 253–254 Existing statistical sources, 170–177 limitations, 173–176 as quantitative data collection technique, 10–11 Expectancy, experimenter, 153–154 Experimental designs, 144–151 comparison of, 150 notation, 151 pre-experimental designs, 148–149 quasi-experimental designs, 149 true, 144–148 classical experimental design, 144 factorial design, 146–147 latin square design, 146 solomon four-group design, 145 two-group posttest-only design, 145 Experimental group, 143 Experimental validity, 152–154 external, 154 internal, 152–154 Experiments, 138–157 and cause-effect relationships, 139–140 design. See Experimental designs double-blind, 153 and ethics, 157 field, 154 limitations of, 140 mortality, 152–153 natural, 155 planning, 143 posttest, 143 pretest, 143 as quantitative data collection technique, 10 random assignment in, 140–142 repeated measures design, 143 research questions appropriate for, 140 results, comparisons of, 156–157 steps in conducting, 143–144 validity. See Experimental validity variables in, 142–143

Explanatory research, 12–13, 44 Explicit knowledge, 213 Exploratory research, 11–12 External validity, 154, 267 F Facebook, unethical research by, 58 Face-to-face interviews, 132 Face validity, 102 Factorial experimental design, 146–147 Falck, R., 86 Fallacy of misplaced concreteness, 175 Farkas, G., 22 Ferree, M. M., 250 Field data, collection of, 222–229 Field experiment, 154 Field notes, 222–227 levels of, 227 maps/diagrams in, 226–227 recommendations for taking, 225–226 Field research, 211 comparative, 253 ethics in, 230–231 ethnography, 211–214 exiting process, 229–230 and historical-comparative research, 239–240 interviewing in, 227–229 naturalism, 214 preparation for, 216 project, parts of, 216–218 rapport in, 219–220 report, 230, 271–273 sampling, 224–225 site, 216–217 stages in, 214–216 success in, strategies for, 220–222 Field site, 216–217 Fine, G. A., 218, 219 Finnigan, R., 179 Focus groups, 231–232 Foddy, W., 128 Foster, G., 162 Fox, C., 262 Freewriting, 265 Frendreis, J., 173 Frequency distribution, 186 normal, 188 skewed, 188 Full-filter question, 126 G Galilei, G., 7 Galton's problem, 252–253 Gamson, W., 250 Gantz, W., 1 Gatekeeper, in field research, 217 Gee, G. C., 96, 100 General Social Survey (GSS), 178 Gerhards, J., 250 Gertz, M., 103 Gillory, J., 58 Goel, V., 58 Gohil, K., 147 Gornick, J. C., 254, 255 Gorski, D., 58 Government reports, as research report source, 28 Grotto, A. R., 254, 255 Grounded theory, 45

GSS. See General Social Survey (GSS) Gurney, J. N., 218 Gusrang, J. L., 17 Gutting, G., 66 Guttman scale, 111–113 Gyamfi, J., 3, 155 H Han, B., 13 Hanco*ck, J., 58 Haphazard sampling. See Convenience sampling Hartley, J., 262 Hawkes, D., 22 Hawthorne effect, 154–155 H-C research. See Historical-comparative (H-C) research Heise, D. R., 110 Hemenway, D., 70 Heneghan, C., 4 Heywood, W., 22 Hicks, J., 5 Hidden populations, sampling of, 86–87 Historical-comparative (H-C) research, 238–239 and cultures, 250–257 ethics in, 258 features of, 240–242 field research and, 239–240 issue of equivalence, 256–257 as qualitative data collection technique, 11 report on, 273 stages of, 243–244 Historical evidence, 245–249 narrative history, 249 primary sources, 245, 247 running records, 247–248 secondary sources, 248–249 Hochschild, Arlie, 211 Hofstetter, C. R., 110 Horne, J., 22 Hoyle, R., 151 Humans, as research participants anonymity, 60 confidentiality, 60–61 ethical issues involved in, 53–64 formal protections for, 62 informed consent and, 57–58 origin of ethical principles for, 53–54 privacy, 59–60 protection from harms, 54–57 special populations, protections for, 61 Hummel, R., 162 Humphreys, L., 56, 57 Hurd, P. D., 262 Hurwitz, J., 129 Hypothesis, 41–42, 122–123, 267 alternative, 42 conceptual, 97 contact, 102–103 empirical, 97 null, 42 I Independent variable, 40, 142, 267 Indexes, 105–106 construction of, 106–108 Indicators multiple, 100 social, 171

296 Index Inductive approach, 37, 93 Informant, 227 Informed consent, 55, 57–58, 133 Institutional Review Board (IRB), 62, 157, 217 Interaction effects, in factorial experimental design, 146–147 Intercoder reliability, 164, 165 Internal validity, 152–154, 201, 267 Internet, and research, 35–36 Interrupted time series, 149 Intervening variable, 40 Interviewer bias, 135 role of, 133 training, 134 Interviews face-to-face, 132 in field research, 227–229 life history, 249 probes in, 134–135 schedule, 121 stages, 133–134 in survey research, 132–136 survey vs. field research, 228 telephone, 132 In-text citation, 35 IRB. See Institutional Review Board (IRB) ISSP (International Social Survey Program), 179 J Jamieson, C., 66 Jeffreys, A., 16 Johnson, R. M., 70 Jon, Jae-Eun, 103 Jones, K., 22 Jotted notes, 227 Journals, scholarly, 25–26 articles, reading, 32 locating, 26–27 studies in, locating/evaluating, 26, 30 Judd, C. M., 138 Judgment sampling, 74 K Kang, M., 22 Keil, F. C., 6 Kersh, R., 3, 155 King, L. A., 5 King, R., 67 Kirk, David S., 84 Kirt, David, 65 Kleg, M., 96 Klitzman, Robert, 58 Knox, D., 22 Kolody, B., 110 Kraemer, H., 2 Kramera, A., 58 Kreft, I. G. G., 14 Kuperberg, A., 203 L Latent coding, 165 Latin square experimental design, 146 Leading question, 123–124 Leal, D. L., 102 Ledford, H., 66 Legal harm, 56–57 Level of statistical significance, 204–207

Levey, T., 12 Levin, S., 103 Lexicon equivalence, 256 Life history interview, 249 Likert scale, 108–109, 118 Linear path, research, 39 Literary Digest, 76 Literature review, 23–24. See also Journals, scholarly; Periodicals procedural steps. See Procedural steps, literature review search plan, 24 Locke, K. D., 147 Lopez, S. H., 215 Lorch, E. P., 151 Lyness, K. S., 254, 255 M Macro level, of analysis, 45, 46 Mail questionnaire, 131 Main effects, in factorial experimental design, 146 Manifest coding, 165 Mann, M., 239 Maps, in field notes, 226–227 Marijnissen, J., 3 Martin, J. A., 22 McRoberts, K., 251 Mean, 188, 189 Measurement, 93 poverty, 93–94 process, parts of, 95–96 qualitative, principles of, 103 quantitative, principles of, 103–105 in social world, 94–95 validity. See Validity variation, 189–190 Measurement equivalence, 257 Median, 188, 190 Metcalfe, C., 103 Meyer, R., 58 Micro level, of analysis, 45–46 Milgram, S., 56 Milgram obedience study, 55, 56 Miller, M., 70 Mills, C. M., 6 Mode, 187 Mooney, C., 67 Mosher, W., 182, 183 Multiple indicators, 100 Multiple regression, 203–204 Multivariate statistics, 186 Murphy, C., 161 Mutheson, D., 2 Mutually exclusive variable, 106 N Narrative history, 249 Natural experiments, 155 Naturalism, 214 Newman, Katherine S., 74 Nolan, K., 224 Nonlinear path, research, 39 Nonreactive research, 160–180 content analysis. See Content analysis ethics and, 180 limitations, 161, 179 physical evidence in, usage of, 161–163 quantitative techniques, 161 strengths of, 179

Normal distribution, 188 Normalize, in field research, 219 Notes recording, 32–34 Null hypothesis, 42 Nuremburg Code, 62 O Objective, in research, 67–68 Observation, in field research, 222–227 One-group pretest-posttest experimental design, 148–149 One-shot case study experimental design, 148 Open-ended response format, 124 Operational definition, 96 Operationalization, 96–99, 120 Opinion poll vs. social surveys, 117 Oral history, 248 Order effects, 130–131 Oreskes, N., 67 Organization for Economic Co-operation and Development (OECD), 29 Orientation reading, 243 Otten, J. J., 3 Oversampling, 82 P Palmgreen, P., 151 Papachristos, Andrew, 65 Papachristos, Andrew V., 84 Paraphrase, 265 Parenthetical citation, 35 Parrillo, V. N., 96 Patrick, K., 22 Pattillo, M., 228 Paul, Anju Mary, 74 Payne, B. K., 142 Peer review, 25–26 Peffley, M., 129 Pennebaker, J., 262 Percentiles, 190 Periodicals defined, 25 scholarly journals. See Journals, scholarly types of, 25, 27 Personal notes, 227 Physical appearance, and social interactions, 223–225 Physical evidence limitations of, 162 and nonreactive research, 161–163 Physical harm, in social research studies, 55 Pickett, J. T., 103 Piliavin, I., 55 Piliavin, J., 55 Placebo, 154 Plagiarism, 52, 265 Policy reports, as research report source, 28–29 Politics, and social research, 66–67 Population, 71, 75 hidden, sampling of, 86–87 parameter, 75, 88 special, protections for, 61 target, 75, 88 Population parameter, 204 Posttest-only nonequivalent group design, 149

Index 297 Poverty, measurement, 93–94 Pre-experimental designs, 148–149 one-group pretest-posttest design, 148–149 one-shot case study design, 148 posttest-only nonequivalent group design, 149 Presented papers, as research report source, 28 Presentism, 247 Presser, S., 127, 128 Prewriting, 265 Primary sources, historical evidence, 245, 247 Principle of voluntary participation, 53 Privacy, 59–60 Probes, in interviews, 134–135 Procedural steps, literature review, 29–37 notes taking, 32–34 reference list, creation of, 35–36 research reports, locating, 30–31 search, designing, 30 topics, refining, 29 writing review, 34–35 Program for International Student Assessment (PISA), 29 Purposive sampling, 73–74 Q Qualitative data, 10 cases, 43 collection techniques, 11. See also specific techniques contexts, 43 existing, 253–254 patterns in, 43–44 research proposal for, 274 Quantitative data, 10 collection techniques, 10–11. See also specific techniques cross-national, 254 patterns in, 43–44 research proposal for, 274 variables, 40 Quasi-experimental designs, 149 Quasi-filter question, 126 Questions/questionnaire, 121–129 in field interviews, 228–229 full-filter, 126 leading, 123–124 length of, 129 order effects, 130–131 quasi-filter, 126 sequence, 129–130 standard-format, 126 Quota sampling, 72–73 R Random assignment, 140–142 Random-digit dialing (RDD), 85–86, 118, 132 Random samples, 71 features of, 76 procedural steps to produce, 76–80 stratified, 82 systematic, 80–81 terminology, 75–76 Random selection process, 71 Range, in statistics, 189 Rathje, W., 161 RDD. See Random-digit dialing (RDD) Recollections, 247–248

Reference list, creation of, 35–36 Reliability, 100 and validity, 102 existing statistics research, 173–174 intercoder, 164, 165 Religion, research and, 7 Repeated measures design, 143 Research covert field, 59 deception in , 58–59 and decision making, 3–6 descriptive, 12 evaluation, 13–15 explanatory, 12–13, 44 exploratory, 11–12 frustrations/misunderstandings about, 5–6 Internet and, 35–36 linear path, 39 nonlinear path, 39 nonreactive. See Nonreactive research procedural steps, 18 proposal. See Research proposal purpose of, 11–15 question. See Research question and religion, 7 reports, 32–34, 261–277 usage of, 15–18 Researcher inference notes, 227 Research fraud, 52 Research grants, research proposal for, 273–277 Research proposal defined, 38 designing, 38–47 preparing, 273–277 for research grants, 273–277 sample, 279–283 study design issues in, 46 Research question, 23, 29, 37, 38–39 good vs. not-so-good, 38 narrowing topic into, 37–38 Research reports, 32–34, 261–277 abstract, 268 executive summary, 269 on field research, 271–272 historical-comparative research, 273 literature review and, 265 need for, 262 organization of thoughts, 264–265 prewriting activities, 265 qualitative, 270–273 quantitative, 268–270 rewriting activities, 266–267 style, 263 tone, 263–264 writing process, 262–267 Resenhoeft, A., 22 Revising, and rewriting reports, 266 Richters, M. K. J., 22 Rideout, V., 1 Robinson, T. N., 2 Rodin, J., 55 Rodriquez, J., 215, 272 Roksa, J., 12 Rosenberg, Morris, 109 Rucht, D., 250 S Sample/sampling, 70–89 cluster, 83–85

in content analysis, 167–168 convenience, 71–72 in difficult/specialized situations, 85–87 distribution, 77–79, 88–89 error, 76, 87–88 field research, 224–225 frame, 75 interval, 80 judgment, 74 need for, 71 purposive, 73–74 quota, 72–73 random. See Random samples ratio, 75, 88 research proposal, 279–283 size, 88 snowball, 74–75 within-housing, 86 Sampling element, 75 Sandfort, T. G. M., 103 Sargent, J. D., 3 Savelsberg, J., 67 Scales, 105–106 Borgadus social distance, 110 construction, 108 Guttman, 111–113 Likert, 108–109, 118 semantic differential, 110–111 Scarce, Rik, 61, 68 Scattergrams, 194–197 Scheitle, C., 13 Schnakenberg, K., 110 Schoenfeld, H., 237 Scholarly journals. See Journals, scholarly Schuman, H., 127 Schwartz, N., 1 Scientific misconduct, 52 Secondary data analysis, 177–178 limitations of, 178–179 Secondary sources, historical evidence, 248–249 Seen, C., 22 Self-administered questionnaires, 131 Semantic differential scale, 110–111 Sherman, D. K., 138 Sherman, R., 213, 217, 222, 223, 230 Shin, H., 102 Shirley, Carla D., 92 Siegal, H., 86 Siennick, S., 22 Silver, B., 135 Silver, E., 22 Silver, S. R., 22 Simpson, J. M., 22 Sinclair, Shana, 103 Sinclair, Stacey, 103 Skerkat, D., 13 Skewed distribution, 188 Smith, A. M. A., 22 Smith, K., 3 Smith, T. W., 129 Snowball sampling, 74–75 Social desirability bias, 126, 128 Social indicators, 171 Social interactions, physical appearance and, 223–225 Social surveys, 117 opinion poll vs., 117 Solomon four-group experimental design, 145

298 Index Source files, creation of, 32 Special populations, protections for, 61 Spencer, M. S., 96, 100 Sponsors, research, and ethical concerns, 64–66 SPSS (Statistical Package for the Social Sciences), 206 Spurious, 201–202, 268 Spuriousness, 46–47 Standard deviation, 190 Standard-format question, 126 Statistical analysis, data for, 183–186 Statistical independence, 194 Statistical significance, 202, 204–207 levels of, 204–207 Stein, J. A., 22, 36 Stephenson, M., 151 Stevenson, R. W., 174 Stone, P., 254, 255 Stratified sample, 82 Stratton, H. H., 14 Stutts, M. A., 3 Style, research reports, 263 Surveys and control variables, 119 cross-national, 254 data, and cause-effect explanation, 119–120 ethics and, 136 formats, advantages and disadvantages of, 131–132 interviews in. See Interviews length of, 129 as quantitative data collection technique, 10 questions. See questions/questionnaire social. See Social surveys stages of conducting, 120–121 web, 132 Sutton, John, 236, 238 Systematic sample, 80–81 T Tacit knowledge, 213–214 Takeuchi, D., 96, 100 Tandon, P. S., 3 Target population, 75, 88 Tatalovich, R., 173 Taylor, S., 56 Tea-room trade study, 56–57

Telephone interview, 132 Thompson, M. K., 14 Thorn, C., 22 Tobler, N., 14 Tone, research reports, 263–264 Topics narrowing into research question, 37–38 refining, 29 selection of, 22–23 True experimental designs, 144–148 classical experimental design, 144 factorial design, 146–147 latin square design, 146 solomon four-group design, 145 two-group posttest-only design, 145 Tuch, S., 270 Tuskegee Study, 51 Two-group posttest-only experimental design, 145 Type I error, 207 Type II error, 207 U Underwood, E., 66 Unidimensional construct, 106 Units of analysis, 45, 96, 164 Univariate statistics, 186–193 displaying, 192–193 frequency distribution. See Frequency distribution mean, 188, 189 median, 188 mode, 187 Universe, 39, 75 Unobtrusive measures, 161 V Validity, 100 content, 102 criterion, 102 existing statistics research, 174 experiments. See Experimental validity face, 102 reliability and, 102 Value-free research, 67–68 Van Boven, L., 138 Vanderbilt, T., 172 Van Laar, C., 103 Van Maanan, J., 211

Van Maanen, J., 56 Van Noorden, R., 66 Variables continuous, 103 control, 46, 119, 201–202, 268 dependent, 142, 267 discrete, 103 exhaustive, 106 in experiments, 142–143 independent, 40, 142, 267 intervening, 40 mutually exclusive, 106 quantitative data, 40 three-variable example, 40–41 Vidich, A., 60 Villa, J., 22 Vriniotis, M., 70 W Wang, J., 86 Weber, J. H., 272 Web surveys, 132 Weitzer, R., 270 Whistle-blower, 64, 66 Whyte, W., 18 Whyte, W. F., 271 Wilking, C., 3 Williams, S., 3 Wilson, D. C., 130 Wiseman, D., 22 Within-household sampling, 86 Writer's block, 264 Wysong, E., 14 X Xie, Y., 112 Y Yamamoto, K., 96 Yong, E., 66 Z Zagumny, M. J., 14 Zank, G., 3 Zhou, X., 112 Zimbardo, P., 55 Z-scores, 191–192 Zusman, J., 22 Zusman, M., 22

William Lawrence Neuman - Understanding Research - Books A La Carte (2nd Edition) (2016, Pearson Education) - PDFCOFFEE.COM (2024)

References