Conference Proceedings 2009 Marketing & Public Policy

MAY 28–30, 2009 THE MADISON HOTEL WASHINGTON, DC 2009 MARKETING and

PUBLIC POLICY 2009 MARKETING & PUBLIC POLICY CONFERENCE PROCEEDINGS

VOLUME 19

Conference Proceedings

VOLUME 19

Contributions, Controversies, and Continuing Challenges in the Worlds of Consumer Protection and Competition

EDITORS Elizabeth S. Moore Janis K. Pappalardo William L. Wilkie Marketing and Public Policy Conference Proceedings 2009

“Contributions, Controversies, and Continuing Challenges in the Worlds of Consumer Protection and Competition”

May 28–30, 2009 The Madison Hotel Washington, D.C.

Editors: Elizabeth S. Moore Janis K. Pappalardo William L. Wilkie

Volume 19

311 South Wacker Drive, Suite 5800 • Chicago, Illinois 60606 PROGRAM COMMITTEE

Alan R. Andreasen, Georgetown University Ronald Paul Hill, Villanova University J. Craig Andrews, Marquette University Debra J. Holt, Federal Trade Commission Stacey Menzel Baker, University of Wyoming Pauline M. Ippolito, Federal Trade Commission Lauren G. Block, Baruch College John Kozup, Villanova University Paul N. Bloom, Duke University Ingrid M. Martin, California State University, Long Beach Paula Fitzgerald Bone, University of West Virginia John D Mittelstaedt, Clemson University Scot Burton, University of Arkansas Patrick E. Murphy, University of Notre Dame John E. Calfee, American Enterprise Institute Debra J. Ringold, Willamette University Les Carlson, University of Nebraska Debra L. Scammon, University of Utah Joel B. Cohen, University of Florida Clifford J. Shultz, Arizona State University Debra M. Desrochers, University of Notre Dame David E. Sprott, Washington State University Pam Scholder Ellen, Georgia State University David W. Stewart, University of California, Riverside George R. Franke, University of Alabama Madhu Viswanathan, Univ. of Illinois, Urbana–Champaign Sonya A. Grier, American University Joshua L. Wiener, Oklahoma State University Gregory T. Gundlach, University of North Florida Jerome D. Williams, University of Texas at Austin

Printed in the United States of America

Composition by Marie Steinhoff, Southeast Missouri State University Cover design by Jeanne Nemcek

ISBN: 0-87757-336-0

All rights reserved. No part of the material protected by this copyright notice may be reproduced or utilized in any form or by any means, including photocopying and recording, or by any information storage and retrieval system without the written permission of the American Marketing Association. PREFACE

The year 2009 marked the twentieth anniversary of the Marketing and Public Policy Conference (MPPC). It was celebrated on May 28–30 in the historic Madison Hotel in Washington, DC, with the theme “Contributions, Controversies, and Continuing Challenges in the Worlds of Consumer Protection and Competition.” Continuing to build on the fine tradition established by previous conferences, the 2009 conference was likely the largest conference in the twenty-year history, with 153 presentations scheduled over a wide variety of interesting, timely, and important topics. This conference is the premier event for marketing scholars and practitioners interested in topics related to social and public policy. It brings together an international roster of academics, marketing practitioners, government officials, consumer representatives, legal professionals, business representatives, and other interested parties to discuss current issues and research results pertaining to significant public policy issues in marketing. This year the MPPC was also preceded by a one-day Emerging Scholars Research Consortium with the theme “Stepping Forward with Marketing in Society Research.” Sessions here covered (1) the six subfields within marketing in society research, (2) key elements of fine research in this area, (3) presentations by editors of the key journals, (4) the work of public policy researchers in government and think tanks, and (5) identification of significant emerging issues in need of much future attention. Some 80 distinguished faculty fellows, doctoral students, and junior faculty members interested in working in this field convened for this event, which was sponsored by the University of Notre Dame and the Association for Consumer Research. The matching of the MPPC planners from Notre Dame and the FTC in 2009 was especially fitting, as the conference itself had begun in 1989 at the University of Notre Dame, with an invitation-only symposium assembled to study the future of the FTC’s operations. The conference then returned to Notre Dame for its tenth anniversary, and now it has convened in Washington, DC, with a series of special “20th Anniversary Sessions” to examine what has changed during that time, what is changing now, and what will change in the future, featuring experts from government, business, and not-for-profit organizations joining academics in addressing these issues. For example: • Given the conference’s history, it was fitting that the first 20th Anniversary session was “A Two Decade Perspective on Changing FTC Priorities, Initiatives, and Impact.”Here, three former directors of the FTC’s Bureau of Consumer Protection and a longtime chief regulator in advertising and consumer protection discussed and debated past developments and future directions. • The next 20th Anniversary session, “Exploring the Mortgage Crisis: Causes and Remedies,” featured experts who described how the situation developed, and they debated the preferred approaches for dealing with it. • This was followed by the 20th Anniversary session, “Behavioral Targeting by Marketers: Boom or Doom?” In this session, leaders of the consumer movement and key speakers from the FTC, industry, and academia explored the ramifications of emerging marketing methods on the Internet and the questions they raise about consumer privacy. • Next was “15 Years Since NLEA Regulations: Emerging Issues, New Variables, and How Nutrition Labeling Really Works,” which presented a series of findings from government research on food information disclosures. • In the last of the 20th Anniversary sessions, “JPP&M’s History and Contributions: Reflections of Its Editors,” the history of this important journal was explicated and assessed by a panel of all its editors. We discuss these separately here because only brief abstracts of the 20th Anniversary sessions appear in this volume. However, the core of the conference is well displayed herein. This featured a wide array of special sessions, competitive papers, and working papers on exciting current topics, including sustainability, consumer credit, and financial services, as well as health-related topics, such as health literacy, obesity, information disclosures, and new forms of health service provisions. For additional, rich panoply of thought represented herein, see the Table of Contents.

iii Several organizations and individuals generously contributed their time and financial resources to help make the 2009 MPPC a success. First, we thank the Marketing Science Institute, the Association for Consumer Research, and the Mendoza College of Business at the University of Notre Dame for their generous financial support of the conference and consortium. Special thanks also go to the planning committee members and reviewers for their thoughtful inputs in the planning process and for their participation in the conference itself. We also thank Jessica Bannister, Cher Dougherty, Francesca Cooley, and Marie Steinhoff of the American Marketing Association for their assistance in conference planning and publication of the conference proceedings. We greatly appreciate the very capable administrative assistance of Elizabeth Pike, ND ’09, who handled a multitude of tasks with skill, insight, and enthusiasm. Thanks also go to our colleagues at Notre Dame and the FTC, particularly John Sherry, Dean Carolyn Woo, Debbie Desrochers, and Pauline Ippolito for their support and encouragement. It was our sincere pleasure to organize this conference, and we greatly appreciate the efforts of all who participated.

Elizabeth S. Moore Janis K. Pappalardo William L. Wilkie University of Notre Dame Federal Trade Commission University of Notre Dame

iv 2009 MARKETING AND PUBLIC POLICY CONFERENCE REVIEWERS

Natalie Ross Adkins, Scot Burton, Drake University University of Arkansas

Alan R. Andreasen, John E. Calfee, Georgetown University American Enterprise Institute

J. Craig Andrews, Michael L. Capella, Marquette University Villanova University

Eric J. Arnould, Les Carlson, University of Wyoming University of Nebraska

Stacey Menzel Baker, Joel B. Cohen, University of Wyoming University of Florida

H.K. Bang, Bettina Cornwell, Villanova University University of Michigan

Barbara Bickart, Brenda J. Cude, Boston University University of Georgia

Lauren G. Block, Kirk Davidson, Baruch College–CUNY Mount St. Mary’s University

Paul N. Bloom, Brennan Davis, Duke University Baylor University

Jean J. Boddewyn, Debra M. Desrochers, Baruch College–CUNY University of Notre Dame

Lisa E. Bolton, William D. Diamond, Pennsylvania State University University of Massachusetts

Paula Fitzgerald Bone, Pam Scholder Ellen, University of West Virginia Georgia State University

Greg Bonner, Scott Erickson, Villanova University Ithaca College

Kevin D. Bradford, Karen R. France, University of Notre Dame West Virginia University

Tonya W. Bradford, George R. Franke, University of Notre Dame University of Alabama

David Brinberg, Dan Freeman, Virginia Polytechnic Institute and State University University of Delaware

v Meryl P. Gardner, Patrick J. Kaufmann, University of Delaware Boston University

Timothy J. Gilbride, Carol Kaufman-Scarborough, University of Notre Dame Rutgers University

Stacy Landreth Grau, Jeremy Kees, Texas Christian University Villanova University

Sonya A. Grier, Kathleen J. Kelly, American University Colorado State University

Joseph P. Guiltinan, Steven W. Kopp, University of Notre Dame University of Arkansas

Gregory T. Gundlach, John Kozup, University of North Florida Villanova University

Manoj Hastak, Gene R. Laczniak, American University Marquette University

Ronald Paul Hill, Russell N. Laczniak, Villanova University Iowa State University

Janet Hoek, Jeff Langenderfer, Massey University Meredith College

Jeanne M. Hogarth, Alan S. Levy, Federal Reserve Board Food and Drug Administration

Debra J. Holt, Kenneth C. Manning, Federal Trade Commission Colorado State University

Elizabeth Howlett, Ingrid M. Martin, University of Arkansas California State University, Long Beach

Bruce R. Hutton, Kelly D. Martin, University of Denver Colorado State University

Eva M. Hyatt, Marlys J. Mason, Appalachian State University Oklahoma State University

Michael R. Hyman, Alan D. Mathios, New Mexico State University Cornell University

Pauline M. Ippolito, Robert N. Mayer, Federal Trade Commission University of Utah

Easwar S. Iyer, George R. Milne, University of Massachusetts University of Massachusetts

Thomas D. Jensen, Jenny Mish, University of Arkansas University of Utah

vi John D. Mittelstaedt, Joel G. Saegert, Clemson University University of Texas at San Antonio

Anthony D. Miyazaki, Neela Saldanha, Florida International University University of Pennsylvania

Carol M. Motley, Debra L. Scammon, Florida A&M University University of Utah

Dennis R. Murphy, Kristin Scott, Federal Trade Commission Oklahoma State University

Patrick E. Murphy, Kathleen Seiders, University of Notre Dame Boston College

Julie L. Ozanne, Clifford J. Shultz, Virginia Tech University Arizona State University

Cornelia Pechmann, David E. Sprott, University of California, Irvine Washington State University

Laura A. Peracchio, Julie V. Stanton, University of Wisconsin–Milwaukee Pennsylvania State University

Vanessa Gail Perry, David W. Stewart, George Washington University University of California, Riverside

Mark Peterson, Mary Sullivan, University of Wyoming George Washington University

Ross D. Petty, John L. Swasy, Babson College American University

Michael Jay Polonsky, Charles R. Taylor, Deakin University Villanova University

Nora J. Rifon, Madhu Viswanathan, Michigan State University University of Illinois, Urbana–Champaign

Debra J. Ringold, Joshua L. Wiener, Willamette University Oklahoma State University

José Antonio Rosa, Jerome D. Williams, University of Wyoming University of Texas at Austin

Randall L. Rose, Terrence H. Witkowski, University of South Carolina California State University, Long Beach

Herbert Jack Rotfeld, Auburn University

vii TABLE OF CONTENTS

PREFACE iii

MANUSCRIPT REVIEWERS v

TABLE OF CONTENTS viii

A TWO DECADE PERSPECTIVE ON CHANGING FTC PRIORITIES, INITIATIVES, AND IMPACT

A Two Decade Perspective on Changing FTC Priorities, Initiatives, and Impact Manoj Hastak, Patrick E. Murphy, J. Howard Beales III, Joan Z. Bernstein, William C. MacLeod, C. Lee Peeler 1

TESTING THE LIMITS: WHEN TO USE EDUCATION DISCLOSURES, OR REGULATION AS CONSUMER PROTECTION STRATEGIES IN THE FINANCIAL SERVICES MARKET

Basic Training: The Impact of Financial Education at the Beginning of a Soldier’s Career Catherine Bell, Daniel Gorin, Jeanne Hogarth 3

Mortgage Originator Compensation and Incentives: Market Function and the Role of Disclosures Ellen A. Merry 5

The One-Two Punch: Using Summary Information to Increase Financial Disclosure Effectiveness John Kozup 7

PROBING THE FRONTIER: VULNERABLE CONSUMERS IN A SOPHISTICATED MARKETPLACE

Political Literacy and Low Literate Consumers Natalie Ross Adkins, Randall E. Adkins 8

Marketing Sex, Drugs, and Desperation: An Ethical and Bi-Cultural Analysis of the “Meth Project” Matt Hettche, Joseph Scarpaci 10

Regulating Product Placements in Children’s Entertainment: Cognitive and Behavioral Effects of Placements on Children Haiming Hang, Susan Auty 21

Immigration Policy and Vulnerability Garrett Coble, Fernando Jiménez, Marlys J. Mason 33

EXPLORING THE MORTGAGE CRISIS: CAUSES AND REMEDIES

Exploring the Mortgage Crisis: Causes and Remedies William L. Wilkie, Karen M. Pence, Barry Zigas, Todd J. Zwicki, John G. Lynch, Jr. 35

viii PURSUING SOCIAL GOALS – LEARNING TO DO BETTER

Warning Consumers about Alcohol Messages in Television Series: Audience Connectedness, Warning Timing, and Emphasis Cristel Antonia Russell, Dale W. Russell, Joel W. Grube 36

Tobacco Branding and Plain Packaging: The New Frontier in Tobacco Control? Janet Hoek, Philip Gendall, Jordan Louviere 38

Impact of Marketing on Infant Formula Choice and Switching Sara B. Fein, Yi Huang, Conrad Choiniere, Judith Labiner-Wolfe 40

Promoting Safety Versus Preventing Aggressiveness: A Comparison Between Gain and Loss Framed Messages in Safe Driving Campaigns Georgiana Craciun, Jason Q. Zhang, Dongwoo Shin 42

CURRENT ISSUES AND RESEARCH ON HEALTH LITERACY: MEASUREMENT AND POLICY DIRECTIONS

Creating a Skills-Based Health Literacy Measure Lauren A. McCormack, Nancy D. Berkman, Linda Squiers, Dean Schillinger, Judith Hibbard, Janet Ohene-Frempong, Rahima Jan Gates, Carla Bann, Claudia Squire, Tania Fitzgerald 45

Understanding Risk Comprehension in Direct-to-Consumer Advertising Kathryn Aikin, Amie O’Donoghue, Helen Sullivan, Jack Swasy 46

Strategies to Advance Health Literacy Policies in the U.S. Linda Neuhauser 47

BEHAVIORAL TARGETING BY MARKETERS: BOOM OR DOOM?

Behavioral Targeting by Marketers: Boom or Doom? Stephen Brobeck, Jessica L. Rich, Pablo Chavez, Susan Grant, George R. Milne 48

RECENT APPLICATIONS OF CONSUMER RESEARCH AT FEDERAL AGENCIES

Developing Mandatory Energy Labels That Do Not Mislead or Bias Consumers: Lessons from Consumer Research at the Federal Trade Commission James Hilger, Janis Pappalardo, Manoj Hastak 52

Exploring the Effects of Decision Structure, Stakes, and Information Provision in a Fraud Case Robert Letzler, Isaac Knowles, Patrick McAlvanah, Manoj Hastak 53

Mercury Advisories and Household Health Trade-Offs Jay P. Shimshack, Michael Ward 54

ix RISK, VULNERABILITY, AND IMPROVISATION IN NATURAL DISASTER PLANNING, MITIGATION, AND RECOVERY: IMPLICATIONS FOR MARKETING & PUBLIC POLICY

The Inevitability of Disasters: The Role of Planning and Improvisation in Sustainable Tourism Deirdre T. Guion, Debra L. Scammon, Leila Borders 56

Understanding the Risk Perceptions and Attitudes of Disaster Subcultures Ingrid M. Martin, Wade E. Martin, Carol Raish 57

Understanding Risk and Vulnerability: Implications for Policy Making and Resource Allocation Stacey Menzel Baker, Ronald Paul Hill 58

ADOLESCENTS AND RISKY BEHAVIORS

Extending the Obesity Discussion: Teen Diets, Behavior, and Use of Nutritional Information Debra M. Desrochers, Elizabeth S. Moore 60

The Effects of Adolescent Age and Anti-Tobacco Ad Campaign Attitudes on Anti-Smoking Beliefs and Intentions to Smoke J. Craig Andrews, Richard G. Netemeyer, Scot Burton 62

Abstinence Education and Public Policy: Effectiveness of Abstinence-Only Programs as Social Marketing Mary Anne Raymond, John F. Tanner, Jr., John D. Mittelstaedt 64

PSYCHOLOGICAL RESEARCH ADVANCES: AT THE FRONTIER

Misperception and Confusion as Unintended Consequences of Government-Mandated Warning Labels Deborah J. MacInnis, Heather Zhong Wan 65

Can Consumers Correct for Subtle Persuasion? Brand Placement Effects and the Important Roles of Disclosure Information and Timing Margaret C. Campbell, Gina S. Mohr, Peeter W.J. Verlegh 73

Vicarious Goal Fulfillment: When the Mere Presence of a Healthy Option Leads to an Ironically Indulgent Decision Keith Wilcox, Beth Vallen, Lauren G. Block, Gavan J. Fitzsimons 75

THE ROLE OF QUALITY INFORMATION IN HEALTHCARE MARKETS: COMPETITION AND CONSUMER PROTECTION ISSUES

The Role of Quality Information in Healthcare Markets: Competition and Consumer Protection Issues Debra L. Scammon, Jack Calfee, Patricia Schultheiss, Kathlene Kendrick, Tom Miller 77

x DEMOCRACY, CIVIC PARTICIPATION, AND CONSUMPTION

Exploring the Emergence of Civic Consumer Culture Through Politicized Shopping Janine Dermody, Stuart Hanmer-Lloyd, Richard Scullion 81

Deliberative Democracy for Multi-Stakeholder Engagement Process Canan Corus 82

The Consumption Strategies of the Poor and the Meaning of Poverty Bige Saatcioglu 83

15 YEARS SINCE NLEA REGULATIONS: EMERGING ISSUES, NEW VARIABLES, AND HOW NUTRITION LABELING REALLY WORKS

The Consumer Impact of Carbohydrate Claims: Wording Effects and Front Label/Back Label Interactions Judy Labiner-Wolfe 86

Consumer Ability to Interpret Trans Fat Disclosures on the Nutrition Facts of Food Products Conrad J. Choiniere, Alan S. Levy 87

Consumer Impacts of Trans Fat Claims in the Presence of Strength of Science and Marketing Primes When Product Nutrient Quality Is Manipulated Alan S. Levy, Conrad J. Choiniere 88

Barriers to Food Label Use among Younger Consumers Conrad J. Choiniere, Anu Mitra 89

POSTER SESSION

Changes in U.S. Home Equity Lending: Evidence from the 1989 Through 2007 Survey of Consumer Finances Jeanne M. Hogarth, Andrew D. Vlietstra 90

Consumer Deskilling and the Credit Card Industry: An Intergenerational Perspective Sarita Ray Chaudhary, Yam B. Limbu 92

Measuring Consumer Perceptions of Fairness in Financial Services James Devlin 94

Online Banking and Identity Theft: Update and Call for Action Ruxandra Niculescu, Mark O’Lexy, Ryan Schiffino, Carol Kaufman-Scarborough 96

A Framework for Social Change: Integrating Marketing and Public Policy David Brinberg, Meghan Pierce, Kim Daniloski, Julien Stauffer 98

Kingdon’s Multiple Streams Framework: a Policy Analysis Tool for Marketing and Society Scholarship Jenny Mish 100

Sustainability, Overconsumption, and Consumer Well-Being: Implications for Marketing’s Role Within a Global Society Michael Bone, Adam Finn 102

xi Consequences of Legislating Packaging Behavior: Germany’s Green Dot Program Lukas Jones, Lynn R. Kahle, Scott Owen 104

Social Values Segmentation in Marketing Residential Solar Power Installations Scott Owen, Lynn R. Kahle 106

The Usefulness of Surveys in Antitrust Carol Miu 108

Determinants of Property Rights: More Empirical Evidence Thomas G. Costello, Ayse Olcay Costello 110

Heading Toward 1984? An Exploratory Study of Business Models and Regulatory Frameworks for Interactive Advertising in the U.S. and the E.U. Olesya Venger 112

The Use of Public Relations to Promote a Plan for Rebuilding a City after a Disaster Owen Kulemeka 114

Contemporary Consumerism: Revealing 21st Century Concerns Kristen L. Walker 116

Why Did I Eat That? Exploring the Effects of Perceived Goal Progress and Rumination on Consumer Health Decisions My Bui, Elizabeth Howlett, Scot Burton, Steven W. Kopp 118

The Relationship Between Homesickness and Compensatory Food Consumption among College Students Elizabeth Crosby, Cele C. Otnes 120

When Food Is More than Nutrition: Regulating Emotions with Food Elyria Kemp, Steven W. Kopp 124

Individualism, Collectivism, and Corporate Social Responsibility: Influence on Obesity Beliefs and Attitudes Toward Public Policy Jung-Sook Lee 126

Improving Effectiveness of Restaurant Menu Labeling: An Exploratory Research Saroja Subrahmanyan 128

Consumer Use of Nutrition Symbols Monique Mitchell Turner, Lucinda Austin, Meryl Lubran, Sejal Patel, Christine Skubisz 130

The Relationship Between Advertising Vulnerability and Weight-Loss Strategy Usage among Obese Consumers Sheri L. Worthy, Kenneth Pilcher, Kristine Lokken, Abbe Boeke 132

The Magical Powers of Supplements: Do Supplements Always Lead to a Healthy Lifestyle? Sayantani Mukherjee, Ingrid M. Martin 134

A Test of Objectification Theory on Female Consumers’ Product Evaluations Chrissy Mitakakis, Stephen J. Gould 136

xii Are You Ashamed? The Role of Shame in Health Communications Kate Pounders 138

Toward a Drug Free America: an Evaluation of Anti-Drug PSA Effectiveness Across Sensation Seeking and Guilt Proneness Karen Becker, Rowena Briones, Brian Keefe 139

Shifting Risk-Taking Preferences to Risk-Avoidance Preferences: The Effects of Marketing Messages on Chronic Regulatory Focus Jane McKay-Nesbitt, Malcolm C. Smith 141

Mothers’ Perceptions of Word of Mouth Marketing When Teens Are Used as Buzz Agents Bob D. Ahuja, Tara A. Michels, Clint S. Schertzer, Mary M. Walker 143

The Role of Values in Cause-related Marketing: A Thai Pharmaceutical Industry Study Apichai Apirattanapimolchai, Peter Vitartas 146

A Comprehensive Analysis of Consumer Targeting Thomas A. Klein 148

Targeting Different Segments of College Student Smokers: Preliminary Results and Proposed Evaluation Strategies William D. Diamond 150

Aberrant Service Consumer Behavior and Coping Sally A. Hibbert, Maria G. Piacentini, Margaret. K. Hogg 152

Could We Have Seen it Coming?: On the Prospect of Mining Online Consumer Review Data in Order to Forecast Product Safety Issues Gavin Huntley-Fenner, Miles Munro 155

A Customer Service Conceptualization of the Ehavioral Advertising Privacy Debate JoAnna Williamson 157

Consumer Complaint Channel Choice in Online and Offline Purchase Environments: If I Buy Online, Will I Complain Online? Seul Lee, Brenda J. Cude 159

Preteen Children’s Online Privacy: How Online Usage Motives Affect Privacy Concerns and Information Disclosure May O. Lwin, Anthony D. Miyazaki, Andrea J.S. Stanaland, Evonne Lee, Mindawati Wijaya 161

It’s Cool: Review and Anticipated Impacts of the New Country of Origin Food Labeling Law Christopher Newman, Steven W. Kopp 163

JPP&M’S HISTORY AND CONTRIBUTIONS – REFLECTIONS OF ITS EDITORS

JPP&M’S History and Contributions – Reflections of its Editors Alan Andreasen, Thomas C. Kinnear, Patrick E. Murphy, Michael B. Mazis, Debra L. Scammon, J. Craig Andrews, Joel B. Cohen, Ronald Paul Hill 164

xiii UNSUSTAINABLE CONSUMPTION HABITS: INSIGHTS AND IMPLICATIONS FOR MARKETING AND PUBLIC POLICY

Construal Level Theory and Unhealthy Consumption Brennan Davis, Cornelia Pechmann 166

Nudging Better Choices in School Cafeterias Brian Wansink, Collin Payne 167

Over-Consumption of Ultra-Concentrated Products Maura L. Scott, Stephen M. Nowlis, Naomi Mandel 168

Sustainable Snacking: Aligning Regulatory Orientation and Snack Decision Strategies for Better Results Leona Tam, Jelena Spanjol, José Antonio Rosa 169

SMOKERS ARE NOT ALL THE SAME: THE IMPORTANCE OF UNDERSTANDING MOTIVATIONAL DIFFERENCES AMONG VARIOUS SMOKER SEGMENTS WHEN DEVELOPING POLICY STRATEGIES

Identifying At-Risk Youth Segments and Reactions to Anti-Tobacco Campaigns Marlys J. Mason, Joshua L. Wiener 173

The Dilemma of Social Smokers Believing That They Are Not Smokers Ingrid M. Martin, Michael Kamins, Aditi Grover 174

Countering Reactance, Entitlement, and Misguided Optimism in Smoking Cessation Strategies Joyce M. Wolburg 175

THE CHALLENGES OF SUSTAINABLE MARKETING

Marketing Claims Regarding Carbon Offset Programs: A Comparison of U.S. and Australian Consumer Knowledge and Behavior Stacy Landreth Grau, Michael Jay Polonsky, Romana Garma 176

Environmentally Friendly Consumption Practices: The Ecology of Cultural Context Marie Hafey DeVincenzo 178

Sustainable Consumer Behaviors: A Conceptual Framework Kristin Scott 180

Generational Cohort Effects on Green Consumer Behavior Gail M. Zank, Karen H. Smith 182

CONSUMPTION CONSTRAINTS: HIGHLIGHTING PERSPECTIVES FROM THE SPRING 2009 SPECIAL ISSUE OF JPP&M WITH DIRECTIONS FOR FUTURE RESEARCH

Consumption Constraints: Highlighting Perspectives from the Spring 2009 Special Issue of JPP&M with Directions for Future Research Stacey Baker, Gene Laczniak, John D. Mittelstaedt, Julie L. Ozanne, Nicholas Santos 192

xiv PATIENT CENTERED MEDICAL HOME: PROSPECTS FOR TRANSFORMATION OF HEALTHCARE

Patient Centered Medical Home: Prospects for Transformation of Healthcare Debra Scammon, Kevin Grumbach, Julie Day, Edwina Rodgers, Sam Butler 194

A GROWING INTERNATIONAL PROBLEM: OBESITY

Loss of Balance? An Examination of Obesity, Individual Responsibility, and Deep Capture in New Zealand Janet Hoek, Rachael McLean 196

When to Say When? Self-Regulatory Rituals for Eating and Drinking Vanessa Gail Perry, Joonwhan David Lee 198

Visceral Cues and Consumer Skepticism: Consumer Reactions to Weight Loss Advertising Clinton Amos, Stacy Landreth Grau 200

An Investigation into Individual’s Repeated Attempts at Behavior Change Courtney Droms 202

POLICY ISSUES WITH CONSUMERS AND CREDIT

Failure to Understand Compound Interest and Mitigating Strategies Eric M. Eisenstein, Stephen Hoch 204

The Effectiveness of Real Estate Foreclosure Auctions: A Case Study and Policy Implications Ekaterina V. Karniouchina, Debra L. Scammon, William L. Moore, Dan A. Fuller 206

Risk-Based Pricing or Excuse-Based Pricing? An Analysis of Penalty Repricing in the Credit Card Industry Joshua M. Frank 217

BLIND MEN AND AN ELEPHANT: TRIANGULATION OF RESEARCH METHODS TO DOCUMENT THE DISCRIMINATION OF MINORITY CONSUMERS AND MINORITY ENTREPRENEURS

An Experimental Scenario-Based Approach to Analyzing Consumer Response to Marketplace Discrimination Jerome D. Williams, Geraldine R. Henderson, Anne-Marie G. Hakstian, Sophia R. Evett 221

The Plight and Pilgrimage of Minority Small Business Entrepreneurs: Exploring the Experience of Vulnerable Populations with ZMET Sterling A. Bone, Glenn L. Christensen, Jerome D. Williams, Elise Riker 223

Market Research for Detecting and Eliminating Racial Profiling and Discrimination in Lending Paul C. Lubin 225

xv ADVANCES ON TNE CLASSIC PROBLEM HOW TO EFFECTIVELY DISCLOSE

A Method to Track Viewing Behavior: Demonstrating the Effect of a Concurrent Disclosure in Dtc Advertising Jack Swasy, Anu Mitra 227

Does Presentation Make a Difference to Risk Perception: Testing Different Formats for Communication of Cancer Risks Sandra C. Jones 229

The Effect of Front-of-Package Nutrition Information and Product Claims on Consumers’ Attitudinal Evaluations and Choice Behavior Ninya Maubach, Janet Hoek, Philip Gendall, Duncan Hedderley 239

Perceptions of Nutrition Symbols by Some U.S. Consumers Chung-Tung Jordan Lin 241

COLLISION? MARKETING, RECIPROCITY, AND PRIVACY IN AN EVOLVING SOCIETY

Beyond the Water Cooler: Gender Differences in Privacy-Related Measures for Facebook Profile Usage Beyond its Original Purpose Mariea Grubbs Hoy, George R. Milne 243

What Californians Understand About Privacy Offline Chris Jay Hoofnagle, Jennifer King 245

Consumer Perceptions of Online Privacy: Behavioral Targeting and the Value of Personal Information Poonam Kumar, Eloise Coupey, Mary Peck 253

Policy Implications of Reciprocity and Independent Judgment Jeff Langenderfer, Marlys J. Mason, Xiang Fang 256

FOOD FOR THOUGHT: APPEALS, CHOICES, (AND CONFUSION)

Consumer Confusion in the Marketplace Paula Fitzgerald Bone, Jeremy Kees, Karen Russo France, John Kozup 258

The Mere Categorization Effect on Perceived Healthiness and Consumption of Food Items Caglar Irmak, Beth Vallen 260

Are Parents Vulnerable, Susceptible or Savvy to Advertising Strategy? An Analysis of Emotion-Based Appeals in Food Ads Deirdre T. Guion, Julie V. Stanton 262

Using the Theory of Planned Behavior as a Framework for Understanding the Organic Food Consumer: An Exploratory Study Sally A. Williams, Karen H. Smith 264

xvi HELPING CONSUMERS TO HELP THEMSELVES (AND OTHERS)

Perceptions of Service Quality with Credence Services Ann M. Mirabito, Leonard L. Berry, Jelena Spanjol 273

Reducing Resistance to Planning for End-of-Life: A Social Marketing Approach Elyria Kemp, Steven W. Kopp 275

What’ S Behind My Good Deeds? Self-Enhancement, Self-View, and Donation Joon Yong Seo, Debra L. Scammon 277

DEVELOPMENTAL PERSPECTIVES ON CHILDREN’S RESPONSE TO MARKETING STIMULI

You Drive a Prius? I Bet I Know What Brand of Cereal You Eat! Chidlren’s Understanding of Consumption Constellations Lan Nguyen Chaplin, Tina Lowrey 281

Do Sales Promotions Persuade Children to Buy? Wendy Attaya Boland, Lance-Michael Erickson 283

Long-Term Effects of Advertising to Children on Judgment in Adulthood Paul M. Connell, Merrie Brucks, Jesper Nielsen 285

AUTHORS 287

xvii

20th ANNIVERSARY SESSION A TWO DECADE PERSPECTIVE ON CHANGING FTC PRIORITIES, INITIATIVES, AND IMPACT

Session Chairs: Manoj Hastak, American University, Washington, D.C. Patrick E. Murphy, University of Notre Dame

Panelists: J. Howard Beales III, George Washington University, Washington, D.C. Joan Z. Bernstein, Bryan Cave LLP William C. MacLeod, Kelley Drye & Warren LLP C. Lee Peeler, National Advertising Review Council

SUMMARY Congressional mandate set out it in 1914 and to determine what type of agency the FTC should aspire to be as it This year we are celebrating the 20th Anniversary Market- enters its second century [see, for example, http:// ing & Public Policy conference. The first conference was www2.ftc.gov/opa/2009/01/ftc100rpt.shtm]. held in 1989 at the University of Notre Dame, as an invitation-only symposium to commemorate the seventy- Given this conference’s history, it is fitting that our first fifth anniversary of the U.S. Federal Trade Commission 20th Anniversary session will revisit the focus of the first (FTC) and to provide a forum for a hard-hitting assess- symposium, again to discuss and debate past development of its past performance and future directions. The ments and future directions. The session will feature four event was non-partisan, with both liberal and conserva- experts on these issues: three are former Directors of the tive views represented among the congregated experts. FTC’s Bureau of Consumer Protection, serving at very Also, there was significant interest in assessing how the different periods of the agency’s operations, and one was academic fields of marketing, consumer research and a chief FTC advertising regulator across a number of years economics might help to improve their contributions to and programs: better public policy decisions. The contents were memo- rialized in the significant book, Marketing and Advertis- J. Howard Beales is currently a professor of strategic ing Regulation: The Federal Trade Commission in the management and public policy at George Washington 1990’s (Murphy and Wilkie 1990). University. He has had a long career in government, including service at the Office of Management and Bud- The Federal Trade Commission, established in 1914, is get and in posts within the Federal Trade Commission, the government agency charged with regulating most of culminating as Director of the Bureau of Consumer Pro- the marketing activity in the U.S. economy. It was created tection from 2001–2004. Howard holds his undergradu- as an independent regulatory commission, having the ate degree from Georgetown University and his PhD. in broad responsibility of providing to the nation a market- Economics from the University of Chicago. place that is efficient and fair to both competitors and consumers (Murphy and Wilkie 1990). The Commission Joan Z. Bernstein is currently of counsel, Bryan, Cave pursues two spheres of regulation, competition policy and LLP. She has had a long career in government, business, consumer protection. and private legal practice, including service as General Counsel of the U.S. Department of Health and Human Given its broad but vague mandate, the FTC is often the Services (including oversight of the FDA) and General center of criticism, both for doing too much and for doing Counsel of the U.S. Environmental Protection Agency. too little. In anticipation of its 100th anniversary in 2014, Recently she led the food industry’s self-regulatory effort the Commission has been engaged in a self-assessment to develop the Children’s Food and Beverage Initiative. process to reflect on how well the agency is fulfilling its She served at the FTC in the active period of the 1970’s,

Marketing and Public Policy Conference Proceedings / 2009 1 20th Anniversary Session: A Two Decade Perspective On Changing FTC Priorities, Initiatives and Impact then returned as Director of the Bureau of Consumer C. Lee Peeler, is currently President and CEO of the Protection from 1995–2001. Jodie holds her undergradu- National Advertising Review Council (NARC) and Ex- ate degree from the University of Wisconsin and her Juris ecutive Vice President, National Advertising Self-Regu- Doctorate from Yale University. lation, Council of Better Business Bureaus (CBBB). He is responsible for leading the advertising industry’s system William C. MacLeod is currently a partner at Kelley Drye of self-regulation. Lee joined NARC and BBB in 2006, & Warren LLP, focusing on competition, trade regula- following a 33-year career at the Federal Trade Commis- tion, and advertising. He has extensive experience in sion. While at the FTC, he served as Associate Director of private practice as well as in government, including ser- the Division of Advertising Practices (1985–2001) and vice with the U.S. Department of Justice and several Deputy Director, Bureau of Consumer Protection (2001– positions within the FTC, where he served as Director of 2006). Lee received his B.A. and J.D. degrees from the Bureau of Consumer Protection from 1986–1990. Bill Georgetown University. holds his undergraduate from Ripon College and his Juris Doctorate from the University of Miami.

2 Marketing and Public Policy Conference Proceedings / 2009 SPECIAL SESSION

BASIC TRAINING: THE IMPACT OF FINANCIAL EDUCATION AT THE BEGINNING OF A SOLDIER’S CAREER

Catherine Bell, Federal Reserve Board, Washington, D.C.1 Daniel Gorin, Federal Reserve Board, Washington, D.C. Jeanne Hogarth, Federal Reserve Board, Washington, D.C.

SUMMARY • More likely to use an informal versus formal budget for spending, Financial education has risen on the agendas and priority lists of a number of agencies and organizations. In fact, the • More likely to comparison shop for major purchases, Journal of Consumer Affairs commissioned a special issue in 2008 (Vol. 42, No. 2) focusing on financial • More likely to understand the differences between literacy and public policy and a second special issue on discretionary and non-discretionary spending, financial literacy is planned for 2010. However, the literature on evaluating the impacts of financial education is • still quite thin. This study is one attempt to move the field More likely to read about money management in forward by addressing two common criticisms of finan- magazines, cial education program evaluation literature: lack of a comparison group and documenting changes over time. • More likely to engage in a moderate to high levels of good behaviors, The Army Emergency Relief (AER), the U.S. Army post at Ft. Bliss in El Paso, TX, and the Federal Reserve Board • Less likely to never pay off their credit card balances, collaborated to provide financial education for young soldiers and to evaluate the impact of that education on the • Less likely to use car title loans, soldiers’ financial management behaviors. Soldiers attending the Army’s air defender advanced individualized • Less likely to buy supplemental life insurance. training (AIT) at Ft. Bliss were offered a two-day financial education course taught by staff from San Diego City There were no significant differences in other behaviors, College; funding for the course is provided by AER.2 At such as over-drawing accounts and bouncing checks, the end of the course, soldiers complete a survey that paying bills late, having an emergency fund, savings serves as a baseline for the evaluation. A second group of habits, using payday loans, or in planning horizons. soldiers at Ft. Bliss, who did not participate in the financial education course, serve as a comparison group. A follow- In mulitvariate analyses, soldiers taking the financial up survey was conducted in January 2008 to provide a education course were more likely than the comparison second data point for those who took the course.3 group to report using informal spending plans and less likely to report using formal spending plans. We also In a Nutshell, What Did We Learn? found that high school financial education programs made a difference: those who had a high school financial edu- • Financial education did seem to have an effect on cation course were more likely to have a savings account specific financial management behaviors. In simple for short-term savings goals and to save regularly and bivariate analyses with the comparison group, sol- were less likely to pay an overdraft fee in the past six diers who had taken the financial education course months. Early financial management experience also seems were: to matter: soldiers who had a high school savings account were more likely to have an emergency fund, more likely • More likely to have retirement savings plans and to read money management articles, and less likely to participate in the Thrift Savings Plan for retirement, “never” pay off their credit card balances.

Marketing and Public Policy Conference Proceedings / 2009 3 Special Session: Testing the Limits: When to Use Education Disclosures, or Regulation as Consumer Protection Strategies In the Financial Services Market

Does this Mean Education “Doesn’t Work?” cial education program, which we plan to investigate in our future work. There are a number of possible explanations for our results. We plan to look at the lapsed time between taking We recognize that our results are not generalizable to the the course and the follow-up survey to explore the effects general public, or even to young people in general. While of timing on behavior change. We know that people can we believe this study helps address the research issues of be at different pre-action stages; thus, we may have having a comparison group and documenting changes missed some of the impacts of the financial education over time, it is not the “gold standard” of a randomized program. We plan to explore possible interactions be- experiment. Finally, we believe that education is neces- tween earlier educational experiences and the financial sary but it alone is not sufficient to establish financially education course as well as explore longer-term impacts. secure families and households. Important complements Substantial portions of the class are devoted to car buying, include access to information, access to financial counsel- which we did not investigate here. Analysis of the vehicle ing and advising, and public policies that provide con- ownership data may reveal additional effects of the finan- sumer protection.

ENDNOTES 2 AIT generally takes place immediately after basic training; depending on the course of instruction, it can last 1 The analysis and conclusions set forth in this presenta- between 6 to 12 weeks. After AIT, the soldiers are tion represent the work of the authors and do not posted to their first official duty station. During the indicate concurrence of the Federal Reserve Board, course of this study, most air defenders stayed at Ft. the Federal Reserve Banks, or their staff. Mention or Bliss or were deployed to Korea or Southwest Asia. display of a trademark, proprietary product, or firm in 3 In January 2009, we conducted a second follow up the presentation by the author does not constitute an survey with the soldiers. This report only includes endorsement or criticism by the Federal Reserve data from the baseline surveys and first follow-up System and does not imply approval to the exclusion survey in January 2008. of other suitable products or firms.

For further information contact Jeanne Hogarth Consumer & Community Affairs, MS 805 Federal Reserve Board 20th & C Streets, N.W. Washington, D.C. 20551 Phone: 202.785.6024 Fax: 202.785.6085 E-Mail: [email protected]

4 Marketing and Public Policy Conference Proceedings / 2009 Special Session: Testing the Limits: When to Use Education Disclosures, or Regulation as Consumer Protection Strategies In the Financial Services Market

MORTGAGE ORIGINATOR COMPENSATION AND INCENTIVES: MARKET FUNCTION AND THE ROLE OF DISCLOSURES

Ellen A. Merry, Federal Reserve Board, Washington, D.C.1

SUMMARY in 2004 the Federal Trade Commission (FTC) released testing results for a broker compensation disclosure pro- This paper explores information problems in the market posed by the Department of Housing and Urban Develop- for mortgage origination services in the U.S. with a focus ment (HUD) and found that the disclosure made it more on the role and incentives of the employees who work difficult for consumers to choose the lowest cost loan and directly with consumers, including both employees of created an anti-broker bias. HUD conducted further test- lenders and independent mortgage brokers who work ing and identified changes that could be made to the with multiple lenders. While the operation of the mort- disclosure to address these problems. HUD included a gage market involves many different parties – consumers, revised presentation of broker compensation in the settle- originators, lenders, investors, insurers, servicers, regula- ment cost disclosures released with their rule issued in tors, and ratings agencies to name a few – originators are November 2008. In December 2007, the Federal Reserve typically the “face” of the mortgage supply chain to Board (FRB) proposed a new disclosure for broker com- consumers. The nature of the originator-consumer rela- pensation, but subsequently withdrew the proposal after tionship is a complex one, combining sales with advice on testing showed problems with consumer comprehension a major financial decision. In addition, many consumers of the disclosure. In addition to these studies that deal find their originators through existing personal and pro- directly with information policy for U.S. mortgage origi- fessional relationships, which can make the nature of the nators, this discussion may be supplemented with results originator’s incentives in the transaction more ambigu- from research on consumer understanding of the role of ous. different types of advisors, such as mortgage originators in the U.K. and investment advisors in the U.S. The paper begins with a discussion of the channels for mortgage origination in the U.S. and the forms of compen- In addition to disclosure, other approaches for improving sation that originators receive. This provides a ground- the market for mortgage origination services have been work for understanding the general workings of the mar- proposed by market participants or policy makers. These ket, the incentives originators face, and some of the include altering the incentives of originators through the challenges consumers may confront when they shop for a structure of compensation; state laws imposing duties on loan. originators; and monitoring efforts to help identify “bad actors.” Some of these non-disclosure strategies are dis- Discussion then focuses on results from consumer testing cussed as well. on disclosures for originator compensation. Specifically,

ENDNOTE display of a trademark, proprietary product, or firm in the presentation by the author does not constitute an 1 The analysis and conclusions set forth in this presenta- endorsement or criticism by the Federal Reserve tion represent the work of the authors and do not System and does not imply approval to the exclusion indicate concurrence of the Federal Reserve Board, of other suitable products or firms. the Federal Reserve Banks, or their staff. Mention or

Marketing and Public Policy Conference Proceedings / 2009 5 Special Session: Testing the Limits: When to Use Education Disclosures, or Regulation as Consumer Protection Strategies In the Financial Services Market

For further information contact: Ellen Merry Consumer & Community Affairs, MS 805 Federal Reserve Board 20th & C Streets, N.W. Washington, D.C. 20551 Phone: 202.736.5627 Fax: 202.785.6085 E-Mail: [email protected]

6 Marketing and Public Policy Conference Proceedings / 2009 Special Session: Testing the Limits: When to Use Education Disclosures, or Regulation as Consumer Protection Strategies In the Financial Services Market

THE ONE-TWO PUNCH: USING SUMMARY INFORMATION TO INCREASE FINANCIAL DISCLOSURE EFFECTIVENESS

John Kozup, Villanova University

SUMMARY to reduce the decision-making burden and allow for learning transfers. The question that emerges is “how?” How can we inform consumers in the most effective With nutrition labels, no matter how healthy or unhealthy manner to help them realize their financial goals? A the food is, consumers can find the information on the worthwhile financial education program starts with hav- label. For financial products as a whole, no such consis- ing the participants set a goal; for example; saving for tency exists. The objective of this discussion is to explore retirement. While choosing how to best invest those whether a summary disclosure impacts consumers’ evalu- savings is one of the most important decisions a consumer ations of a financial product. Parallels between nutrition can make; this can be an especially challenging task given education and disclosure programs and their financial the current financial information disclosure environment. counterparts will be the focus of this discussion with Researchers and policy makers seem to support some results from both areas disseminated. standardized way to present financial information in order

For further information contact: John Kozup College of Commerce and Finance 2051 Bartley Hall Villanova University 800 Lancaster Ave. Villanova, PA 19085–1678 Phone: 610.519.6441 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 7 POLITICAL LITERACY AND LOW LITERATE CONSUMERS

Natalie Ross Adkins, Drake University, Des Moines Randall E. Adkins, University of Nebraska – Omaha

SUMMARY We conduct an analysis of adult learners’ writings and group discussion transcripts solicited by a U.S. federal Throughout United States history, state and local govern- agency, the National Institute for Literacy (NIFL). The ments inhibited and, at times, prohibited the participation analysis focuses on discovering how the low literate of low literate adults in the political process. Up until the consumers generally view participation in the political passage of the Voting Rights Act of 1965, many states marketplace and specifically the impact of their limited required potential voters to pass literacy tests before being literacy skills on that world view (McCracken 1988). allowed to vote. Today, the federal government protects Using a hermeneutic approach we tack back and forth the civil rights of low literate voters by acting to prevent between the data and the relevant literature on literacy, efforts to discriminate against them because of their political literacy, and low literate consumers. limited literacy skills. Consistent with previous research (cf., Adkins and Ozanne While literacy can no longer be used as a litmus test to 2005), the data provides support for both the sociocultural qualify participants in the political process, literacy skills and functional perspectives of political literacy. The data undoubtedly impact voters’ participation. According to suggest low literate consumers who engage the political the 2003 National Assessment of Adult Literacy (NAAL) system begin seeing themselves as possessing agency adults without a high school degree constitute the largest with the potential to participate in and transform the group with “below basic” literacy skills in the measured political marketplace (Fingeret and Drennon 1997; Friere prose, document and quantitative literacy assessments 1986). Other findings more closely align with a functional (Kutner, Greenberg, and Baer 2005). Approximately 45 perspective of political literacy in which consumers’ percent of these adults scored in the lowest level for limited literacy abilities actively inhibit political partici- document literacy which is extremely relevant for politi- pation in part out of fear of voting “the wrong way.” Under cal participation. Although low literate adults likely pos- both sociocultural and functional perspectives of political sess limited political knowledge their right to participate literacy, a need for political marketing collaterals written in the political process remains. However, an analysis of in clear, plain language fashion clearly exists. U.S. Census Bureau current population reports on voter registration and actual voting behavior participation by educational attainment reveals about one-third of the Implications for Public Policy and Marketing eligible voters with the lowest levels of education register to vote and less than one-quarter do participate as voters Through this exploratory study we suggest two main (Stanley and Niemi 2006). Thus, although low literate avenues warranting additional research by marketing adults have the basic right to vote and participate in the academicians and public policymakers. political marketplace, a vast majority choose not to exercise this right given to U.S. citizens. Ethically Targeting and Engaging Low Literate Con- sumers. More research focused on the factors which We provide a brief overview of the historical literacy prevent low literate voters from engaging in the political requirements for political participation in the United States marketplace could shed light on how various channels of as well as recent research on low literate consumers in the communications are perceived and used. For example, in generalized marketplace to introduce two primary re- addition to the traditional channels of communications search questions: such as newspapers, radio, and television how might voter 1. What does it mean – at least symbolically – to a low workshops and door-to-door canvassing be used to in- literate consumer to exercise political participation; crease political participation? Further, the costs associ- and ated with face-to-face campaigning raises serious questions about the need for reaching out to low literate voters 2. What motivates or encourages low literate consum- and the limitations on campaign spending imposed by ers to participate? campaign finance reforms.

8 Marketing and Public Policy Conference Proceedings / 2009 Educating and Empowering Low Literate Consumers. early as the presidential election of 1796 (see Adkins With the increased potential for misleading and misin- 2008). Similarly, how low literate voters acquire and use forming low literate adults, public policy makers should political information raises concern for highly complex be attuned to the marketing practices being employed by issues at the state and/or local levels such as ballot initia- candidates in the political marketplace. Academicians tives, referenda, and amendments to state constitutions or might research how low literate consumers interpret and city charters where voters must find information from use political marketing collaterals. Such research could print (often internet and direct mail) sources. In such aid in the development of additional federal guidelines to cases, participation in the political process could yield guard against misleading forms of communications. Un- significant impact or consequences for the low literate fortunately, the practice of distributing deceptive political consumer. References are available upon request. marketing collaterals in the United States dates at least as

For further information contact: Natalie Ross Adkins College of Business & Public Administration Drake University 2507 University Avenue Des Moines, IA 50311–4505 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 9 MARKETING SEX, DRUGS, AND DESPERATION: AN ETHICAL AND BI-CULTURAL ANALYSIS OF THE “METH PROJECT”

Matt Hettche, Virginia Tech, Blacksburg Joseph Scarpaci, Virginia Tech, Blacksburg

ABSTRACT BACKGROUND TO THE METH EPIDEMIC

This paper examines the ethical and cultural dimensions Methamphetamine is a psycho stimulant drug that is of a social marketing campaign called “The Meth Project.” pharmacologically prescribed to treat a narrow range of We examine some of the more pressing ethical objections disorders such as obesity, attention deficit disorder, and to the program’s marketing strategy, analyze its Hispanic narcolepsy. Although recently discovered to have a natu- (Spanish language) ads for cultural sensitivity, and dis- ral derivation, methamphetamine in its most prevalent cuss its implications for public policy. form is synthesized from chemicals that are readily available from a host of consumer products (Clement et al. INTRODUCTION 1998). Manufactured through a relatively simple procedure of chemical distillation, methamphetamine is one of Drug addiction and substance abuse are perennial threats the most addictive and destructive illegal drugs found in to public health and safety. The economic and human our communities. Its street names in English are meth, costs of these social problems not only have lasting effects crank, ice, and crystal and in Spanish, tina, hielo, tiza, and on individual lives but they also tear at the very fabric of vidrio. productive society. Understanding the role that marketing (and social marketing) can play in combating these corro- Meth’s key ingredient is pseudoephedrine (or ephedrine), sive forces is an important step to their elimination. the active ingredient of many mass marketed cold and allergy medicines. Before national legislation in 2006, Our focus in this short paper is with the ethical and cultural through a provision of the revised Patriot Act limiting the dimensions of the “The Meth Project’s” advertising strat- distribution of pseudoephedrine as a “behind-the-counter” egy. The Meth Project, originating in 2005 in Montana non-prescription drug, pseudoephedrine was widely avail- with a grant from a private citizen (Thomas Siebel), is a able to the consumer in virtually unrestricted quantities national drug awareness program that has the specific aim (USA PATRIOT Reauthorization Act 2006). The primary to reduce the incidence of first-time methamphetamine pharmacological effect of pseudoephedrine is the shrink- use among teens and young adults. Although meeting ing of nasal membranes that provides relief from sinus with initial success and even acclaim from law enforce- congestion. Typical users of the drug will ingest 120 mg ment for its positive impact on crime, some of the program’s and experience up to 12 hours of decongestion. For many techniques to communicate its core message have been users, there is also an increase in energy and focus expe- sharply criticized. The program’s use of fear, graphic rienced (e.g., similar to the stimulant effects of caffeine). depiction of violence, and preteen sexuality, for example, Since the late 1980s, however, at least in the United States, has some critics questioning the unintended consequences an increasing unintended use of pseudoephedrine has of the program’s advertising approach. been the production of meth.

The three questions that direct our research in this paper Meth is ingested in a variety of ways (e.g., injected, are as follows: (1) How should a social marketing cam- snorted, or smoked) and it is reported to give its users a paign that is designed to influence drug use in society be “euphoric rush that produces a tingling feeling throughout integrated into public policy decisions? (2) What is the the body that is happy and giddy,” and lasts anywhere potential collateral damage of using graphic fear and between 12 and 24 hours (Frontline 2006). Similar to shock-based persuasion tactics with a predominately ju- other hard drugs, such as heroin and cocaine, the effects venile audience? and (3) are there significant differences of coming-down from meth are both physically and psy- between Anglo and Hispanic speaking meth users when chologically taxing, often resulting in erratic and unpre- retelling their personal accounts of meth addition? dictable behavior. Because of meth’s crash, users can be

10 Marketing and Public Policy Conference Proceedings / 2009 hedonically prompted into a several day binge of use that demand. Mexican organized crime filled this production typically ends in complete exhaustion, mental depression, gap, and by the end of 2006, Mexico appeared to outpace and/or death. Prior to aggressive public policy measures U.S. suppliers of methamphetamine. In the former Meth in the last five years, coordinated at the State and National Mecca of Kansas, for instance, U.S. government seizures levels, meth abuse was quickly becoming the fastest of meth originating in Mexico soared over 1000 percent growing addictive drug epidemic in the nation. between 2004 and 2005. And although 22 pounds of Mexican methamphetamines were seized in 2004, a year The social costs of meth abuse are staggering. In the state latter that amount reached 271 pounds. International data of Montana, for example, “50 percent of [the] prison assembled by The United Nations Office on Drugs and population has been cited on charges of meth use or crimes Crime report that methamphetamine confiscation on the related to meth use,” and “50 percent of all children in U.S.-Mexican border reached 1.1 tons in 2002, 1.7 tons in foster care . . . are placed there because one or both parents 2003, and 2 tons in 2004. By 2007, it was safe to say that are involved with a meth addiction” (Meth Project News- 99 percent of Mexican meth was bound for the United letter 2007). And perhaps most disturbing, “77 percent of States. Tijuana is a principal transshipment point for U.S.- first time users of the drug are age 15 and younger” (Roper bound Pacific coast users and Michoacán serves as a 2008). Given the strong likelihood of causing serious major production center for drugs heading to the U.S. health problems and even death among its users, recent Midwest, southeast, and Atlantic seaboard (Logan and public policy efforts have explored several avenues to Kaires 2007). address the problem. By 2007, a dozen U.S. governors declared meth as their THE GEOGRAPHY AND DISTRIBUTION OF number one elicit drug problem (Table 1), and 40 states METH IN THE USA AND MEXICO enacted tougher meth laws. Unlike other drugs such as marijuana, cocaine, and even heroine, meth users are Methamphetamine use has soared in both the United usually so strung out that their rehabilitation costs fall to States and Mexico. Latin-American drug production and the state, whose already thinly stretched Medicaid pro- transit routes have targeted Mexico for its proximity to grams must incur the costs (Spencer 2006). U.S. consumers. Although the drug was little known south of the U.S. border in the 1990s, production in Survey data on illegal substance abuse is highly unreliable Mexico has climbed dramatically in the 2000s. The tradi- (either because respondents often are fearful to fully tional pattern of spread identified in the 1990s was from disclose their behavior and/or they misperceive the full California to the Midwest and beyond. A 2006 survey of extent of their use). Despite these limitations, however, 500 county law enforcement officials by the DEA in 44 there appear to be consistent differences among ethnic states reported that meth continues to be the number one groups in the USA. Roper (2007) survey data revealed drug problem – 48 percent of the counties report that meth that 71 percent of youth get their information on the evils is the primary drug problem, more than cocaine (22%), of meth from the web. Peers, however, are a significant marijuana (22%), and heroin (3%) combined. Up until source for getting the drug. One quarter of all Whites and 2005, methamphetamine production was scattered among Hispanics say it would be easy or somewhat easy to get local producers, or “mom and pop” labs, particularly in meth, versus 16 percent for Blacks. Germane to our study, rural areas in the Midwest states. For instance, Kansas the same survey reports that 14 percent of all Hispanics earned the reputation as the “Meth Mecca” because of its state that they have been offered meth, versus 10 percent large supply of meth users and producers (Logan and and 7 percent for whites and blacks, respectively (Roper Kairies 2007). In 2001, Kansas retailers responded by 2007, p. 51). establishing the Meth Watch Program. This public-private partnership clamped down on the unscrupulous sale There are also key differences in actual consumption by and stealing of everyday household products used to ethnicity. Whites are more likely than either African manufacture methamphetamine. So successful was the Americans or Hispanics to perceive risk in taking meth Kansas Meth Watch that it became a national model three once or twice a week (70%, 56%, and 59%, respectively) years later (Logan and Kairies 2007). or even regularly (89%, 69%, and 77%). Four of ten (42% of African Americans) and 37 percent of Hispanics mis- U.S. measures appear to have forced clandestine opera- takenly believe there is only a slight or no risk to using the tions to Mexico where regulations were more lax. Once in drug once or twice. As well, 28 percent of African Ameri- effect, the Combat Methamphetamine Act was effective cans think there is little risk in taking the drug fairly in curbed supply in the U.S., but had no effect on user regularly (Roper 2007, p. 51).

Marketing and Public Policy Conference Proceedings / 2009 11 TABLE 1 Selected Anti-Meth Prevention Strategies by State, 2006

State Administrator Anti-Meth Initiative

Georgia Governor Sonny Purdue Pledged money for rehabilitation and enforcement programs.

Idaho Governor Dirk Kemphthoren Anti-meth initiative shut down 760 labs.

Nebraska Governor Dave Heinmenan Reported a 60 percent drop in the number of meth labs during the first 90 days following the sale of pseudoephedrine.

Arizona Governor Janet Napolitano Requested a “tough meth bill” that she could sign.

New Mexico Governor Bill Richardson Developed statewide registry of properties affected by meth production so homebuyers and neighbors could be apprised of risk.

Hawaii Governor Linda Lingle Proposed $10.2 million to enhance “Healthy Lifestyles” and “About Face” programs to deter youth consumption and promote health life-styles.

Data source: Spencer (2006).

UN-SELLING METH: THE SOCIAL room for drugs, and repeats, “I’m gonna try meth just MARKETING STRATEGY OF once” (Oudesma and Wedell 2007). “THE METH PROJECT” Siebel’s strategy of using consumer-marketing techniques “The Meth Project,” perhaps best known for its graphic ad to address a social problem is certainly not new. There is, campaign that both features and targets teenagers, utilizes after all, a well established branch of marketing, known as the techniques of consumer marketing to battle the ill social marketing, which explores the application of using effects of meth abuse. Founded by Thomas Siebel in the traditional marketing mix (product, place, price, and Montana in 2005, the Meth Project aims to reverse the promotion) as tools to create positive social outcomes prevalence of meth use by impacting attitudes about the (Andreasen 2001). Although social marketing (itself) is drug. Siebel’s prior experience in marketing involves the ethically inert or value neutral, there are distinct theoreti- creation of a sophisticated consumer tracking software cal contexts where normative concerns factor in. Michael program (known as Siebel Systems) that was sold in 2005 Rothschild, for example, describes three levels where to Oracle for 5.8 billion dollars (PBS Now Interview, ethical concerns figure into a meta-evaluation of social March 2007). Issuing a grant of $2 million from his own marketing, and for our purposes in this paper, are particu- philanthropic foundation, Siebel spearheaded a market- larly helpful to review (Rothschild 2001). First, at the ing research project on how best to reach teens and young ground level, are questions about the point or purpose of adults about first-time meth use. a marketer’s goal and whether it, as a plan or ideal, can be ethically permitted. Like science and technological ad- The ads first aired in Montana in 2005 and consisted of 30- vance, marketing can be used for both noble and sinister second spots that chronicled the travails of teenagers who purposes. Examples of ethical concerns that surface at this declare in the first scene, “I’m gonna try meth just once.” ground level might include the promotion of a particular One girl promises, “I’m gonna smoke this just once.” religious or political ideology for self-interested reasons. Next, she says, “I’m gonna steal just once.” It is followed by, “I’m going to sleep with him for meth just once.” In the At the middle level of Rothchild’s analysis, ethical ques- last scene, her little sister furiously searchers through her tions about social marketing concern how marketing is

12 Marketing and Public Policy Conference Proceedings / 2009 utilized and juxtaposed with other social institutions to The Meth Project approaches methamphetamine as a influence human behavior. Specific questions to emerge consumer products marketing problem. Meth is a at the middle level include how marketers should both consumer product. It is readily available. It is reinforce and compliment existing institutions, such as affordably priced. It is distributed through an effec- law and education, within society. That is to say, how and tive, although alternative, distribution channel. It has to what extent does marketing remain subordinate to law product attributes that are perceived as attractive. and education and what exactly is the proper role of each Initial meth consumption decisions are made – simi- in influencing behavior? To help clarify the differences lar to other consumer product decisions – after an among these social measures, Rothschild offers the fol- evaluation (perhaps subconscious) of benefit, risk, lowing definition of marketing: social acceptance, price, and availability (Roper 2008, p. 7). Marketing refers to attempts to manage behavior by offering positive reinforcing incentives and/or con- The research techniques utilized by the Meth Project to sequences in the environment. The choice environ- develop its primary brand message also relies on tradi- ment is made favorable for appropriate behavior tional marketing communication strategy to an English through the development of alternatives with com- speaking audience. The guiding idea of the campaign is parative advantage (products and services), favor- that the destructive behavior of Meth use can be “un-sold” able cost-benefit relationships (pricing), and time to the highly impressionable and apparently tolerant popu- and place utility enhancement (channels of distribu- lation of teens and young adults. Beginning with a baseline tion). Positive reinforcement is given when a transac- survey of 1258 respondents (ages 12 to 24), Siebel and tion is completed (Rothschild 2001, p. 19). fellow researcher Siobhan O’Conner, a former marketing executive from Hewlett Packard, measured perceptions At the top level of his analysis, the ethical concerns are the and attitudes of Montana teens to determine how Meth techniques and tactics of marketing practitioners. Three compared with other illegal recreational drugs (Oudsema prominent issues often debated at this level include: (1) and Wendel 2007, p. 71). After finding an exceptionally What role should fear play in promoting a certain desired high rate of accepting attitudes about first time recre- behavior or outcome; (2) Is false information and deliber- ational use, Siebel and his team began to shape its core ate misinformation ever ethically justified; and (3) To message through input gained from focus groups and what extent should marketers preserve and respect con- interviews. From the emergent themes of violence, sex sumer privacy when collecting information? While gen- abuse, stealing, and family dissolution, two messages eral philosophical questions about the ethics of illegal consistently tested well among the representatives of their drug use certainly have interesting implications for a targeted population: “Not even once” and “This isn’t discussion of the Meth Project, a concern that plausibly normal, but on Meth it is.” occupies Rothschild’s “ground level,” the ethical focus we attend too here in this paper is primarily with issues at Driving the theoretical framework of the Meth Project’s the middle and top levels. research methodology is the work of Catalano and Hawkins, from the University of Washington’s Social Siebel’s approach to social marketing in the Meth Project, Development Research Group (Roper 2008). Catalano as evidenced in his flagship program the Montana Meth and Hawkins believe that communication with teenage Project, is innovative in a number of respects. Notewor- populations on the destructive aspects of drug use must thy, at least in hindsight, was Siebel’s hypothesis that meet two conditions: (1) the message must come from a Montana, given its size, relative isolation and demo- credible source and (2) the message must in some way ally graphic makeup, could serve as a predictive test-market itself with existing currents of social approval/disap- for a larger national anti-drug advertising campaign. proval. Siebel’s idea here was to develop a baseline measure of attitudes and message effectiveness that could be con- The first generation product of Siebel’s research is a series structed, revised, and implemented in a relatively short of advertisements distributed through traditional medi- amount of time, while at the same time, have an immediate ums. The Meth Project employs what it calls “saturation- impact when broadcasted to a larger population. Argu- level” advertising which is a simultaneous broadcast of ably, the strategic assumption motivating Siebel’s ap- television, radio, internet, and print ads, that aims to reach proach is how illegal methamphetamine use is identified the largest audience through several communication chan- as a consumer products marketing problem. The position- nels. In 2007, for example, the Montana Meth Project was ing of this assumption is described in the Meth Project considered the largest advertiser in the state, estimated to National Survey 2008 as follows: have reached “70 to 90 percent of its teen audience three

Marketing and Public Policy Conference Proceedings / 2009 13 to five times a week” (Meth Project Newsletter 2007). program is planned and carried out. Cultivating healthy According to the Montana Meth: Use and Attitudes Sur- relationships between study subjects and researchers that vey 2008: “Anti-meth advertising is ubiquitous in Mon- extend beyond the information gathering stage of a project tana, having been seen or heard at least once a week by to a more long-term partnership (where cultural, educa- two-thirds (66%) of teens. Almost all teens (88%) report tional, and economic resources are shared) is one ex- seeing or hearing an ad in the past month” (Roper 2008, ample. While there are dimensions of the Meth Project p. 5). that certainly resemble a PAR approach (most notably the “Paint the State” art initiative and talent contest), there are The graphic depiction of the harm suffered by meth- arguably additional ways the Meth Project can expand in addicts, utilizing the persuasion tactics of fear and shock, this capacity and converge with a PAR research paradigm. turns out to be particularly effective among teens. Not only have teens long been considered loyal consumers of TWO ETHICAL OBJECTIONS horror and action adventure films, but given the developmental changes that accompany adolescence, it is be- The persuasion tactics of the Meth Project are in some lieved that teens tend to steer their attention to issues of sense double-edged. On the one hand, utilizing themes of body image and external beauty more so than the general violence, sexual abuse, rape, and physical demise are population. The graphic and visceral nature of the Meth highly effective as “attention getters” and preambles to Project’s ads certainly plays to this profile. In fact, initial attitude formation. This is particularly the case for the results and statistical feedback measures from the Mon- Meth Project’s target audience. The appeals presented in tana Meth Project suggest the approach is having a posi- the Meth Project ads are rooted in fear and represent basic tive impact (Meth Project Newsletter 2007, p. 2). threats to human well-being and happiness. Any of the nightmare situations depicted in the ads, if were actually In an interview with PBS, Montana Attorney General, to occur in one’s life, would have long-lasting, if not Mike McGrath, maintains that the presence of the Meth permanent, consequences. Yet, on the other hand, be- Project in his state has been instrumental in drawing cause the ads play so heavily on the persuasion tactics of attention to the meth problem and motivating action fear and shock, there are genuine questions about what among legislators and community leaders. McGrath states: other effects the ads may precipitate. That is to say, the “[The Meth Project’s advertising campaign] isn’t a pro- measures of success that the Meth Project has detected for gram that could have been done by government or com- curbing first time meth use may in fact be only telling a mittee” . . . [given the edgy nature of the ads]. . . . “I can’t partial story. Worrisome, therefore, are the unintended imagine a school board or legislative committee that consequences the Meth Project ads may have. Our focus would put together an ad campaign like this” (Now 2007). in this section is to raise two worries that follow from an For our purposes, McGrath’s insight reveals an important a priori analysis of what the Meth Project attempts to do connection between public perception and social change. by using fear and shock as advertising appeals. McGrath’s point is that antecedent to the innovative programs of reform to take off in his state, such as new Before presenting the details of our “unintended conse- drug treatment facilities, drug courts, and tough legisla- quences objections,” it is perhaps important first to clarify tion curbing the sale and distribution of pseudoephedrine, the theoretical assumptions that we believe ground or it was necessary to have a shift in the public’s attitude justify them. What we are not attempting to argue, for about the importance and urgency of the meth epidemic. example, is that the Meth Project ads are somehow ethically objectionable by the mere presence of violent, shock- The tight integration of the Meth Project’s advertising ing, or sexually themed images. That is to say, it is not the campaign with public policy initiatives make it an ideal mere content of the ads that may perhaps violate a person’s candidate for what is recently identified in the academic religious or social sensibilities that we premise or voice marketing literature as Participatory Action Research our objections. There is little question, it can be argued, [PAR hereafter]. PAR assumes an integral relationship that the content of the ads is not suitable for all viewers and between the goal of improving consumer welfare, on the the ideas and images often prove to be either offensive or one hand, and the research methods employed to gain inappropriate for some audience members. Although this knowledge of consumer preferences, attitudes, and be- type of worry has merit, and should perhaps be explored havior, on the other (Ozanne and Saatcioglu 2008). A within a broader context of how violence and human central premise to the approach is that consumer well- sexuality are integrated into public service messages, this being is not a static variable to be measured once through- is not an angle of concern we will address. Our primary out the research process, but rather is a dynamic variable concern is with the potential additional effects or unin- that can be significantly impacted by how a research tended consequences the ads may have. The ugly images

14 Marketing and Public Policy Conference Proceedings / 2009 and brutality depicted in the ads are undoubtedly objec- Since the Meth Project ads are so widely broadcast to a tionable, at least on some level, but most viewers presum- target demographic that closely borders the children seg- ably tolerate them because of the impact and positive ment, it also stands to reason that young children are the social good that results. Our focus is to press and question unfortunate recipients of these graphic messages. The whether such tolerance is really justified in the end. compromise of a child’s imagination, however, is only part of the worry. The imaginations of teens and young The first objection focuses on the effects of a sustained adults are also subjected to graphic and vivid ads that elicit advertizing campaign that uses fear and violence to im- strong emotional and ethical evaluations. If the traumatic pact a teenager’s attitudes. We call this objection the scenes are not properly situated into a boarder context of “imagination assault objection.” The basis of the worry socially responsible behavior, there is a potential that the pulls from two insights, one ancient and the other modern. regularity of the ads will have a dulling effect on the way The first insight comes from Aristotle’s Rhetoric and it a person reacts to depraved, abusive and mentally ill amounts to the claim that whenever we entertain a dra- behavior. A quick comparison with pornography might matic representation there is a momentary suspension of help reinforce the point. There is a fair amount of empiri- disbelief, which involves an imaginative engagement and cal evidence that suggests as consumers of pornography powers of emotional and moral evaluation. How we grow familiar with certain depictions and representations, psychologically react to a particular dramatic representa- there is a tendency to seek out more and more radical tion, including how far we are willing to extend our moral depicts of the medium (Eaton 2007, p. 713). Similar approbation/disapprobation, is ethically relevant insofar observations are cited with the use of fear in advertising, as it conditions and trains our emotions for the future. In suggesting that there is often significant interest in fear Aristotle’s view, we desire to have a strong and well- driven ads at first, but its effect is limited over time (Ray adjusted character so that we appropriately respond to the et al. 1970; Sternthal et al. 1974). If the Meth Project is right events and objects in our life with the proper inten- affecting teen and young adult attitudes about meth use, sity. we might pause to consider if the assault on the imagination is actually worth it. The more modern insight for our objection comes from the guidelines established by the Council of Better Busi- The second objection to point out has to do the collateral ness Bureau’s Children Advertising Review Unit (CARU). damage of how the Meth Project ads might unwittingly In the 2006, eighth edition of the guidelines, under the affect attitudes about the subornation of women and section “Product Presentation and Claims,” the guideline children (as sexual objects). While we do not deny that reads: female meth users and female children meth users, in particular, are vulnerable to sexual abuse, it is somewhat Claims should not unduly exploit a child’s imagina- biased of the advertising campaign to simply highlight tion. While fantasy, using techniques such as anima- these relationships. This objection of “unequal time for tion and computer-generated imagery, is appropriate sex crimes against female and children” essentially ques- for both younger and older children, it should not tions why there are a disproportionate number of ads create unattainable performance expectations nor targeting female and teen sexuality. If the ad campaign is exploit the younger child’s difficulty in distinguish- really addressing realities of meth use rather than sensa- ing between the real and the fanciful (CARU 2006). tionalizing certain types of sex crimes, we might expect to see a more representational (i.e., gender neutral) presen- The basis of the guideline is that children, in particular, tation of sexual compromise. Certainly one opportunity of have difficulty understanding the difference between fic- the Meth Project to embrace a more PAR research ap- tion and reality and so special care should be exercised proach/perspective would to be elicit a dialogue with when presenting merchandise and ideas to them. The role teenagers and young adults about not only illicit drug use of the imagination is extraordinarily important for learn- but also the issue of female subornation and the recent ing and there is growing evidence to emerge in the past trend in western culture to sexualize childhood. few decades that “imaginative play” is necessary for the development of speech and abstract reasoning. If the SOCIAL MARKETING TO HISPANICS development of language and other cognitive skills are closely aligned with a child imaginative capacity, it stands Guendelman and Wagner (2000) have shown a shift in to reason that a precautionary principle should be en- how Hispanics interact with the U.S. health-care system, dorsed when targeting certain advertising messages to both on the curative and preventive sides. Initial research children. suggests that demographic traits and cultural beliefs de-

Marketing and Public Policy Conference Proceedings / 2009 15 termined access into the system. However, research in the and Latino target markets. Our sampling frame consisted 1990s and 2000s shifted the paradigm to one in which of nonprofit and government agencies that produced health delivery system barriers, access to insurance, and English-language ad campaigns with some Spanish-lan- low socio-economic status are better predictors of His- guage translation or separate Spanish production. Data panic health-care accessibility (Guendelman and Wagner came from Mothers Against Meth (www.MaMa.org) who 2000, p. 19). They argue, “Studies consistently show that sell four videos, only two of which are available in Hispanics are less likely than any other group to be linked Spanish, the Montana Meth Project, the Arizona Meth to a regular source of care” (Guendelman and Wagner Project, and several federal websites. All but a handful of 2000, p. 19). Not surprisingly, Hispanics with English- spots were directed to English-speaking audiences. An speaking abilities were more likely to interact with the appropriate proportion given the demographic profile for health-care system than those without language skills (p = unselling meth in both regional and national markets (e.g., 0.05), underscoring the importance of language-appro- there are simply more primary English-speakers at risk). priate media in targeting unhealthy behaviors. The au- The pioneer site, The Montana Meth Project, had only one thors conclude that “culturally appropriate personal inter- 30-second video titled “Crash” (Acidente) out of 17, and actions” are paramount to monitoring the health care it was a verbatim translation of its English-language status of Hispanics in the USA (Guendelman and Wagner source video, which aired first (Figure 1). Video content 2000, p. 44). and use of Mexican-American pronunciation by the pro- tagonist (victim) were all appropriate. The only weakness Consumption of methamphetamine during pregnancy is we observed was the use of the project logo that appeared of concern because of the drug’s prevalence among women at the end (Not even once) that lacked a parenthetical of childbearing ages. Chronic users suffer permanent translation (ni una vez). changes in brain functions. Studies of rat brains exposed to the drug indicate serious implications for human fetal The Arizona Meth Project carried a different tact; 30 brain development (Volkow and Lee 2007, p. 273). More second radio ads, seemingly unscripted comments by broadly, the knowledge of maternal transfer of drugs in English- and Spanish-language youth, from age 15 to 21. utero remains a perennial public health challenge. Given Our inquires to the ad agency (Revolucion.com) about the that most Hispanics in the American west are of Mexican ideas behind the project, the level of editing that was done, origin, we reviewed the literature on meth use among or whether the meth users were prompted or not with Mexican women. Guicha-Alba and Miranda-Chavez specific questions, went unanswered. Therefore, we con- (2007) found that one-third of Sonoran pregnant women ducted a content analysis of these testimonies, looking for at the state’s largest hospital suffered from the abstinence latent and frequency content of terms that are germane to syndrome: Some mothers assumed that merely stopping the unselling of illicit drugs. We collapsed categories into use of meth during the moment they found out they were broader categories and then conducted a Chi-square test to pregnant until the delivery would relieve the fetus or determine whether there were significant differences in newborn of all ill-health effects. They did not consider the how the Anglo and Latino youth described their torment time between conception and the pregnancy notification, with meth. We arranged our categories based on a simple nor the role of lactation in transferring meth to the new- locus of control dichotomy: primary groups (individual born. and family) and secondary groups. Coding measured the number of times references (Table 2) were made to The model of health promotion to alter the informational individual responsibility or shortcomings and pain in- environment has long been a hallmark of the United States flicted on family members (primary groups), and troubles and Western Europe. Of course, “not everyone believes with society at large (secondary groups, or employers, that public funds should be expended, or the veneer of police, school). Findings show that Anglos reported more government legitimacy used, to prescribe particular social primary group problems (77.5%) than did Latinos (55.5%) orthodoxies” (Gostin and Stone 2007, p. 71). To date, (Figure 2). Latinos, on the other hand, were almost equally little policy or content analyses have examined the effi- concerned about secondary-group problems such as cacy of reaching at-risk Hispanic drug users, nor has these schools, jobs, and the police than were Anglos (Chi- been a full vetting of the ethical implications of using Square = 4.599, p = .034). “shock” advertisements in unselling meth.

We surveyed that extant public information spots that We next conducted a t-test on the phrases that we coded appeared on television, radio, print, and Internet with an to determine whether were significant differences in the eye toward significant possible differences between Anglo kinds of words used by Latinos and Anglos in their

16 Marketing and Public Policy Conference Proceedings / 2009 FIGURE 1 A Frame from the Arizona Meth Project TV Spot “Crash” (Acidente)

TABLE 2 Coding Used Strips in Content Analysis Radio Spots, in English and Spanish (English Translations and Ideas Shown Here) almost died introduced others to meth pawning school problems can’t use just once job problems personal responsibility shooting up family hurt loneliness responsibility STDs/Hepatitis family introduced to lost friends picked blemishes stealing meth memory loss police problems suicidal thoughts family theft meth steals your self pressure to use teeth problems guns morals public shame tracks headache paranoia resort to prostitution violence heart palpitations impure ingredients running away

Data source: Arizona Meth Project. Coding conducted by authors.

Marketing and Public Policy Conference Proceedings / 2009 17 testimonies/radio spots. A pooled Student t-test value of secondary groups has not been examined in the social 2.36 was significant at the 0.036 level. marketing literature, and is worthy of future research. To be sure, there are patterns of behavior in the assimilation We conclude that there are significant differences in the process that transcend the use of English-language skills, ways Anglos and Latinos express their journeys through dress, public displays of emotion, and personal values the methamphetamine nightmare. Despite the highly ille- (Gordon 1964, p. 79). If the sample of Hispanic and Latino gal nature of meth use, we attribute the Hispanic preoccu- accounts of the anti-meth radio spots are both unscripted pation with secondary group problems as a part of their and representative of meth users’ attitudes toward the assimilation desires and the norms and expectations that drug, then it would be well for focusing on the extrinsic residency and citizenship entail (Alba and Nee 1997). forces of secondary group metrics of behavior (complet- Social theorist Milton Gordon (1964) differentiated being school, maintaining a job, not getting in trouble with tween acculturation and “structural assimilation.” In the the law) in the social marketing media directed at unselling latter, he referred of the entry of ethnic minorities into meth to Hispanics. primary-group relationships with the cultural majority. Such a distinction, with its focus on the nature of an IMPLICATIONS AND CONCLUSIONS FOR immigrant’s primary-group association, however, indi- FUTURE RESEARCH cates one of the weaknesses of Gordon’s scheme: It is a microsociological interpretation that is not theoretically Social marketing aspires to unsell the consumption of integrated into broader social forces such as the fluidly of goods and services. In so doing, it raises several ethical ethnic boundaries (Barth 1956). Our data sources, for and cultural issues. Our assessment of regional and na- instances, do not reveal whether the Latino speakers are tional marketing strategies of the Meth Project questions first- or second-generation American residents. Never- the appropriateness of using fear and shock toward juve- theless, Gordon’s differentiation between primary and niles. Specifically, the portrayal of violence and sexual

FIGURE 2 Content Analysis Results from Anglo (English-Language) and Latino (Spanish-Language) Anti-Meth Radio Spots According to Speakers’ Reference to Primary and Secondary Groups

18 Marketing and Public Policy Conference Proceedings / 2009 exploitation in several TV spots begs the question: Is groups than did Anglos. This suggests a one-model-fits- raising the level of shock to such an extreme an ethical and all marketing strategy may be ill advised given our explor- effective means of unselling meth? While we have only atory research. Drug prevention strategies that more heavily begun to grapple with these issues, our ethical analysis rely on more traditional values (i.e., parents, siblings, about meth advertisements suggest that they are perhaps family life) may be misguided. Our preliminary findings too taxing on the imagination and contribute to negative also show that Anglos meth users are significantly more attitudes about gender inequality and teenaged sexuality. concerned about primary group relations than their His- panic counterparts are. This finding challenges the stereo- Our analysis of thirteen Arizona Meth Project radio spots type that Anglos are “colder and more detached about show there are significant differences in how Anglo family and family relations” than Hispanics. Future re- (English speakers) and Hispanic (Spanish speakers) meth search will require a larger sample size of media promo- users describe their experiences. Surprisingly, Hispanics tional spots as well as longitudinal and controlled studies voice more concern about violating norms of secondary to tease out these effects.

REFERENCES Chávez (2008), “Informe Preliminar sobre Abuso Fetal en el HIES 2002–2007,” Boletín Clínico del Alba, Richard and Victor Nee (1997), “Rethinking Hospital Infantil Eduardo Son, 24 (1), 15–21 Assmilation Theory for a New Era of Immigrantion Logan, Sam and Kate Kairies (2007), “U.S. Drug Habit Policy,” International Migration Review, 31 (3), (Win- Migrates to Mexico,” AlterInfos, Accessed October ter), 826–74. 23, 2008 [http://www.alterinfos.org/spip.php]. Barth, F. (1956), “Ecologic Relationships of Ethnic Groups Meth Project (2007), Meth Project Newsletter. (Summer). in Swat, North Pakistan,” American Anthropologist, Meth Project (2008), Meth Project Newsletter. (Summer). 58. 1079–89. Now PBS, Show 335 (2007), Accessed November 12, CARU (2006), The Children’s Advertising Review Unit 2008, [http://www.pbs.org/now/shows/335/video. Self-Regulatory Program for Children’s Advertis- html]. ing, Assessed November 12, 2008 [http:// Ozanne, Julie L. and Bige Saatcioglu. (2008), “Participa- www.caru.org/guidelines/guidelines.pdf]. tory Action Research,” Journal of Consumer Re- Clement, B.A. and C.M. Goff, and T.D.A. Forbes (1998), search, 35 (August). “Toxic Amines and Alkaloids from Acacia Rigidula,” Ouesma, Suzy and Rick Wedell (2007), “Unselling Meth: Phytochemistry, 49 (5), 1377–80. A Graphic Ad Campaign Shocks Teens Away from Eaton, A.W. (2007), “A Sensible Antiporn Feminism,” Drugs,” Stanford Social Innovation Review, (Sum- Ethics, 117 (3), (July), 647–715. mer). Accessed September 6, 2008, [http://www. Frontline, PBS (2006), The Meth Epidemic, written and ssireview.org/images/articles/2007SU_whatworks_ produced by Carl Byker (originally aired February meth.pdf]. 14, 2006). Ray, Michael and William L Wilkie (1970), “Fear: The Gordon, Milton (1964), Assimilation in American Life. Potential of an Appeal Neglected by Marketing,” New York: Oxford. Journal of Marketing, 34 (January), 54–62. Gostin, Lawrence and Leslie Stone (2007), “Health of the Roper Public Affairs & Media (2007), National Meth Use People: The Highest Law?” in Ethics, Prevention, & Attitudes Survey. (September). and Public Health, Angus Dawson and Marcel ______(2008), Montana Meth Use & Attitudes Verweij, eds. Oxford, 59–77. Survey. (April). Guendelman, S. and T. Wagner (2000), “Hispanics’ Ex- Rothschild, Michael (2001), “Ethical Considerations in periences Within the Health Care System,” in Minor- the Use of Marketing for the Management of Public ity Health in America: Findings and Policy Implica- Health and Social Issues,” in Ethics in Social Market- tions from the Commonwealth Minority Health Sur- ing, Alan R Andreasen, ed. Washington, DC: vey, C.J.R. Hogue, M.A. Hargraves, and K.S. Collins, Georgetown University Press. eds. Baltimore and London: Johns Hopkins Univer- Sternthal, Brian and C. Samuel Craig (1974), “Fear Ap- sity Press, 19–46 peals Revisited and Revised,” Journal of Consumer Guicho-Alba, Evangelina and Víctor Manuel Miranda- Research, 1 (December), 22–34.

Marketing and Public Policy Conference Proceedings / 2009 19 Volkow, N.D. and T.K. Li (2007), “Drugs and Alcohol: M.T. Tsuang, W. Stone, and M.J. Lyons, eds. Wash- Treating and Preventing Abuse, Addiction, and Their ington, D.C. American Psychiatric Publishing, 297– Medical Consequences,” in Recognition and Preven- 318. tion of Major Mental and Substance Use Disorders,

For further information contact: Matt Hettche and Joseph Scarpaci Department of Marketing Pamplin School of Business Virginia Tech Blacksburg, VA 24061 Phone: 540.231. 9621 Fax: 540.231.3076 E-Mail: [email protected]

20 Marketing and Public Policy Conference Proceedings / 2009 REGULATING PRODUCT PLACEMENTS IN CHILDREN’S ENTERTAINMENT: COGNITIVE AND BEHAVIORAL EFFECTS OF PLACEMENTS ON CHILDREN

Haiming Hang, Lancaster University, United Kingdom Susan Auty, Lancaster University, United Kingdom

ABSTRACT vulnerable to commercial messages even if they understand the concept of persuasion (Moses and Baldwin Our research indicates that product placement has an 2005). Moreover, according to Moore and Rideout (2007), effect on children even after brief exposure. When back- little consideration has been given to children’s under- ground placements are used, children are likely to be standing of and vulnerability to embedded marketing influenced under their awareness threshold. When promi- techniques like product placement in media other than nent, children are more likely to process placements TV, although some evidence of effect is now emerging consciously, with limited preference effects. Implications (Mallinckrodt and Mizerski 2007). In order to provide for policy makers are discussed. more insight into this issue, two experiments were done with children to explore the effects of different types of INTRODUCTION product placement within a film. Unlike TV programs, films on DVD, like advergames, are likely to be viewed With the FCC currently collecting evidence on embedded many times. advertising in children’s TV programming, the issue of product placement within children’s entertainment media Current Regulations on Product Placement is clearly a live one. The Campaign for a Commercial- Free Childhood (CCFC 2008) points out that the practice Compared with their counterparts in Europe, the United is widespread and increasing, calling for “an explicit States has not regulated product placement despite the ban . . . in all children’s programs” (p. 20). But will shift in communications investment from traditional ad- banning the practice on TV be enough? As Moore and vertising to product placement (Hood 2005). The FCC has Rideout (2007) point out, advergames are enthusiastically requested comments on the practice of embedded adver- filling the gap, and children’s films on DVD are another tising, thereby indicating a possible need for legislation source of repeated exposure to brands. Moreover, as (MB Docket 08–90 cited in CCFC 2008). Commercial Moore and Rideout go on to point out, defining brand Alert’s proposal of pop-up notification of the paid nature appearances can be a contentious issue, with advertisers of product placement whenever a brand appears in the seeking ever more creative and subtle ways to embed their program seems impractical and disruptive; however, there products within children’s entertainment. is no agreement about whether disclosing the sponsor of the placements at the beginning of the program would be There may be a need for control across media to protect a sensible solution. Opponents of this suggestion have children from these subtle forms of advertising. The argued that viewers could still dip in and out of the question is, do children respond to very subtle place- program and miss the notification (Hood 2005). CCFC ments? And, given their capacity – albeit incomplete – to (2008) notes that disclaimers would effectively advertise understand the persuasive nature of ads (Owen et al. twice to children, once when the disclaimer was shown 2007), do children respond differently to obvious place- and once when the placement appeared, acting as a prim- ments than they do to ads? In order to provide more ing agent rather than a warning. information to policy makers, two experiments were done with children to explore the effects of, first, subtle product THEORETICAL BACKGROUND placement and, second, the effects of prominent placements compared to ads for the same product. Previous Product Placement Research

Academic research on how to measure the behavioral Although Model T Fords appeared in Mack Sennett effects of embedded advertising is sparse, especially with comedies as early as in the 1910s (Lehu 2007), the extant regard to children who are believed to be especially literature on product/brand placement is just over a de-

Marketing and Public Policy Conference Proceedings / 2009 21 cade old. Researchers have documented viewers’ ethical elicited by the fluency as brand preference (Lee 2004). acceptance of product placement (e.g., Brennan, Such effects are strongest when consumers lack cognitive Rosenberger, and Hementera 2004; Gould, Gupta, and capacity, as in the case of children, or they lack the Krauter 2000; Gupta and Gould 1997; McKechnie and motivation to process the stimulus in sufficient detail to Zhou 2003; Tiwsakul and Hackley 2005); practitioners’ make a judgment (Winkielman et al. 2003). According to views of product placement effectiveness (e.g., Karrh, the processing fluency account, prior exposure to brands McKee, and Pardun 2003); and have focused on cognitive in the media would help children more easily to process results like brand recall and recognition to measure its the stimulus when reminded at test (Auty and Lewis effectiveness (e.g., Babin and Carder 1996a, 1996b; 2004). Thus, the subjective ease of processing a stimulus Brennan, Dubas, and Babin 1999; Gupta and Lord 1998; could lead children unconsciously to misattribute their Karrh 1994; Law and Braun 2000; Morton and Friedman fluency as brand preference and lead to subsequent choice. 2002; Russell 2002). Even among the limited number of studies, the results are still inconclusive. For example, Advertising to Children some have reported that the effectiveness of the placement is moderated by its size (Gupta and Lord 1998; Schneider During the past decades, traditional advertising to chil- and Cornwell 2005) while others have not found this dren has received increasing attention (Roedder John (Nelson 2002). Law and Braun (2000, 2004) argue that 1999), and an impressive body of research has accumu- the contradictory results of previous research into product lated on this issue (see Wright, Friestad, and Boush 2005 placement occurred because the studies employed only for a review). However, in the case less traditional forms explicit tests of memory that are designed to measure such as product placement, the work of Auty and Lewis cognitive results (brand recall and recognition). As Nelson (2004) appears to be the first published research focusing (2002) pointed out these measures seem appropriate ways on children. Children who were exposed to a film clip given that increased brand awareness is one of the goals of showing a brand of soft drink were more likely to choose product placement. However, Law and Braun (2004) that brand than a control group was. When the clip acted argue that these explicit tests of memory fail to measure as a reminder of a prior exposure to the film, compared the implicit memory from incidental exposure. Different with a single exposure or prior exposure without a re- measures are required to detect implicit memory, as expo- minder, the effect was even stronger, suggesting that sure may influence consumer brand preference without repeated viewings compound the processing fluency. any recollection of the exposure (Lee 2002; Shapiro 1999). More recent work by Mallinckrodt and Mizerski (2007) looked at the effect on children aged 5–8 of playing Implicit Memory and Processing Fluency advergames. They found that regardless of the children’s understanding of the game’s persuasive intentions, chil- Recent developments in cognitive psychology suggest dren who played the game showed a significantly greater that consumer decisions may be influenced in many ways, preference for the brand – Froot Loops – than those in the some of which do not need explicit memory of the previ- control group. The difference was greatest in the 8-year- ous exposure (e.g., Janiszewski 1993; Shapiro, MacInnis, old children, who should be more advanced in their and Heckler 1997). While explicit memory refers to a information processing (Roedder John 1999). So the person’s conscious recollection of previous events to evidence to date suggests that children are responding to perform certain tasks, which may be tapped by direct tests these covert messages and would do so even if warned. like recall or recognition (Lee 2002; Shapiro and Krishnan 2001), implicit memory is characterized by an enhance- Research by Braun-LaTour and LaTour (2004) suggests ment on the performance of a subsequent task that may be that childhood exposure to advertising images leads to traced to a recent stimulus exposure, even if a person is not long-lasting positive associations toward a brand, and aware of it (Lee 2002). One explanation for the enhance- even source confusion when another brand later uses a ment is processing fluency, or the subjective ease of similar message strategy. These positive feelings of warmth processing stimulus information. Processing fluency has “can remain relatively permanent and influential” (p. 57), been found to bias consumer judgment as it appears to be implying that when this exposure occurs in childhood hedonically marked; that is, the sense of familiarity gen- negative deliberative thoughts are less likely to interfere erated by the stimulus is interpreted as a positive feeling with the associative learning than when exposure occurs (see Schwarz 2004; Winkielman et al. 2003 for a review). in adulthood. Given the age profile of games players, and Unaware of the source of fluency, consumers uncon- the familiarity of brands such as McDonald’s and 7 Up sciously misattribute the positive evaluative response that are integrated into games and films viewed by chil-

22 Marketing and Public Policy Conference Proceedings / 2009 dren, policy makers need to understand the effect on these and internal reliability has been widely approved (e.g., less savvy consumers, especially given the prevalence of Akhter, Andrews, and Durvasula 1990). However, as it soft drinks and snack food brands in such placements. was originally written in English, it was translated into Chinese and then it was back translated by three other RESEARCH OVERVIEW Chinese adults who were unaware of the purpose of this research in order to validate that the original meanings of Little is known about how different types of visual place- the scale items were retained. Finally, the Chinese version ments are processed by children: is there a difference questionnaires were shown to the school teachers to between more or less prominent placements in the level of ensure that children at the age of 9–11 could understand consciousness and choice? Fluency theory suggests that the vocabulary. less obvious placements might be more effective than obvious ones because of the positive affect that arises In the scale, children were asked to rate the following from an unconscious misattribution of liking, which oc- product categories on six measures of relevance to their curs only when the memory is implicit, not explicit. Thus, lives: TVs, laptops, soft drinks, fast food restaurants, two experiments were carried out to investigate types of sports footwear, and chocolate bars. Sports footwear and visual product placements (background versus promi- fast food restaurants were chosen because they achieved nent) in the same medium (film) on children, with specific the highest mean among all product categories in the scale regard to the difference between the level of recall and the (Footwear: M = 3.6 out of 6; Fast Food Restaurant: M = level of choice. The effect of prominent placements within 2.96 out of 6). In order to avoid ceiling and floor effects, an engaging context is also compared to that of advertis- a new pre-test was done on another 43 children from the ing, in particular advertising that is highly engaging and same sample pool to choose the appropriate brands, that mood-enhancing (i.e., transformational) rather than merely is, well known but not market leaders. In the pre-test for presenting information (informational). The focus was on the sports footwear the brand share for Nike was 40 children aged 9–11, still in the analytical stage of develop- percent, Adidas 49 percent, and other brands 11 percent; ment, but old enough to understand the persuasive intent for the fast food restaurant the brand share for Pizza Hut of advertising (Owen et al 2007) and to have a wide was 28 percent, McDonalds 28 percent, and KFC 44 knowledge of brands. percent. Therefore, Nike and Pizza Hut were chosen as the brands for the experiments. Sample and Stimulus Selection EXPERIMENT 1 Since no research of implicit effects of advertising to our knowledge has been carried out in China, a country with The purpose of Experiment 1 was to explore the effect of a vast population of children, the research site of Shanghai visual product placement on children when it is loosely was believed to have intrinsic interest. The first author had connected to the plot. According to Russell’s (2002) connections with schools there, and thus arranged with Tripartite Typology of Product Placement, when visual teachers at several schools with similar socio-economic placements are loosely connected to the plot such as props intakes (lower to middle class) to conduct the research. shown in the background, they are perceived as congruent The teachers had full knowledge of the research at every placements as they are natural and can enhance the realism stage and acted in loco parentis as is the usual case in of the film (see, for example, Gould, Gupta, and Krauter China. Children were offered the opportunity to withdraw 2000). Under this condition, children are more likely to without penalty at each stage and were kept informed process the brand at relatively low levels of conscious- about procedures and the general intentions of the re- ness, thus leading to limited brand recall (Balasubramanian, search. Karrh, and Patwardhan 2006).

The target brands were selected on the basis that they Roedder John (1999) suggested that children in the ana- should not only be relevant to children at the age 9–11 but lytical stage (7–11 years old) have limited knowledge of also children at that age would be likely to have some basic products in the market and only process brands at a knowledge of different brands in the product category (as perceptual level when making judgments and choices they would be asked to make a choice among different (Achenreiner and Roedder John 2003). Furthermore, pre- brands in the same product category). A pre-test was done vious research suggests that when choice tasks are very on 36 children in the same sample pool as the main complex and people have limited knowledge on the sub- experiments on the product involvement scale developed ject, they resort to heuristics to make a judgment (e.g., by Zaichkowsky (1985) to identify the appropriate prod- Bettman, Luce, and Payne 1998). Therefore, when asked uct categories. This scale was chosen because its validity to make a choice among different brands in the same

Marketing and Public Policy Conference Proceedings / 2009 23 product categories, with their limited brand and product Previous researchers have argued that most experiments knowledge, children would rely on simple cues to make a that have tried to demonstrate an unconscious effect were choice. In this situation, we expect the hedonically marked contaminated by conscious effects (or vice versa), as both processing fluency arising from prior exposure to be of them could contribute to overall performance on cer- misattributed by children as brand preference and lead to tain tasks. This research therefore followed Shapiro (1999) choice. in using an exclude-recollect procedure to separate conscious effects from unconscious effects (based on the Experimental Design and Sample procedures of Curran and Hintzman 1995). In other words, children were asked to choose a brand if they preferred Fifty-five children (aged 9–11) from one primary school this brand to others, but not if they remembered seeing it in Shanghai were recruited for this experiment. Each child in the film clip (exclude procedure). In the recollection was randomly allocated by enrollment number by their procedure, children were asked to write down any brands teacher to watch either a short movie clip with product they specifically remembered seeing. placement or a movie clip without product placement. The dependent variable was children’s choice and brand re- RESULTS call, with exposure to product placement as the independent variable. Gender was elicited as a possible moderat- Manipulation Check ing variable. Children’s understanding of the exclude-recollect proce- Materials and Procedure dure was checked before the data analysis. No child who recalled seeing a brand in the recollect procedure chose Children in groups of 10–12 were asked to watch the the same brand in the exclude procedure or vice versa. movie clip with or without product placement in their Therefore, all the children’s responses were retained for school computer clusters for about 2–3 minutes. In the the analysis. computer cluster, there were about 60 computers with 6 in a line. After children entered the computer room, they No three or four way interactions among all the variables were asked by the researcher to sit two in a line (one on the (condition, recall, choice and gender) were identified left side and the other on the right side) to avoid influenc- using loglinear analysis (four way interaction: L. R. χ2 = ing each other. They were told that they would be asked to 0, p = 1 > 0.05; three way interaction: L. R. χ2 = 0.56, p = describe what they saw in a questionnaire after the view- 0.76 > 0.05). Therefore, the following analysis will just ing. The children in the product placement condition focus on the interactions between two variables. watched a movie clip from I Am Sam, a story about a father and his young daughter, for 3.5 minutes in which Nike Brand Recall was used as a prop in one scene (displayed on the team t- shirt worn by the daughter playing in a school football As we expected from the covert nature of the placement, match). This film has not been distributed in China and there was no difference in recall between these two children were therefore unlikely to have seen it previ- conditions. To be exact, in both conditions, no child could ously. During the whole scene, the Nike logo appeared recall being exposed to Nike. seven times, each time for less than a second. Children in the control group watched an edited clip of a football Brand Choice match without any brands for a similar length. Our prediction that product placement will have a clear When children finished watching, a questionnaire was effect on children’s choice was correct. There was a distributed to each child. As instructed, children in the significant difference between the control group and product placement condition were asked to describe the product placement group in their Nike choice (n = 55, χ2 = story they saw in the movie while children in the control 4.583, df = 1, p = 0.032 < 0.05). To be exact, while in the group were asked to describe the match they just saw in the placement condition 50 percent chose Nike, only 22 clip. Then in a cover story about helping a new hypermarket percent did so in the control group (see Figure 1). manager to decide which brands to make available, children were asked to make a single choice for each of five When the issue of gender was checked, there was a product categories including sports footwear and four significant difference among boys’ choice when they filler product categories from a set of brands (e.g., choco- were exposed to the brand compared to not being ex- late bars) with the names (in Chinese characters) and posed: n = 33, χ2 = 4.332, df = 1, p = 0.037 < 0.05. In the logos as stimuli. control group only 3 (out of 15) boys chose Nike, whereas

24 Marketing and Public Policy Conference Proceedings / 2009 FIGURE 1 The Effect of Low-Connected Product Placement on Children’s Choice Figure 1: The Effect of Low-Connected Product Placement on Children’s Choice

100%

80%

60% Nike Choice 40% Others

20%

0% Control Group Placement Group

10 (out of 18) did so if they were exposed to the brand. But higher elaboration threshold than girls and therefore be there was no difference in girls’ choice (p = 0.66 > 0.05). more likely to make judgments heuristically because of Hence the effect seems largely confined to boys (see their schema-based processing (Meyers-Levy and Sternthal Figure 2). 1991). In other words, boys are more likely to rely on fluency to make preference judgments rather than using a Discussion detailed processing strategy unless their attention at the time of exposure has been engaged sufficiently to induce This experiment demonstrates that the effect of product elaboration. The Nike logo on a girl’s T-shirt is unlikely placement upon children even though they watched a to have been processed explicitly by boys, even though movie clip for only a few minutes and were exposed to the the logo, being meaningful to them, seems to have been brand for just a few seconds. This finding supports the implicitly noted. earlier work of Auty and Lewis (2004) and Mallinckrodt and Mizerski (2007). When visual product placement is EXPERIMENT 2 only incidental and fleeting, its effectiveness was evident in the children’s choice, but not their brand recall, even What happens when a product is so prominently displayed though the children knew that they would be asked about that children have a high chance of recalling it? Will their the clips afterwards. responses differ in any way from advertising in terms of recall and choice? The purpose of Experiment 2 was to These findings support the idea that low recall does not explore the effect of a visual product placement on chil- mean low persuasion, and they are in line with previous dren when it is in the foreground and connected to the research that has shown that consumer brand preference characters or storyline. may be influenced without the recollection of earlier exposure to the brand (e.g., Lee 2002; Shapiro 1999). When placements go beyond just being a prop, they may Thus when children were asked to make a decision of little become visible to the audience by virtue of size and personal relevance, like helping a new store manger to position in the scene (Gupta and Lord 1998). In particular, decide which brands to list, they appeared to rely on when visual placements are more integrated into the film, simple cues to make a judgment. This effect was particu- they may have the same effect as “transformational” ads larly evident among boys who are believed to have a where the symbolic nature of the consumption experience

Marketing and Public Policy Conference Proceedings / 2009 25 FIGURE 2 Boys’ and Girls’ Choice of Nine Figure 2: Boys’ and Girls’ Choice of Nike

60%

40%

Nike Choice (boy Nike Choice (girls

20%

0% Control Group Placement Group

is the only focus rather than any product or brand informa- clip in Experiment 1 and those exposed to the Pizza Hut tion (Russell 1998; Balasubramanian, Karrh, and clip. Patwardhan 2006). Thus, we expect there to be no difference in terms of brand recall between children who watch Although Experiment 1 suggested that boys may process transformational ads and children who watch a movie clip product placements differently from girls, Meyers-Levy with a very prominent product placement. and Sternthal (1991) argued that such a difference would only occur when the message or task factors do not encourage a detailed processing strategy. The stimulus in Although previous research suggests that both brand Experiment 1 was below the elaboration threshold of both recall/recognition and product judgment and brand choices genders. However, when a personally relevant placement are influenced by information retrieved from memory is prominent and offers insight into one of the main (e.g., Alba, Hutchinson, and Lynch 1992), Monroe and characters, both boys and girls should engage in process- Lee (1999) argued that consumers often make purchasing ing the brand above their elaboration threshold, especially decisions not on what they can explicitly remember about as they know they will have to describe what they have the advertisement but on what they “know” about the seen. product. In the case of this research, although prior exposure to the brand in the product placement or in an ad will Experimental Design and Sample increase the brand’s accessibility in the memory and have an advantage over those brands that are less accessible Fifty-five children (aged 9–11) from the same subject (Nedungadi 1990), there should be no misattribution of pool as Experiment 1 were recruited for this research. As the fluency as preference because children will be aware in Experiment 1, children were randomly allocated by of the source of the familiarity and should discount the enrollment numbers by their teachers either to watch TV exposure as not relevant to the preference task (Schwarz advertising or watch a short movie clip with product 2004). As such, when making a choice, children should placement. The dependent variable was children’s choice rely on what they know about the products (perhaps from and brand recall, with exposure to product placement and personal experience) rather than prior exposure to make a gender (boys = 33; girls =22) as independent variables. In judgment. Therefore, there should be no difference in addition, as a check there was an unexposed control group choice between the ad group and the product placement (n = 27), although the focus of this experiment was on the group. However, we do expect there to be a difference in differences between prominent placements and transfor- the rate of recall between the group exposed to the Nike mational ads.

26 Marketing and Public Policy Conference Proceedings / 2009 Materials and Procedure will focus on the interactions between two different variables. The whole procedure was similar to Experiment 1. The children (in groups of 10–12) were asked to watch the Brand Recall movie clip with product placement or TV advertising in their school computer clusters for about 2–3 minutes. As predicted, there was no difference in recall between They were asked to sit with some distance in between to these two conditions (n = 53, χ2 = 1.481, df = 1, p = 0.224 > avoid influencing each other. The same as in Experiment 0.05). However, due to the long exposure time in the 1, the children in the product placement condition watched advertising group, slightly more children recalled Pizza a movie clip from I Am Sam for about 3.5 minutes in which Hut brand (56%) than children in the placement condition Pizza Hut was the setting and was especially dominant in (39%) although the difference is not significant (see one scene (lasting approximately 20 seconds). Pizza Hut Figure 3). was where the father (who hopes to get custody of his young daughter) worked, and the story focuses on his Brand Choice competence in the job, showing him serving and cleaning. Children in the TV advertising condition watched four 30- The second prediction was that product placement will second TV ads, one of which was for Pizza Hut. In this ad, have the same effect as TV advertising on children’s there is no direct message about product information but choice. The result supports this expectation. There was no rather Santa Claus claims that Pizza Hut is the restaurant significant difference between the advertising group and that only the best children can go to during Christmas. the product placement group in fast food restaurant choice Although not widely celebrated in China, the images of (n = 53, χ2 = 0.853, df = 1, p = 0.356 > 0.05). Christmas and Santa Claus are widely recognized and understood by Chinese children. The order of different Gender Processing ads was randomized and counter-balanced between different children. Both groups were told they would be As expected, there was no difference between boys and asked to describe what they had seen after the viewing. girls in their recall of Pizza Hut, when both were exposed to the plot-connected placement: n = 28, χ2 = 0.003, When children finished viewing the stimulus material, a df = 1, p = 1. questionnaire was distributed to each child. Children in the TV advertising condition were first asked to comment DISCUSSION on which ad they thought was most humorous and why, while children in the product placement condition were Although Experiment 1 demonstrated that when visual asked to describe the story they had seen in the movie. placements are shown in the background children are Then they were asked to help a manager make a choice for mainly influenced unconsciously, this experiment shows brand outlets in different product categories including fast that when visual placements are prominent and related to food restaurants to appear in a new hypermarket and to a main character, children are more consciously influ- write down any brand recalled as in Experiment 1 using enced to the extent of showing high levels of brand recall the exclude-recollect procedure as in Shapiro (1999). (see Figure 4). RESULTS The high rate of recall made measurements of choice Manipulation Check using the exclude-recollect procedure difficult, because that technique is designed to measure unconscious choice As in Experiment 1, children’s understanding of the only. Yet even given the exclusion procedure, which exclude-recollect procedure was checked. In the ad con- meant that only children who did not recall seeing Pizza dition, two children who recalled Pizza Hut also chose Hut could choose it, there was no difference between the Pizza Hut, and were therefore excluded from the analysis, placement group and the ad group, suggesting that promi- reducing the sample to 53. nent placements work in the same way as ads that are not recalled. In explaining a similar result in an experiment No three or four way interactions among all the variables involving word completion tasks, Debner and Jacoby (condition, recall, choice and gender) were identified (1994) contend that unconscious influences of memory using loglinear analysis (four way interaction: L.R. χ2 = (as in forgetting that one has been exposed to a stimulus) 0.33, p = 0.565 > 0.05; three way interaction: L.R. χ2 = are closely related to unconscious perception “and may 7.939, p = 0.16 > 0.05). Therefore, the following analysis even share common mechanisms” (p. 310).

Marketing and Public Policy Conference Proceedings / 2009 27 FIGURE 3 Recall of Product in Advertising versus Highly Connected Product Placement Figure 3: Recall of Product in Advertising versus Highly-Connected Product Placement

90%

60%

Recall Not Recall

30%

0% Ads condition Placement Condition

FIGURE 4 Recall of Product in Low-Connected Placement versus High-Connected Placement Figure 4: Recall of Product in Low-Connected Placement versus High-Connected Placement

100%

90% 80%

70% 60% 50% Recall

40% Not Recall

30%

20%

10%

0% loosely connected highly connected

28 Marketing and Public Policy Conference Proceedings / 2009 Intriguingly, an unexpected finding in Experiment 2 was more evident on behavior rather than cognitive outcomes that among those children who did not recall seeing Pizza like brand recall. In other words, when product placement Hut in the film or in an ad (and thus were free to choose it is not prominent or integrated into the storyline, children within the exclusion procedure) there was nevertheless a are more likely to be influenced under their awareness lower choice of that brand than in a control group who threshold and perceptually rather than conceptually. This were not exposed to the brand: 18 percent who saw the is because children, especially boys, may process the product placement, 9 percent who saw the ad compared to stimulus implicitly and thus show limited brand recall 37 percent who were in the control group (n = 82, χ2 = (Auty and Lewis 2004). Then, when asked to make a 4.00, df = 2, p = .135). The comparison between the ad choice, without realizing the true source of their prefer- group and the control group reaches significance at the 90 ence, children may misattribute the processing fluency percent level of confidence (p = .085), but not 95 percent. derived from the brand logo as being a positive feeling There was no significant difference between the two toward the brand and therefore choose the “target” brand treatment groups. (Winkielman et al. 2003).

These findings suggest that some form of implicit pro- Yet this research found that when product placement is cessing may be occurring in both groups in addition to the more prominent, children are more likely to process the explicit processing recorded in the levels of recall, be- placement consciously, in the same way that they process cause although these children are not conscious of having transformational ads which attempt to influence their seen the brand, they still appear to choose the brand less affect toward the brand. Such processing may lead to often than would be expected (although not at the 95 higher cognitive results (brand recall), but not necessarily percent confidence level). Further research would be to higher levels of choice or more positive attitudes required to ascertain if this lower rate of choice is the result (Cowley and Barron 2008). For policy makers, an impor- of some form of unconscious counterargument owing to tant point is that our research indicates that the level of implicit conceptual processing. That is, once a placement choice for those exposed to prominent product placement becomes obvious in the way ads are obvious – even if it is is no different from those exposed to transformational ads. not recalled – it is no longer as influential as when it is Thus, banning ads alone, as some countries have done, is covert because the influence is not merely perceptual. unlikely to have the desired effect of protecting children. Far from it: a ban will most likely shift ever more of the Are boys more susceptible to placements than they are to communications budget over to product placement, where advertising because of their tendency shown in Experi- its influence is much less easy to detect because of implicit ment 1 to process brands implicitly? In this experiment the processing and the effect of background placements on very prominent placement stimulus and the recall task brand choice may actually exceed that of ads. Boys may encouraged both boys and girls to apply a detailed pro- be more susceptible to this unconscious influence than cessing strategy, and thus, as suggested by Meyers-Levy girls, because of their tendency to eschew a detailed and Maherswaran (1991) no difference was found be- processing strategy unless encouraged otherwise. tween boys and girls in their recall of the placement or their choices. LIMITATIONS AND FUTURE RESEARCH

GENERAL DISCUSSION This research has some limitations arising from the decision to focus only on unconscious influences using the Although children’s vulnerability to product placement exclusion-recollection procedure. Some comparison of has raised great concerns among policy makers around the choice not requiring exclusion of recalled brands might world, the academic evidence to help them set regulations allow greater insight into the difference between plot- to protect children from the implicit influence of product connected and incidental placements, and between plot- placement is still lacking. Furthermore, even those coun- connected placements and ads. Furthermore, this research tries that do ban product placement in TV are limited in compared placements and transformational ads that em- their ability to ban it from other media such as film and phasized the symbolic meaning of the brand but did not video games. This research has explored the influence of provide factual information about the products subtle product placements on children even after brief and (Balasubramanian, Karrh, and Patwardhan 2006). How- understated exposure. Different types of product place- ever, as Sheehan and Guo (2005) pointed out, the brand ments appear to influence children through different may become the “star” of a program (product assimila- mechanisms. For example, when product placement is in tion) where detailed information about the brand/product the background, appearing only incidentally, its effect is is provided and demonstrated. In this situation, place-

Marketing and Public Policy Conference Proceedings / 2009 29 ments may function like informational ads and may be same limitations as advertisements, as prominent place- processed differently, with less implicit influence on ments appear to affect children in the same way as adver- children but perhaps having more explicit influence. tising. However, it would not be practical or even desirable to have a pop up warning (itself a form of priming) The role of the product category in the type of processing, every time a brand appears in the background (imagine the especially by boys, is not clear from our research. Did screen clutter with a scene set in a supermarket!), and yet boys process the sports logo, albeit with very limited this form of placement seems to be more effective in attention, because of intrinsic interest in sports and there- affecting preference than more obvious ones. fore act on their processing fluency? Would their choice have been affected by a category of less personal rel- Our somewhat unexpected finding that boys are more evance in that the exposure might not have been processed likely to be influenced by subtle placements than girls even minimally? In any case, one can assume that brand suggest that policy makers might need to develop specific owners would seek placements of relevance to the target intervention programs targeted especially at these more audience, so this result merely confirms that implicit vulnerable children to cope with this hidden form of processing can be effective below the threshold of aware- advertising. Further research would be required to sup- ness. port this contention.

GUIDANCE FOR POLICY MAKERS We acknowledge that other factors may have influenced the effect on choice, for example, existing strong prefer- This research demonstrates that different types of product ences (Lee 2002) and program liking (Cowley and Barron placements influence children through different mecha- 2008). However, what we are particularly interested in are nisms. While children are mainly able to identify promi- the unconscious effects of exposure to brands in entertain- nent placements, they are more likely to be influenced ment and the possibility that repeated placements become under their awareness threshold when the placements are associated with positive affect implicitly through that just used as props in the background. Therefore, when exposure (Braun-LaTour and LaTour 2004). While more setting the regulations, policy makers may need to use research is needed to understand fully how produce place- different guidance toward different placements with re- ments influence children, our research suggests it is time gard for size, centrality and plot connection (Russell to err on the side of caution by putting in place guidelines 2002). There is a case for subjecting placements to the for these placements in children’s entertainment.

“Viewers’ Recognition of Brand Placed Within a REFERENCES Film,” International Journal of Advertising, 15 (2), 140–51. Achenreiner, Gwen B. and Deborah Roedder John (2003), Balasubramanian, Siva K, James Karrh, and Hemant “The Meaning of Brand Names to Children: A Devel- Patwardhan (2006), “Audience Response to Product opmental Investigation,” Journal of Consumer Psy- Placements,” Journal of Advertising, 35 (Fall), 1151– chology, 13 (3), 205–19. 141. Akhter, S., C. Andrew, and S. Durvasula (1990), “A Bettman, James R., Mary Luce, and John W. Payne Framework for Conceptualizing and Measuring the (1998), “Constructive Consumer Choice Process,” Involvement Construct in Advertising Research,” Journal of Consumer Research, 25 (3), 187–217. Journal of Advertising, 19 (4), 27–41. Braun-LaTour, Katherine and Michael LaTour (2004), Alba, Joseph W., J. Wesley Hutchinson, and G. Lynch, Jr. “How and When Advertising Can Influence Memory (1992), “Memory and Decision Making,” Handbook for Consumer Experience,” Journal of Advertising, of Consumer Behavior, Thomas S. Robertson and 33 (4), 7–25. Harold H. Kassarjian, eds. Englewood Cliffs, NJ: Brennan, Ian, Khalid M. Dubas, and Laurie A. Babin Prentice Hall, 1–49. (1999), “The Influence of Product Placement Type Auty, Susan and Charlie Lewis (2004), “Exploring and Exposure Time on Product Placement Recogni- Children’s Choice: The Reminder Effect of Product tion,” International Journal of Advertising, 18 (3), Placement,” Psychology & Marketing, 21 (9), 697– 323–37. 713. Brennan Stacey, Philip J. Rosenberger III, and Veronica Babin, Laurie A. and Sheri Thompson Carder (1996a), Hementera (2004), “Product Placement in Movies:

30 Marketing and Public Policy Conference Proceedings / 2009 An Australian Consumer Perspective on Their Ethi- Mahwah, NJ: Lawrence Erlbaum Associates, 63–78. cality and Acceptability,” Marketing Bulletin, 15, 1– Lee, Angela Y. (2002), “Effects of Implicit Memory on 16. Memory-Based Versus Stimulus-Based Brand CCFC (2008), “Comments” by the Center for a Commer- Choice,” Journal of Marketing Research, 39 (4), cial-Free Childhood to FCC, submitted September 440–54. 28, 2008 (accessed on October 30, 2008) [available at ______(2004), “The Prevalence to Metacognitive http://www.commercialexploitation.org/pdf/ Routes to Judgment,” Journal of Consumer Psychol- fcccomments.pdf]. ogy, 14 (4), 349–55. Cowley, Elizabeth and Chris Barron (2008), “When Prod- Lehu, Jean-Marc (2007), Branded Entertainment: Prod- uct Placement Goes Wrong: The Effect of Program uct Placement and Brand Strategy in the Entertain- Liking and Placement Prominence,” Journal of Ad- ment Business. London: Kogan Page. vertising, 37 (1), 89–98. Mallinckrodt, Victoria and Dick Mizerski (2007), “The Curran, Tim and Douglas L. Hintzman (1995), “Viola- Effects of Playing an Advergame on Young Children’s tions of the Independence Assumption in Process Perceptions, Preferences, and Requests,” Journal of Dissociation,” Journal of Experimental Psychology: Advertising, 36 (2), 87–100. Learning, Memory, and Cognition, 21, 531–47. McKechnie, Sally A. and Jia Zhou (2003), “Product Debner, James A. and Larry L. Jacoby (1994), “Uncon- Placement in Movies: A Comparison of Chinese and scious Perception: Attention, Awareness, and Con- American Consumers’ Attitudes,” International Jour- trol,” Journal of Experimental Psychology: Learn- nal of Advertising, 22 (3), 349–74. ing, Memory, and Cognition, 20 (2), 304–17. Meyers-Levy, Joan and Brian Sternthal (1991) “Gender Gould, Stephen J., Pola B. Gupta, and Sonia Grabner- Differences in the Use of Message Cues and Judg- Krauter (2000), “Product Placement in Movies: A ments,” Journal of Marketing Research, 28, 84–96. Cross-Culture Analysis of Austrian, French, and ______and Durairaj Maheswaran (1991), “Ex- American Consumers’ Attitudes Towards This ploring Differences in Males’ and Females’ Process- Emerging International Promotion Medium,” Jour- ing Strategies,” Journal of Consumer Research, 18 nal of Advertising, 29 (4), 41–58. (1), 63–71. Gupta, Pola B. and Kenneth R. Lord (1998), “Product Monroe, Kent B. and Angela Y. Lee (1999), “Remember- Placement in Movies: The Effect of Prominence and ing vs. Knowing: Issues in Buyers’ Processing of Mode on Audience Recall,” Journal of Current Is- Price Information,” Journal of the Academy of Mar- sues & Research in Advertising, 20 (1), 47–60. keting Science, 27 (Spring), 207–25. Hood, John (2005), “Warning: Coke, Though Refreshing, Moore, Elizabeth S. and Victoria J. Rideout (2007), “The Is a Sponsor of this Program,” Reason Magazine, Online Marketing of Food to Children: Is it Just Fun November 22 (accessed October 30, 2008), [avail- and Games?” Journal of Public Policy and Market- able at http://www.reason.com/news/show/ ing, 26 (2), 202–20. 33001.html]. Morton, Cynthia R. and Meredith Friedman (2002), “‘I Janiszewski, Chris (1993), “Preattentive Mere Exposure Saw It in the Movies’: Exploring the Link Between Effects,” Journal of Consumer Research, 20 (3), Product Placement Beliefs and Reported Usage Be- 376–93. havior,” Journal of Current Issues and Research in Karrh, James A. (1994), “Effects of Brand Placement in Advertising, 24 (2), 33–40. Feature Films,” in Proceedings of the 1994 Confer- Moses, L.A. and D.A. Baldwin (2005), “What Can the ence of the American Academy of Advertising, Karen Study of Cognitive Development Reveal about Whitehill King, ed. Athens, GA: American Academy Children’s Ability to Appreciate and Cope with Ad- of Advertising, 90–96. vertising?” Journal of Public Policy and Marketing, ______, Kathy Brittain McKee, and Carol J. Pardun 24 (Fall), 186–201. (2003), “Practitioners’ Evolving Views on Product Nedungadi, Prakash (1990), “Recall and Consumer Con- Placement Effectiveness,” Journal of Advertising sideration Sets: Influencing Choice Without Altering Research, 43 (2), 138–49. Brand Evaluations,” Journal of Consumer Research, Law, Sharmistha and Kathryn A. Braun (2000), “‘I’ll 17, 263–76. Have What She’s Having’: Gauging the Impact of Nelson, Michelle R. (2002), “Recall of Brand Placements Product Placements on Viewers,” Psychology & in Computer/Video Games,” Journal of Advertising Marketing, 17 (12), 1059–75. Research, 42 (2), 80–92. ______and ______(2004), “Product Owen, Laura, Susan Auty, Charlie Lewis, and Damon Placements: How to Measure Their Impact,” in The Berridge (2007), “Children’s Understanding of Ad- Psychology of Entertainment Media, L.J. Shrum, ed. vertising: An Investigation Using Verbal and Picto-

Marketing and Public Policy Conference Proceedings / 2009 31 rially Cued Methods,” Infant and Child Develop- ______and H. Shanker Krishnan (2001), ment, 16 (6), 617–28. “Memory-Based Measures for Accessing Advertis- Roedder John, Deborah (1999), “Consumer Socialization ing Effects: A Comparison of Explicit and Implicit of Children: A Retrospective Look at Twenty-Five Memory Effects,” Journal of Advertising, 30 (3), 1– Years of Research,” Journal of Consumer Research, 13. 26 (3), 183–214. Sheehan, Kim Bartel and Aibing Guo (2005), “Leaving Russell, Cristel A. (1998), “Towards a Framework of on a (Branded) Jet Plane: An Exploration of Audi- Product Placement: Theoretical Proposition,” Ad- ence Attitudes Towards Product Assimilation in Tele- vances in Consumer Research, 25, 357–62. vision Content,” Journal of Current Issues and Re- ______(2002), “Investigating the Effectiveness search in Advertising, 27 (1), 79–91. of Product Placements in Television Shows: The Tiwsakul, Rungpaka, Chris Hackley, and Isabelle Szmigin Role of Modality and Plot Connection Congruence (2005), “Explicit, Non-Integrated Product Placement on Brand Memory and Attitude,” Journal of Con- in British Television Programmes,” International sumer Research, 29 (3), 306–18. Journal of Advertising, 24 (1), 95–111. Schneider, Lars-Peter and T. Bettina Cornwell (2005), Winkielman, Piotr, Norbert Schwarz, Tedra A. Fazendeiro, “Cashing in on Crashes via Brand Placement in and Rolf Reber (2003), “The Hedonic Marking of Computer Games,” International Journal of Adver- Processing Fluency: Implications for Evaluative Judg- tising, 24 (3), 321–43. ment,” in The Psychology of Evaluation: Affective Schwarz, Norbert (2004), “Metacognitive Experiences in Processes in Cognition and Emotion, J. Musch and Consumer Judgment and Decision Making,” Journal K.C. Klauer, ed. Mahwah, NJ: Lawrence Erlbaum, of Consumer Psychology, 14 (4), 332–48. 189–216. Shapiro, Stewart (1999), “When an Ad’s Influence is Wright Peter, Marian Friestad, and David M. Boush Beyond Our Conscious Control: Perceptual and Con- (2005), “The Development of Marketplace Persua- ceptual Fluency Effects,” Journal of Consumer Re- sion Knowledge in Children, Adolescents, and Young search, 26 (1), 16–37. Adults,” Journal of Public Policy and Marketing, 24 ______, Deborah J. MacInnis, and Susan E. Heck- (2), 222–33. ler (1997), “The Effects of Incidental Ad Exposure Zaichkowsky, Judith Lynne (1985), “Measuring the In- on the Formation of Consideration Sets,” Journal of volvement Construct,” Journal of Consumer Re- Consumer Research, 24 (1), 94–105. search, 12 (December), 341–52.

For further information contact: Susan Auty Management School Lancaster University Lancaster LA1 4YX United Kingdom Phone: +44.1524.593914 Fax: +44.1524.593928 E-Mail: [email protected]

32 Marketing and Public Policy Conference Proceedings / 2009 IMMIGRATION POLICY AND VULNERABILITY

Garrett Coble, Oklahoma State University, Stillwater Fernando Jiménez, Oklahoma State University, Stillwater Marlys Mason, Oklahoma State University, Stillwater

SUMMARY there. To better understand the impact of this state’s policy, we conduct in-depth interviews with key infor- Immigration has long been a focal, and at times conten- mants working in the Hispanic community (e.g., business tious, topic in public policy in the United States. Today, owner, educator, healthcare worker, minister) and with nearly 30 million immigrants live in the U.S., of which an Hispanic immigrants. estimated 11.8 million are unauthorized immigrants. Of this segment, 35 percent are estimated to have entered Our secondary analysis of immigration policy reveals after 2000, resulting in an increase of 3.3 million unautho- themes related to constraints being targeted to specific rized persons (Hoefer, Rytina, and Baker 2008). Given areas including employment, education, healthcare, and these numbers, it is no surprise that legislators are working other basic consumer services (transportation, housing). on policies to address the growing trend of unauthorized Our analysis of the interviews reveals the emergence of immigration. three focal themes related to the impact of the restrictive policy including: (1) an initial fear of the unknown conse- In recent years, immigration policy has received much quences of the policy, (2) withdrawal and restrictions in attention at the state level because the federal government consumption, and (3) a renewed resistance and resolve has not taken an active role in amending national policy within in the community. (NCSL 2007). As a consequence, immigration laws issued by states have rapidly expanded (NCSL 2008). Most The initial theme that emerged from the interviews is that bills and resolutions attempt to discourage unauthorized of the unknown and fear. For our informants, living as an immigration in their respective territories by making it unauthorized immigrant (or family member of the immi- harder to access employment and services. Although grant) means living with the ongoing fear of deportation these restrictions explicitly imply consumption alterations, of themselves or family members. The unknown effects no marketing studies have examined the impact of such brought about rumors, stories, and confusion, all which policy. Does immigration policy contribute to consumer resulted in an elevated level of fear. The fear was so vulnerability among immigrants? If so, what adjustments prominent that they began adjusting their consumption, do immigrants make in their consumption patterns, in how their interactions in the marketplace, their driving pat- they attain services important for quality of life, and in terns, their willingness to attend school, go to church, seek how they interact in the marketplace and community? health services, and so on. The broader community also What concerns underlie their shifts in consumer behav- reported feeling this withdrawal (and migration out of the ior? To date, marketing studies that examine immigration state) and witnessing the impact through businesses clos- tend to focus on acculturation and identity issues (Penaloza ing, community events not being attended, Spanish lan- 1994, 1995) with little consideration to policy’s role in guage media withdrawing from the market, etc. marketplace participation. Our research addresses this gap. A second theme emerging in our analysis is an increased restriction in consumption. For most informants, lan- This paper investigates the impact of immigration policy guage barriers and financial resources already created following the passage of bills which restrict ability to limits on consumption. Now they also felt the need to interact in the marketplace and access important services. change how they accessed and consumed services in order First, drawing on secondary sources (legal documents, to not violate the new law. Informants reported five main newspapers) we review the current status of immigration areas (transportation, spending, healthcare, education, policy at the state level, identify the focus of these re- and housing) where restriction was felt strongest follow- forms, and highlight particular aspects using selected ing the new policy and adjustments needed to be made. states. Then, we describe the case of a specific southern state that has by some been noted to have the “toughest” The third emergent theme is resistance and resolve within immigration reform in the U.S. and conduct the case study the community. This topic emphasizes the collective

Marketing and Public Policy Conference Proceedings / 2009 33 response to the immigration law. While individuals and but not all left. The ones who stayed in Oklahoma wit- families felt fear and vulnerability, it was through their nessed a change in their external community and in the community they found renewed strength. The Hispanic way they would have to interact in that community. Our community united and is attempting to resist what they findings contribute to the conceptualization of consumer feel as an aggression. Leaders have emerged within the vulnerability a greater understanding of the effect policy community. They have organized, developed a public can play on the external conditions that contribute to relations campaign (newspaper ads, editorials, billboards), vulnerability. Our analysis also provides insights into marched on the state capitol, worked with non-Hispanic resiliency and that empowerment can occur when com- business owners for reform, and reached out to vulnerable munities respond collectively to policies that impede members in the community. As a result, our informants consumer participation in the marketplace. In the future, report a renewed sense of connection within the commu- greater caution should be taken in immigration policy nity. development and enforcement, and a deeper understanding of the ramifications of marketplace constraints is House Bill 1804 accomplished part of its goal when large needed. References are available upon request. numbers of Hispanics left the state. Large numbers left,

For further information contact: Garrett Coble Oklahoma State University 405 C Spears School of Business Stillwater, OK 74078–4011 Phone: 405.744.8624 Fax: 405.744.5180 E-Mail: [email protected]

34 Marketing and Public Policy Conference Proceedings / 2009 20th ANNIVERSARY SESSION

EXPLORING THE MORTGAGE CRISIS: CAUSES AND REMEDIES

Session Chair: William L. Wilkie, University of Notre Dame

Panelists: Karen M. Pence, Federal Reserve Board Barry Zigas, Consumer Federation of America Todd J. Zwicki, George Mason University School of Law John G. Lynch, Jr., Duke University

SUMMARY Barry Zigas is the Director of Housing Policy for the Consumer Federation of America. He also works on The current financial crisis in the United States was policy development and consumer advocacy through triggered in part by a dramatic rise in delinquent mort- his firm Zigas and Associates LLC. He is a former gages and subsequent foreclosures, with major adverse senior vice president for Fannie Mae and prior to that consequences for consumers, banks and financial markets was president of the National Low Income Housing around the world. A large decline in the value of homes Coalition. and in the capital of many banks, and tightening credit has been the result. The crisis has exposed pervasive weak- Todd J. Zywicki is Professor of Law at George nesses in financial industry regulation and the global Mason University School of Law and Senior Scholar financial system. Questions about potentially unfair and of the Mercatus Center at George Mason University. deceptive practices have led to calls by some for stronger He has published legal and economic analyses in a consumer protections. Recent events also highlight the number of areas such as consumer bankruptcy law, importance of informed consumer choice at each stage of consumer credit and subprime lending. the mortgage lending process. Our final speaker will address the role of consumer This session features experts from varying academic and decision-making in the crisis, and how this might be professional backgrounds, who examine how the mort- improved in the future: gage crisis developed, and debate preferred approaches for dealing with it. John G. Lynch Jr. is the Roy J. Bostock Professor of Marketing at the Fuqua School of Business at Duke Karen M. Pence is Senior Economist in the House- University. An expert in the study of consumer infor- hold and Real Estate Finance Section of the Federal mation processing, his recent research focuses on Reserve Board: she will provide a detailed look at the consumer decision making in markets for financial causal factors and timeline of the unfolding crisis. services such as retirement savings plans and mortgages. Our next two speakers will provide very different analyses of the problem, as well as different approaches to resolving it:

Marketing and Public Policy Conference Proceedings / 2009 35 WARNING CONSUMERS ABOUT ALCOHOL MESSAGES IN TELEVISION SERIES: AUDIENCE CONNECTEDNESS, WARNING TIMING, AND EMPHASIS

Cristel Antonia Russell, University of Auckland, New Zealand Dale W. Russell, Prevention Research Center, Berkeley Joel W. Grube, Prevention Research Center, Berkeley

SUMMARY about advertising messages in the story should raise viewers’ attention to them and in turn improve their ability This research investigates whether warning viewers about to discount them. Similarly, a warning about health mes- alcohol messages embedded in the content of a TV pro- sages should raise viewers’ attention to the negative gram affects viewers’ drinking beliefs and if audience consequences of drinking. Messages that emphasize the connectedness moderates the warning’s impact (Russell negative consequences of drinking should resonate with and Russell 2008). Despite the alcohol industry’s self- highly connected viewers. Conversely, highly connected regulation, positive alcohol messages are increasingly viewers may discount warning messages about advertis- present, often in the form of product placements ing content because they view the program content as part (Christensen et al. 2000). Negative consequences of drink- of real life. Post-warnings have not been tested in previous ing are also sometimes depicted, as policymakers advo- research but they should allow viewers to reflect on the cate the use of entertainment-education to communicate overall message of the episode and to modify their beliefs the risks associated with unhealthy behaviors (Singhal accordingly. et al. 2004). Often, pro- and anti-alcohol messages occur within the same series, and even the same episode (Blair Method et al. 2005). Warning messages have been advocated for to disclose the presence of embedded messages (CA An experiment was conducted based on a 2 (connected- 2007). This research tests the impact of different warning ness level) X 2 (timing of the warning: before or after conditions on viewers’ alcohol beliefs following expo- episode) X 2 (emphasis of warning: advertising vs. health sure to alcohol messages embedded in program content. message) between-subjects design with a control condition (no warning). Connectedness was measured and the Conceptual Model timing and emphasis of the warnings were manipulated. The stimulus was a real episode of Two and a Half Men, Warning messages placed before or after a persuasive the primetime series with the most prominent alcohol communication may allow individuals to access persua- messages (Russell and Russell 2009). The storyline in- sion knowledge and counter argue the message (Friestad cluded both positive and negative alcohol messages. Two and Wright 1994). However, pre-warnings may increase hundred fifty university students participated in the ex- interest in and reactance to the subsequent message (Bush- periment which consisted of viewing an episode and man and Stack 1996). Therefore, warning individuals completing a questionnaire. A first series of questions after exposure to the program message may be more assessed attitude toward the episode, involvement with, effective. Audience connectedness is likely to moderate and connectedness with the program. Then, general alco- warnings’ impact because it reflects engagement with the hol beliefs were assessed: respondents indicated how TV program (Russell et al. 2004) and may preempt likely eight positive and eight negative consequences viewers from treating embedded messages as persuasive would happen to them if they were to drink three or four attempts (Chen et al. 1992). whole drinks of alcohol (Grube et al. 1995).

This research proposes that the effectiveness of a warning Results depends on viewers’ level of connectedness. Amongst low connected viewers, the timing and emphasis of the The two alcohol belief measures were analyzed in a 2 warning should have different impacts on viewers’ alco- (connectedness level, median split) X 2 (timing) X 2 hol beliefs. Pre-warnings should highlight the nature of (emphasis) MANCOVA with alcohol consumption, age, the alcohol messages in the episode. Hence, a forewarning and gender as covariates. There was a significant main

36 Marketing and Public Policy Conference Proceedings / 2009 effect of connectedness on positive beliefs qualified by However, several important conditions must be met for the three-way interaction. The advertising pre-warning the warning to effectively lower viewers’ positive beliefs reduced positive beliefs but only for the low connected about alcohol and its consequences. First, warnings about viewers, as expected. For the high connected viewers, the advertising messages in the episode only affected low opposite effect was observed, with even greater positive connected viewers, lowering their positive beliefs and beliefs than in the control condition. The post-warning increasing their negative beliefs about drinking. However, with a health emphasis produced lower positive beliefs highly connected viewers, who have developed parasocial amongst low connected viewers but not amongst highly relationships with the TV characters, were not affected by connected viewers. Neither the health pre-warning nor the the advertising warnings and even demonstrated a advertising post-warning affected positive beliefs. For reactance effect in the pre-warning condition. Second, negative beliefs, there was an interaction between con- even amongst low connected viewers, advertising warnings nectedness and warning emphasis. Health emphasis warn- can produce greater negative beliefs about alcohol but ings produced more negative beliefs about drinking reducing positive beliefs about alcohol requires the right amongst highly connected viewers, regardless of timing. combination of timing and emphasis of warning. Only the In contrast, warnings that emphasized the pro-alcohol pre-advertising and the post-health message warnings message did not affect highly connected viewers’ nega- generated significantly lower positive beliefs about alcohol tive beliefs but produced more negative beliefs amongst compared to the no warning condition. Advocates of low connected viewers regardless of timing. message warnings should be conscious of the differential impact of forewarning on high and low connected viewers. Discussion A forewarning about alcohol placements in a TV program may reduce the effect of those placements amongst low Warning viewers about embedded messages in the content connected viewers but it will instead increase their impact of a TV series can impact their beliefs about alcohol. on highly connected viewers.

For further information contact: Dale W. Russell Prevention Research Center 1995 University Avenue, Suite 450 Berkeley, CA 94704 Phone: 510.833.5722 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 37 TOBACCO BRANDING AND PLAIN PACKAGING: THE NEW FRONTIER IN TOBACCO CONTROL?

Janet Hoek, University of Otago, New Zealand Philip Gendall, Massey University, New Zealand Jordan Louviere, University of Technology Sydney, Australia

SUMMARY emotional attributes. These attributes then act as heuristics that do not require systematic or detailed processing, Although several countries have restricted tobacco mar- but elicit learned responses that may reduce the salience keting, smoking prevalence remains higher than govern- and effectiveness of competing stimuli. ment targets. Tobacco control researchers have argued that, while tobacco advertising and sponsorship may no In summary, theories of warning salience and behavior longer be possible, extensive covert marketing supports modification suggest plain packaging would enhance the smoking initiation. Undertaken primarily via packaging effectiveness of PWLs by making these more visually and branding, covert marketing undermines the effective- impactful and salient. By contrast, branding would ame- ness of advertising bans, enables tobacco companies to liorate the impact of PWLs as its reassuring imagery maintain the salience of their brands, and allows them to would compete with the PWL. We therefore hypothesized continue attracting new smokers. that connotations elicited by familiar brands would offset negative stimuli, such as PWLs. However, we posited that The importance of maintaining a strong brand presence is unfamiliar brand imagery would not have developed the evident from marketing literature; tobacco industry docu- emotional pairings necessary to offset PWLs, although ments also detail strategies that could be used to counter even unfamiliar branding would be more attractive than any marketing restrictions that might be introduced. Pack- generic packs, which earlier research has suggested are aging rates highly among these options as it has high regarded as unattractive. We further hypothesized that a utility and thus considerable potential for exposing brand familiar brand with a PWL would be more attractive than imagery, including colors, shapes, words, and logos. a generic pack featuring a text warning, while an unfamil- Tobacco researchers concluded this imagery reassures iar brand with a PWL would be less attractive than a smokers (Pollay 1995) and reduces the visibility and generic pack with a text warning. impact of health warnings (Mahood 1999; Carter 2003b). Their findings informed the Framework Convention on We tested these hypotheses using a 3*2 experimental Tobacco Control, which called for two regulatory mea- design that featured three brand options (familiar (Holi- sures to reduce packaging’s potential as a marketing day), unfamiliar (Kool) and plain pack) and two warning medium: stronger and more salient on-pack warnings, and formats (text and PWL). These were tested in a Best- plain (or generic) packaging. Worst (BWS) experiment where respondents evaluated the utilities of different attribute combinations. Face-to- Pictorial warning labels (PWLs) are more specific, more face interviews were conducted with a systematically personal, and less easily avoided than text-only warnings; selected sample of 245 young adults who were asked to they reduce the attractiveness of tobacco packages by imagine a scenario featuring an individual who had just disrupting the connotations evoked by brand imagery. moved to university and was living in a hostel (college). Argo and Main (2004) suggested warning labels attracted This individual had developed a new group of friends and consumers’ attention more effectively when they had had started smoking socially with the group. Respondents “vividness-enhancing characteristics,” such as large font, were asked to select the pack they thought this individual color, symbols, and specific information. Their findings would be most likely (best) and least likely (worst) to buy are consistent with behavior modification theory (Nord to share with the new social group. and Peter 1980), which implies that large, full-color, pictorial warnings will be more effective than text-only The familiar brand, Holiday, was the most preferred warnings, since the former are more likely to gain atten- option when paired with a text warning, followed by the tion. Respondent conditioning, a component of behavior unfamiliar brand (Kool), also when combined with a text modification, explains how tobacco branding pairs brand warning. The PWL depressed pack attractiveness, al- names, colors and other imagery with psychological and though brand familiarity offset this negative stimulus and

38 Marketing and Public Policy Conference Proceedings / 2009 even when the Holiday brand was paired with a PWL, featuring a text warning. PWLs and plain packaging had respondents were still more likely to prefer it to a generic similar effect sizes with respect to smokers; however, pack with a text only warning. However, when the unfa- plain packs had a stronger effect on non-smokers than did miliar Kool brand featured a PWL, it was slightly less the PWLs. This suggests branding attracts non-smokers’ attractive than the generic pack with a text warning. These attention, and implies that plain packs could reduce the findings suggest familiar brand imagery overrides nega- attractiveness of tobacco products to non-smokers, thus tive stimuli, even when these are intrusive and disconcert- potentially deterring initiation. ing. More generally, the findings support concerns that branding reduces the impact of health warning informa- Removing brand imagery decreases the attractiveness of tion. Furthermore, the difference between the familiar tobacco products and eliminating this marketing medium brand and plain pack was greater than the difference could deter smoking initiation. A PWL decreased the between the text and PWL warning; this suggests that attractiveness of tobacco packages, particularly among plain packaging could decrease the attractiveness of smok- smokers, and generic packages were less attractive than ing beyond the effects of PWLs. branded packages, especially when they featured a PWL. However, the study estimated package attractiveness, not Smokers had a stronger preference for branded packs with actual behavior and further work is required to test behav- text warnings than did non-smokers and were more likely ioral outcomes. Nevertheless, the findings illustrate the than non-smokers to identify packs with PWLs as their importance of tobacco branding demonstrate support ex- “worst” choice. While familiar branding offset the PWL tending the frontier of tobacco control to include plain among non-smokers, smokers had similar preferences for packaging. References are available upon request. the familiar brand paired with a PWL and the plain pack

For further information contact: Janet Hoek Department of Marketing University of Otago P.O. Box 56 Dunedin New Zealand Phone: +643.479.7692 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 39 IMPACT OF MARKETING ON INFANT FORMULA CHOICE AND SWITCHING

Sara B. Fein, Food and Drug Administration, College Park Yi Huang, University of Maryland, Baltimore Conrad Choiniere, Food and Drug Administration, College Park Judith Labiner-Wolfe, Food and Drug Administration, College Park

SUMMARY mothers, as well as the impact of marketing efforts to induce switching. Until 1990, infant formula in the U.S. was marketed to physicians, who were expected to recommend a formula Methods to parents. Since 1990, direct to consumer (DTC) marketing of infant formula has become widespread in the U.S. Data are from the Infant Feeding Practices Study II, a The World Health Organization views the marketing of longitudinal mail study of 2000 pregnant women from infant formula as a threat to breastfeeding and, since 1981, 2005–2007 (Fein et al. 2008). Mothers were sent ques- has promoted an International Code of Marketing Breast tionnaires in late pregnancy and 10 times during their Milk Substitutes that restricts all advertising of formula. infants’ first year. The sample is limited to mothers of The Code has not been adopted by the United States. healthy singletons born at 37 weeks gestation or greater. Breastfeeding advocacy and professional organizations, The survey collected data on all types of formula promo- including the American Academy of Pediatrics, have tion in the decision model of Cutler and Wright (2002), called for elimination of some DTC marketing activities including mothers’ exposure to prenatal formula adver- because of their adverse effects on breastfeeding (AAP tisements, direct sample mailings, and reasons for choos- 2005; Howard et al. 2000; Rosenberg et al. 2008). Efforts ing a particular formula (physician recommendation, brand to limit infant formula marketing will benefit from a fed to the baby at the hospital, brand given by the Supple- greater understanding of the effectiveness of various mental Feeding Program for Women, Children, and In- types of DTC marketing on mothers. fants (WIC), having coupons, information from family and friends, and previous experience with formula). In Infant formula has characteristics that make it both an addition, the survey collected data on exposure to market- experience and credence product. Consumers learn the ing at discharge via hospital gift packs that included quality of experience products (e.g., whether the infant formula samples and/or coupons. The questionnaire pro- tolerates the formula) after consuming it. However, con- vided six groups of formula from which the mother could sumers cannot detect the quality of credence products indicate the one used. Mothers were asked how many (e.g., whether the formula has an impact on the infant’s times they switched the formula fed to their infants and long term neural development), so the utility of consum- whether the formula was switched for health reasons. In ing the product may never be determined. These attributes this analysis, switching was measured in three ways: are characterized by high search and switch costs for the change in formula group, switches for non-health reasons, consumer (Nelson 1970; Darby and Karni 1973). and number of switches in a month.

Economic theory posits that search and switch costs make We hypothesize that implicit and explicit expert opinions markets less competitive. Expert opinions, such as doctor increase brand loyalty and thereby reduce search costs recommendations and hospital endorsements, reduce and increase the likelihood that mothers use the same search costs for the mother. Offering free samples of a formula used in the hospital (hospital brand) when their product at a time of particular need promotes brand infant is a month old. We also hypothesize that exposure loyalty, which in turn increases switching costs. Firms to infant formula advertising and marketing prenatally may use marketing and promotion to reduce those costs and at hospital discharge reduces switch costs, and thereby and encourage mothers to switch brands of infant formula counteracts brand loyalty and increases the likelihood that (Shum 2004). This analysis tests the impact of implicit a mother switches formula. To test the hypotheses we use hospital endorsements and doctor recommendations on logistic regression and longitudinal data analysis methods the development of brand loyalty to infant formula among adjusting for covariates such as demographics of the

40 Marketing and Public Policy Conference Proceedings / 2009 mother, infant sex and birth weight, WIC participation, web were more likely to switch according to all three breastfeeding initiation, percent of milk feeds that were switching measures. In addition, mothers who did not breast milk, breastfeeding duration, and measures of choose the hospital brand at month one and who received infant health and tolerance of formula. samples of formula in the mail were more likely to switch formula for reasons other than infant health. Results Policy Implications The statistically significant results support the hypothesis that implicit doctor recommendations and hospital en- Our analysis provides evidence that two types of infant dorsements reduce search costs and increase brand loy- formula marketing may be effective in promoting formula alty. Those receiving free samples in hospital discharge choice or switching. Endorsement by the mother’s medi- gift packs were more likely to use that formula when their cal care givers (the birth hospital via gift packs or a doctor) infant was a month old, and mothers who chose the encourages choice of the promoted formula and loyalty to formula fed at one month because of a doctor’s recom- it; web marketing discourages brand loyalty. Only com- mendation were less likely to make non-health related panies which have trained staff and other resources needed switches. to market to hospitals and doctors can engage in professional endorsement types of marketing. In contrast, any The significant results also support the hypothesis that formula company can engage in web advertising. Advo- some marketing activities increase the likelihood that a cates working to eliminate hospital gift bags should ex- mother switches the formula, i.e., counteract brand loy- pect significant opposition from the companies that en- alty, presumably by reducing switch costs. Mothers who gage in this type of marketing. Those working to restrict chose a formula for marketing reasons at month one and other forms of marketing are likely to be met with resis- mothers who both chose the hospital formula at month one tance from a broader contingent of companies that sell and had prenatal exposure to formula information on the formula in the U.S.

REFERENCES Howard, C.R., F.M. Howard, R. Lawrence, E. Andresen, E. DeBlieck, and M. Weitzman (2000), “Office Pre- American Academy of Pediatrics Section on Breastfeeding natal Formula Advertising and its Effect on (2005), “Breastfeeding and the Use of Human Milk,” Breastfeeding Patterns,” Obstetrics and Gynecology, Pediatrics, 115 (2), 496–506. 95 (2), 296–303. Cutler, B.D, R.F. Wright (2002), “The U.S. Infant For- Nelson, P. (1970), “Information and Consumer Behav- mula Industry: Is Direct-to Consumer Advertising ior,” Journal of Political Economy, 78 (2), 311–29. Unethical or Inevitable?” Health Marketing Quar- Rosenberg, K.D., C.A. Eastham, L.J. Kasehagen, and terly, 19 (3), 39–55. A.P. Sandoval (2008), “Marketing Infant Formula Darby, M.R. and E. Karni (1973), “Free Competition and Through Hospitals: The Impact of Commercial Hos- Optimal Amount of Fraud,” Journal of Law & Eco- pital Discharge Packs on Breastfeeding,” American nomics, 16 (1), 67–88. Journal of Public Health, 98 (2), 290–95. Fein, S.B., J. Labiner-Wolfe, K. Shealy, R. Li, J. Chen, Shum, M. (2004), “Does Advertising Overcome Brand and L.M. Grummer-Strawn (2008), “Infant Feeding Loyalty? Evidence From the Breakfast-Cereals Mar- Practices Study II: Study Methods,” Pediatrics, 122 ket,” Journal of Economics & Marketing Strategy, 13 (Supple 2), S28–S35. (2), 241–72.

For further information contact: Sara Fein Food and Drug Administration 5100 Paint Branch Parkway College Park, MD 20740–3835 Phone: 301.436.1824 Fax: 301.436.2637 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 41 PROMOTING SAFETY VERSUS PREVENTING AGGRESSIVENESS: A COMPARISON BETWEEN GAIN AND LOSS FRAMED MESSAGES IN SAFE DRIVING CAMPAIGNS

Georgiana Craciun, Loyola College in Maryland, Baltimore Jason Q. Zhang, Loyola College in Maryland, Baltimore Dongwoo Shin, University of Seoul, Korea

SUMMARY achieve the same end goal by focusing on avoiding possible negative consequences of driving such as miss- In 1988, California initiated the longest running aggres- ing a highway exit or being involved in an accident. sive driving media campaign, Smooth Operator, which has since been adopted and is still running in the District According to regulatory focus theory (Higgins 1996, of Columbia, Maryland and Virginia. In 1998, nine U.S. 1997), people strive to achieve their goals through two states introduced legislation that provided penalties for distinct self-regulatory systems – promotion and preven- aggressive driving. According to a 2003 AAA Mid- tion – that oversee their actions. When people are more Atlantic Transportation Poll, motorists in the Washington focused on their ideal self (e.g., aspirations), they develop area rated aggressive driving as the number one threat to the promotion system and rely on approach strategies to highway safety and nearly eight out of ten motorists move closer toward positive end states. On the other hand, considered aggressive drivers to be a greater danger than people with a strong ought self (e.g., responsibilities) terrorists. develop the prevention system and rely on avoidance strategies to stay away from negative end states. There- Aggressive driving is defined as the operation of “a motor fore, it is expected that an individual with a strong promo- vehicle in a manner which endangers or is likely to tion predisposition may focus on the positive conse- endanger people or property” (Martinez, U.S. House of quences of driving (e.g., reaching the destination faster), Representatives 1997). It includes speeding, tailgating, while ignoring any negative consequences (e.g., possible failure to yield, weaving in and out of traffic, making accident), which may lead to more aggressive driving improper and unsafe lane changes, running stop signs, and behaviors. On the other hand, an individual with a strong red lights, making hand and facial gestures, honking and prevention predisposition may focus on the negative headlight flashing. Research in both social psychology consequences of driving, which may lead to safer driving. and marketing studied the antecedents of aggressive driv- Study 1 measured the correlation between self-regulation ing with mixed results. Little work has been done in the traits and aggressive driving intentions. Sixty-one under- marketing literature addressing the effectiveness of mes- graduate students completed first a questionnaire that sages targeted at aggressive drivers. This paper uses measured regulatory trait (Higgins et al. 2001), and then regulatory focus theory (Higgins 1997) to understand a questionnaire that measured aggressive driving intent aggressive driving and identify the most effective mes- (Bone and Mowen 2006; Goldstein and Mosel 1958; sages targeted at preventing this behavior. Krahe and Fenske 2002) and demographic characteristics. Data were examined using a series of regression Driving is seen here as a goal-oriented behavior. When analyses with self-regulatory trait, age, and gender as individuals prepare to drive to a certain destination, they predictors. When only the demographic predictors were tend to set and follow specific goals. They may develop used, gender had a significant impact on aggressive driv- driving plans about the time of departure and arrival, other ing intent (ß = -.260, p = .05). Results indicated that male necessary stops, and identify the route most suitable for respondents were more likely to engage in aggressive their trip, all of which involve various stages of goal- driving than female respondents, replicating prior find- directed actions and require an effective self-regulatory ings (Krahe and Fenske 2002). Nevertheless, when self- system. For example, some drivers may achieve their end regulatory trait was also introduced in the regression goal (arriving at the destination) by focusing on positive model, the effect of gender became non-significant (ß = strategies such as staying ahead of the other drivers or -.092, p = .48) and self-regulatory trait was the only speeding to catch the green light. Other drivers may significant predictor (ß = 0.418, p = .001). Results show

42 Marketing and Public Policy Conference Proceedings / 2009 that promotion oriented individuals are more likely to F(1, 170) = 4.716, p = 0.027. Promotion oriented partici- drive aggressively, while the opposite is true for preven- pants (M = 3.81) were more persuaded by the gain-framed tion oriented individuals. message than prevention oriented participants (M = 3.22, F(1, 170) = 7.958, p = 0.005). On the other hand, preven- Study 2 focused on determining the most effective mes- tion oriented participants (M = 3.564) showed higher sages in targeting these two types of drivers. Two types of average attitude scores toward loss-framed message than messages were tested: (1) loss-framed messages that promotion oriented participants (M = 3.215), but the focus on the negative consequences of aggressive driving difference was statistically not significant (F(1, 170) = (“Speeding puts other lives on the road at risk”), and (2) 0.087, p = .769). gain-framed messages that highlight the positive consequences of avoiding aggressive driving (“Slowing down In conclusion, these two studies extend our understanding saves other lives on the road”). Regulatory focus theory of aggressive driving and the type of messages that are (Higgins 1997; Higgins et al. 2001) posits that people with effective in preventing it. There are two major findings. a promotion orientation are more likely to focus on ideals First, it is the people who are more promotion oriented and, therefore, may find gain-framed messages more who are also more likely to drive aggressively. It is persuasive. On the other hand, people with a prevention possible that gender was a significant predictor of aggres- orientation focus on responsibilities and obligations and, sive driving in previous studies because of a weak corre- therefore, may perceive a loss-framed message that de- lation between gender and regulatory focus. That is, on scribes the negative consequences of aggressive driving average, male students might be more promotion oriented to be more persuasive. One hundred and seventy-five because the typical socialization process of boys tends to undergraduate students were randomly assigned to one of emphasize adventure and achievement. On the other hand the four conditions in a 2 (regulatory orientation: promo- girls are more likely to undergo a socialization process, tion vs. prevention) x 2 (message framing: gain vs. loss) which emphasizes safety and security. Such a relationship between-subjects experimental design. Attitude toward between gender and regulatory focus might be one of the the advertisement was measured on 7-point semantic underlying reasons why gender became non-significant differential scales. Data were analyzed using a series of when the regulatory orientation variable was entered in ANOVAs. Gender was a significant predictor (p = .008) the model. Second, prevention campaigns should mainly and was left in the model as a covariate. Female partici- target people who are promotion oriented and do that by pants appeared to be more persuaded (M = 3.72) by the using gain-framed messages, which have been demon- messages than male participants (M = 3.37). Results strated to work more effectively with these individuals. support the hypothesized interaction between an References are available upon request. individual’s regulatory orientation and message framing,

For further information contact: Georgiana Craciun Loyola College in Maryland 4501 N. Charles St. Baltimore, MD 21210 Phone: 410.617.5727 Fax: 410.617.2117 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 43 SPECIAL SESSION

CURRENT ISSUES AND RESEARCH ON HEALTH LITERACY: MEASUREMENT AND POLICY DIRECTIONS

Session Chair: Jack Swasy, American University, Washington, D.C.

Moderator: Kathryn Aikin, Food and Drug Administration, Washington, D.C.

Presenters: Linda Squiers, RTI International, Washington, D.C. Amie O’Donoghue, Food and Drug Administration, Washington, D.C. Linda Neuhauser, University of California, Berkeley

SUMMARY improper product use. Compliance with medication is poor, as patients are unable to read or understand The Department of Health and Human Services, in its prescription labels and drug information sheets (Kalichman Healthy People 2010 initiative, defines health literacy as et al. 1999; Williams et al. 1998). This topic has significance “the degree to which individuals have the capacity to not only for government policy, as illustrated by the joint obtain, process, and understand basic health information call for further literacy research by the National Institutes and services needed to make appropriate health deci- of Health, the Association for Healthcare Quality and sions.” A 2004 study by the IOM indicated that low health Research and the Centers for Disease Control, but also for literacy is a wide spread problem, affecting more than 90 companies that produce healthcare materials and products. million adults in the United States. It is estimated that about 43 percent of adults demonstrate only the most basic The objectives for this session are to: (1) Discuss the or below-basic levels of prose literacy (U.S. Department concept of health literacy and the evolution in approaches of Education 2005). to the measurement of literacy, (2) Discuss the usefulness of health literacy measures in the testing of marketing Illiteracy has recognized consequences for health of the communications such as direct-to-consumer advertising, individual and health care resources (Gordon et al. 2002). and (3) Explore strategies to translate health literacy Patients with low health literacy may be at risk for higher measurement and communication interventions into medication-related adverse events associated with policies.

44 Marketing and Public Policy Conference Proceedings / 2009 Special Session: Current Issues and Research on Health Literacy: Measurement and Policy Directions

CREATING A SKILLS-BASED HEALTH LITERACY MEASURE

Lauren A. McCormack, RTI International, Research Triangle Park Nancy D. Berkman, RTI International, Research Triangle Park Linda Squiers, RTI International, Washington, D.C. Dean Schillinger, University of California, San Francisco Judith Hibbard, University of Oregon, Eugene Janet Ohene-Frempong, J.O. Frempong & Associates, Inc., Elkins Park Rahima Jan Gates, University of California, San Francisco Carla Bann, RTI International, Research Triangle Park Claudia Squire, RTI International, Research Triangle Park Tania Fitzgerald, RTI International, Research Triangle Park

SUMMARY website) that can be used to measure the task, and lastly, (3) the mode of data collection is chosen based on what is Instruments commonly used to measure health literacy required to measure the task using the stimuli. We re- include the Wide Range Achievement Test (WRAT), the viewed tasks and stimuli both individually and as a set to Rapid Estimate of Adult Literacy in Medicine (REALM), ensure representation of different health literacy domains, and the Test of Functional Health Literacy in Adults levels of difficulty, and stimuli format (e.g., print, inter- (TOFHLA). These instruments emphasize reading ability personal communication) that most Americans typically with an inattention to other dimensions of literacy such as encounter. We categorized health literacy tasks into three document literacy, verbal competency, and numeracy. levels: (1) understanding; (2) applying to specific situa- Other drawbacks of existing instruments include the tions; and (3) integrating into more comprehensive care lengthy time needed to administer them; the need to have management. We identified a range of stimuli to measure an interviewer administer them; and their focus on mea- whether individuals can accomplish these tasks. suring health literacy in relation to acute, episodic care – a decreasingly frequent context of the current health Next Steps and Applications system. The National Assessment of Adult Literacy (NAAL) overcomes some of these limitations, but is not We will test the items using multiple, iterative rounds of publicly available. The purpose of this project is to de- cognitive interviews with modifications made based on velop an instrument to measure health literacy that over- these interviews. Then, we will conduct a pilot test with a comes existing limitations and will be made widely avail- representative sample of 300 participants using an internet- able. based panel. We will assess the reliability and validity of the items as well as their psychometric properties with From our review of the literature and input from experts, scales and subscales proposed based on factor analyses. we defined the domains of health literacy to include print The resulting items could be used in population-based (prose, document, and numeracy), oral (listening, speak- surveys to assess health literacy levels and in clinical ing), and computer/web. We established a hierarchical intervention studies to promote improvements in health approach to instrument development: (1) tasks are the literacy. Results from these studies could help policy most essential element as they measure skills, (2) fol- makers develop social marketing interventions to better lowed by selection of stimuli (e.g., reading passage, chart, meet the needs of their intended audiences.

Marketing and Public Policy Conference Proceedings / 2009 45 Special Session: Current Issues and Research on Health Literacy: Measurement and Policy Directions

UNDERSTANDING RISK COMPREHENSION IN DIRECT-TO- CONSUMER ADVERTISING

Kathryn Aikin, Food and Drug Administration, Washington, D.C. Amie O’Donoghue, Food and Drug Administration, Washington, D.C. Helen Sullivan, Food and Drug Administration, Washington, D.C. Jack Swasy, American University, Washington, D.C.

SUMMARY tures of this study. First, viewing time on the first page (the “promotion” or “display” page) and the second page (the This presentation will discuss findings from the first of brief summary) of the prescription drug advertisement three studies conducted by FDA which examined con- was recorded as well as participant’s (back and forth) sumer use and comprehension of the brief summary page paging behavior between the two pages of the drug in print “direct-to consumer” (DTC) prescription drug advertisement. Additionally, the participant’s reading advertisements. speed was estimated using times on several instructional pages. Second, at the end of the session, the Rapid Estima- In study one, 800 participants – sufferers (80%) and tion of Adult Literacy in Medicine (REALM) was admin- caregivers (20%) – in three medical conditions (asthma, istered to provide a measure of health literacy. We believe cholesterol and weight control) were recruited at 16 that this research is the first to utilize REALM in a study locations where they viewed print advertisements on a of DTC advertising and to assess viewing behavior and computer in a self-paced manner. One of the advertise- comprehension of risk information contained in the pro- ments was a two-page DTC advertisement which pre- motion page as well as more detailed risk information sented a new prescription treatment for their condition. contained in the brief summary of DTC print advertise- Participants were randomly assigned to one of two ver- ments. sions of the advertisement (either a “low” or “high” risk profile). About one third of the sample was selected to The second study extends the first investigation to include have an educational level of high school or less. numeracy skills and consumer comprehension of quantitative displays of side effect information in the brief Risk comprehension was assessed along with other vari- summary. Time permitting, preliminary results of this ables of interest such as interest in the product, health study may also be discussed. status and demographics. There are two important fea-

46 Marketing and Public Policy Conference Proceedings / 2009 Special Session: Current Issues and Research on Health Literacy: Measurement and Policy Directions

STRATEGIES TO ADVANCE HEALTH LITERACY POLICIES IN THE U.S.

Linda Neuhauser, University of California, Berkeley

SUMMARY 4. Setting an organizational mandate that communication meet the literacy and other needs of the intended The approximately 43 percent of U.S. adults with limited audiences. literacy are among those with the poorest health outcomes and those who incur the highest health care costs. Signifi- 5. Establishing “clear health communication” principles cant scientific evidence indicates that health literacy fac- and assessment criteria drawn from the literature and tors are critical to effective health communication, but are from US Department of Health and Human Services not always well assessed or addressed. As discussed in guidelines. this session, new measures and interventions are helping to determine people’s health literacy competencies, and 6. Evaluating progress toward goals, with organiza- ways to meet their needs for better communication. How- tional oversight. ever, over 250 studies indicate that health communication materials are written at reading levels greatly exceeding Examples of Actions to Advance Communication Rel- the average literacy skills of adults (7–9th grade) in the evant to Users’ Health Literacy U.S. It has been difficult to translate knowledge about health literacy into interventions and policies that will 1. Federal: Surgeon General’s Office’s Workshop on improve health communication on a large-scale. Guid- Health Literacy and Town Halls; National Cancer ance from translational research and systems science Institute’s usability guidelines; Food and Drug provide useful frameworks and approaches to meet this Administration’s Risk Communication Advisory goal. This presentation focuses on approaches that can be Committee; and the Centers for Disease Control and used to move health literacy measurement and interven- Prevention’s Simply Put guidelines. tions – including those of the workshop co-presenters – into action at multiple levels. 2. California: Office of the Patient Advocate; Depart- ment of Managed Health Care; and Department of Suggested Strategies Health Care Services.

1. Convening an expert group to advise on setting and Conclusions meeting health literacy goals. Limited health literacy is a critical national challenge. 2. Assessing the organization’s capacity to design health New measurement and communication approaches have communication relevant to the skills of the intended great potential to address the problem – if implemented on audiences. a large-scale. The U.S. Surgeon General’s Office should develop and implement a “Call to Action” to advance 3. Examining the current quality of health communica- national health literacy policies. References are available tion and defining a strategic plan for improvement. upon request.

For further information contact: Jack Swasy American University 4400 Massachusetts Ave. Washington, D.C. 20016–8044 Phone: 202.885.1974 ♦ Fax: 202.885.1992 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 47 20th ANNIVERSARY SESSION

BEHAVIORAL TARGETING BY MARKETERS: BOOM OR DOOM?

Session Chair: Stephen Brobeck, Consumer Federation of America, Washington, D.C.

Panelists: Jessica L. Rich, Federal Trade Commission, Washington, D.C. Pablo Chavez, Google, Inc., Washington, D.C. Susan Grant, Consumer Federation of America, Washington, D.C. George R. Milne, University of Massachusetts, Amherst

SUMMARY online advertisers can best protect consumers’ privacy while collecting information about their online activities. Online behavioral advertising employs the Internet’s capacity to track an individual’s online behavior (e.g., In this session, leaders of the consumer movement and key websites visited, clicks and pages viewed) then to use this speakers from the FTC, industry, and academia will information to directly deliver advertising content deemed explore the ramifications of emerging marketing methods to be relevant to his or her interests. For marketers, on the Internet and the questions they raise about con- behavioral targeting clearly affords the potential to de- sumer privacy. velop and implement more efficient media plans. It is no surprise, therefore, that marketer investments in this mode • Stephen Brobeck, is Executive Director of the Con- of delivery are increasing significantly, with behavioral sumer Federation of America, which represents over advertising expected to increase by over ten times in the 300 consumer groups and is the nation’s largest current six-year period (from $350 million in 2006 to an consumer advocacy organization. He holds a PhD. in expected $4.4 billion in 2012 (eMarketer 2009). American Studies from the University of Pennsylva- nia, and frequently offers testimony before Congress However, the detailed tracking of consumers’ activities and interviews to national media. Dr. Brobeck has makes this a controversial marketing practice. Debates also served as a university professor, as a director of have ensued about the benefits and costs to consumers, as the Federal Reserve Bank of Richmond, and as a well as the level and nature of consumer protections that board member of several nonprofit organizations. He are needed. Consumer advocates, industry representa- has published considerable research, including edit- tives, technology experts, government regulators and ing the significant tome, The Encyclopedia of the academics are all involved in the discussion. On the one Consumer Movement. hand, behavioral advertising offers consumers content that is personalized to their needs and interests which • Jessica L. Rich is Assistant Director in the Federal many people value and also enables free online content Trade Commission’s Division of Privacy and Iden- that is supported by advertising. However, it also raises tity Protection. She oversees a wide variety of matters privacy concerns, such as the lack of transparency to related to consumer privacy and data security, includ- consumers of marketers’ data collection practices and the ing investigations, enforcement, rulemakings, work- risk that the information that is collected including per- shops, surveys, reports, testimony to Congress, con- sonal or sensitive information (e.g., data related to fi- sumer and business education, and general policy nances, health) might be misused. The issue of behavioral development. In recent years, such matters have targeting is one that the Federal Trade Commission (FTC) included development of the FTC’s policies and has examined periodically over the last several years, and guidance regarding online behavioral advertising. recently issued a report outlining a set of revised principles to govern the self-regulation of online behavioral • Susan Grant is Director of Consumer Protection at advertising (FTC 2009). The key issue concerns how the Consumer Federation of America. She has many

48 Marketing and Public Policy Conference Proceedings / 2009 20th Anniversary Session: Behavioral Targeting By Marketers: Boom or Doom?

years of experience in consumer protection and be- the town hall entitled “Ehavioral Advertising: Track- fore joining CFA served as the Vice President and ing, Targeting, and Technology” that was held by the Director of the Fraud Center at the National Consum- Federal Trade Commission in 2007. ers League. At CFA, she focuses on privacy, deception, online safety and security, fraud as well as • Pablo Chavez is Senior Policy Counsel for Google. electronic and mobile commerce. He advises on, helps to develop, and advocates in favor of a range of public policy issues including • George R. Milne is Associate Professor of Marketing privacy, content regulation, and Internet censorship. at the Isenberg School of Management at the Univer- Before joining Google, Pablo worked in the U.S. sity of Massachusetts. His research focuses con- Senate as a legal and legislative counsel to the Senate sumer information privacy, database marketing and Commerce Committee and to Senator John McCain. interactive marketing issues. He has investigated Prior to this, he was an attorney at the law firm of exchange relationships between marketers and con- Wilson Sonsini Goodrich & Rosati. He received his sumers, privacy notice compliance and readability, law degree from Stanford Law School and his under- as well as identity theft. Professor Milne testified at graduate degree from Princeton University.

Marketing and Public Policy Conference Proceedings / 2009 49 SPECIAL SESSION

THE ROLE OF CONSUMER RESEARCH IN POLICY DEVELOPMENT AND EVALUATION AT FEDERAL AGENCIES: SOME RECENT EXAMPLES

Session Chair: Robert Letzler, Federal Trade Commission, Washington, D.C.

Discussant: Jeanne Hogarth, Federal Trade Commission, Washington, D.C.

SESSION OVERVIEW tion, medical products, product safety, mortgages, and other credit products. Government uses consumer strate- This session focuses on how consumer research can assist gies to help environmentally conscious people act on their the work of public agencies. Agencies need a practical ideals and reduce the spread of communicable diseases. understanding of how consumers understand and react to Income tax withholding is a consumer-aware decision communications in order to accomplish their missions. structure that helps people avoid writing a massive check Consumer research can benefit government programs that in April and avoids a challenging budgeting problem. The take actions like: Medicare Prescription Drug benefit program required consumers to take action to get benefits and to navigate a • Informing consumers about options and consequences large and confusing menu of options. In law enforcement so that they can make better choices and so that where scholars report that telling specific trouble makers markets reach better outcomes. Sometimes govern- that their next transgression is likely to lead to a harsh ment mandates disclosures like automobile fuel punishment can be highly effective. The FTC does signifi- economy labels or provides resources like cant consumer research to determine what claims con- consumer.gov. In other contexts, government re- sumers take away from advertising campaigns before it quires that firms disclose certain facts – like drug side determines whether the ad is deceptive as the reasonable effects – or regulates the use of terms like “recy- consumer interprets it. clable” or “cheese,” but leaves firms significant lati- tude in how they present those claims. Marketing academics are interested in how familiar consumer research techniques can be applied in new, policy- • Limiting the claims or decision structures offered in relevant contexts and how agencies’ work reveals new the marketplace; and deciding when marketing is so opportunities to use these techniques and how those deceptive or decision structures so unfair that they techniques have to be supplemented or positioned in a should take enforcement action. policy advising context. Practitioners are interested in seeing how experimental and econometric techniques are • Designing program communication materials and brought to bear on practical policy problems to yield decision structures to deliver program benefits to the insights about opportunities to improve programs. We clients it targets and otherwise achieve the program’s believe this session will lead to cross-fertilization of goals. interest to scholars inside and outside the academy.

• Evaluating when information or communication strat- The session looks at how consumer research can be egies – like crashworthiness labels – are sufficient applied to improve effectiveness and to spot and avoid and when more intrusive strategies – like automobile unintended consequences in a variety of contexts: safety regulations – are necessary. • Agencies seek to require product labels that will help These challenges are ubiquitous. Effective consumer strat- consumers make well-informed choices about things egies are important in policy areas including food, nutri- like health and energy conservation.

50 Marketing and Public Policy Conference Proceedings / 2009 Special Session: The Role of Consumer Research In Policy Development and Evaluation at Federal Agencies: Some Recent Examples

• Agencies need to design choice structures and com- “Exploring the Effects of Decision Structure, Stakes, and munications that will help people make appropriate Information Provision in a Fraud Case” analyses a data- choices about when to claim benefits. base documenting which customers opted out of subscription products from a fraudulent firm during the course of • Agencies need to understand how practices in the its normal operations, and then after customers received marketplace distort consumer decisions in ways that either an opt in or opt out notice during an FTC lawsuit are so troubling that they merit enforcement action. against the firm. The study explores how termination rates vary depending on whether the customer got (1) solely a To these ends, agencies both commission proactive con- small or large charge on their bank statement; (2) an opt in sumer research to design effective interventions and do notice; or (3) an opt out notice. Far fewer customers retrospective analysis of cases to better understand les- remained when their choice was structured as an opt out sons from their work load. This session includes both than did similar customers whose choice was an opt in – prospective and retrospective policy design and analysis so customers’ outcomes were very sensitive to the court work. Two of the three papers take a retrospective look at ordered decision structure. It also explores how these opt field data, while one study is a prospective copy test. out rates vary depending on a customer’s demographic group. Session Presentations “Mercury Advisories: Information, Education, and Fish “Developing Mandatory Energy Labels That Do Not Consumption” examines responses to a national FDA Mislead or Bias Consumers: Lessons From Consumer advisory that urged at-risk individuals – women who may Research at the Federal Trade Commission” reports on a become pregnant, pregnant women, nursing mothers, and recent randomized, controlled experiment conducted at young children – to limit store-bought fish consumption the FTC to assess alternative EnergyGuide labels. The due to the dangers of methyl-mercury. We investigate EnergyGuide label provides energy efficiency and oper- consumer response using both parametric and nonpara- ating cost information for a large number of common metric methods. Some targeted consumers significantly household appliances, such as refrigerators, dishwashers, reduced canned fish purchases as a result of the advisory, and water heaters. Researchers involved in the FTC’s suggesting that information-based policies can achieve EnergyGuide research project will discuss the purpose, the issuing agency’s goals. Education and newspaper design, and outcome of the recent consumer research. readership were important determinants of response, sug- Two sets of analyses, with broad implications for labeling gesting that information acquisition and assimilation are policies across numerous products will be featured. First, key factors for risk avoidance. While some groups re- the effect of label design on reported willingness-to-pay duced consumption as a result of the advisory, we do not estimates for energy efficiency will be explored. Second, find a response among the relatively large group of at-risk the role of dollars as a metric for making decisions among households which met neither the education nor reader- substitute products and across different products will be ship criteria. The advisory also had unintended spillover discussed. Finally, consistent with analyses presented in effects; some consumers not considered at-risk reduced the Federal Register, the researchers will discuss how the consumption in response to the advisory. research was used in the policy process.

Marketing and Public Policy Conference Proceedings / 2009 51 Special Session: The Role of Consumer Research In Policy Development and Evaluation at Federal Agencies: Some Recent Examples DEVELOPING MANDATORY ENERGY LABELS THAT DO NOT MISLEAD OR BIAS CONSUMERS: LESSONS FROM CONSUMER RESEARCH AT THE FEDERAL TRADE COMMISSION

James Hilger, U.S. National Marine Fisheries Service, National Oceanic and Atmospheric Administration, Washington, D.C. Janis Pappalardo, Federal Trade Commission, Washington, D.C. Manoj Hastak, American University, Washington, D.C.

SUMMARY treatments were included to study the effects of the EPA and DOE’s Energy Star insignia on qualifying, hypotheti- For decades, marketing and public policy researchers cal products. have questioned the value of information remedies insti- tuted without the benefit of consumer research. And for Differences among label treatments were examined for decades, policy makers have required labels without such dimensions, such as accurate comprehension of energy research. This is unfortunate, because disclosures adopted use characteristics, accurate comprehension of operating with the best of intentions can actually backfire. costs, product quality perceptions, perceptions about the ease of label use, and reported willingness-to-pay esti- Fortunately, consumer research is sometimes undertaken mates. Several open-ended questions are also included in by policy makers to inform mandatory disclosure policy. the study to understand the findings from the quantitative One example of such research is a recent randomized, research. For example, the study included follow-up ques- controlled experiment conducted at the FTC to assess tions to understand how different label designs might alternative EnergyGuide labels. The EnergyGuide label affect differences in reported willingness-to-pay mea- provides energy efficiency and operating cost informa- sures. tion for a large number of common household appliances, such as refrigerators, dishwashers, and water heaters. The The researchers will report on statistical analyses con- research helped to guide the design of a new label. ducted to compare differences within and among competing label treatments and on findings from the open-end Researchers involved in the FTC’s EnergyGuide research questions. Two sets of analyses with broad implications project will discuss the purpose, design, and outcome of for labeling policies across numerous products will be the research, which is based on findings from 4,000 featured. First, the effect of label design on reported respondents of an internet panel. The tested labels in- willingness-to-pay estimates for energy efficiency will be cluded (1) the incumbent label, which highlighted ex- explored. Second, the role of dollars as a metric for pected appliance annual energy usage in kilowatt hours making decisions among substitute products and across (kWh), (2) a new label highlighting annual operating cost different product will be discussed. Finally, consistent measured in dollars, (3) a categorical label utilizing a five- with analyses presented in the Federal Register, the re- star energy efficiency rating, and (4) a modified label searchers will discuss how the research was used in the featuring kWh as the primary disclosure. In addition, policy process.

For more information contact: Janis Pappalardo Federal Trade Commission Mail Drop NJ–4136 600 Pennsylvania Avenue NW Washington, D.C. 20580 Phone: 202.326.3380 E-Mail: [email protected]

52 Marketing and Public Policy Conference Proceedings / 2009 Special Session: The Role of Consumer Research In Policy Development and Evaluation at Federal Agencies: Some Recent Examples EXPLORING THE EFFECTS OF DECISION STRUCTURE, STAKES, AND INFORMATION PROVISION IN A FRAUD CASE

Robert Letzler, Federal Trade Commission, Washington, D.C. Isaac Knowles, Federal Trade Commission, Washington, D.C. Patrick McAlvanah, Federal Trade Commission, Washington, D.C. Manoj Hastak, American University, Washington, D.C.

SUMMARY explore who opts out, when, and why. We examine whether customers with higher priced subscriptions opted The FTC routinely notifies customers about their rights to out sooner. We will analyze how neighborhood demo- choose whether to continue a subscription or apply for graphics affect opt out rates. redress. Our case work confronts us with a pool of customers who have enrolled in inadequately disclosed After FTC sued the firm, a magistrate ordered that cus- “negative option” subscriptions that charge consumers tomers enrolled for less than five months receive mailings periodically until they opt out. This project explores these allowing them to opt-in to continue their enrollments, policy issues using field data from an FTC fraud case. It while customers enrolled longer got similar, opt-out no- uses evidence from one fraud case to better understand tices requiring a reply to discontinue membership. There what demographic groups experience the most harm, how were large and striking differences between the percent- the size of charges affect opt out rates, and how the age of opt-out customers who cancelled and the percent- decision structure and information content of court-or- age of opt-in customers who let their memberships lapse. dered notices change outcomes. Further, we measure the effects of the court ordered opt- We have customer account data about people who re- out notice by comparing the rate of customer exits when sponded to fraudulent phone sales calls and subscribed to they received an opt out notice in the Nth billing cycle of programs that debited their bank accounts monthly until their subscription to the exit rate for an earlier cohort of the customer cancelled their subscription. We focus on customers who solely received a charge to their bank customers who were paying for products and services that statements during their Nth billing cycle. We examine the they would never use. For example, we will study opt out effects of demographics and subscription type on the decisions by people who were paying for unlimited long impact of these opt-out notices. distance service yet making no free calls. Opting out as soon as possible was generally the right choice. We

For more information contact: Robert Letzler Federal Trade Commission Mail Drop NJ–4136 600 Pennsylvania Avenue NW Washington, D.C. 20580 Phone: 202.326.2912 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 53 Special Session: The Role of Consumer Research In Policy Development and Evaluation at Federal Agencies: Some Recent Examples MERCURY ADVISORIES AND HOUSEHOLD HEALTH TRADE-OFFS

Jay P. Shimshack, Tulane University, New Orleans Michael Ward, Australian National University, Australia

SUMMARY two years after the advisory (2000–2002). Detailed information on more than 5,300 unique products, including In the last decade, concern has arisen that methylmercury species type, allows us to translate home fish consumption from fish consumption may pose a significant threat to into household mercury and omega-3 intakes. In order to children’s neurological development. In 2001, the U.S. isolate the effect of the advisory, we control for confound- Food and Drug Administration (FDA) reacted to mount- ing factors using the quasi-experimental changes-ining evidence of risk by releasing an advisory instructing changes model. Changes-in-changes is a non-parametric pregnant women and households with young or nursing extension of the commonly utilized difference-in-differ- children to limit fish consumption. However, as the advi- ences method, but instead of identifying an average treat- sory noted, a moderate amount of fish consumption pro- ment effect alone it identifies the entire distribution of vides significant health benefits to both adults and young treatment effects. children, particularly in the form of omega-3 fatty acids. We find that at-risk consumers significantly reduced their The conventional wisdom is that improved information mercury exposure, and there is a particularly strong de- will enhance consumer welfare. However, the inherent cline at the upper tail of the per capita mercury intake tension between seafood benefits and mercury risks has distribution. If there were no health trade-offs, this is caused many to speculate that the FDA advisory may have positive for public health. However, there is a substantial harmed public health, on net, if individuals reduced con- countervailing omega-3 cost to this benefit. We find that sumption of all fish rather than just high mercury fish. at-risk consumers significantly reduced their omega-3 Despite the debate, no systematic revealed preference intake, and this decline occurs everywhere along the per analysis of the ensuing health trade-offs yet exists. We capita omega-3 distribution, including the lower tails. therefore analyze advisory-induced changes in the quan- This is negative for public health. In short, at-risk consum- tity and composition of household-level fish consump- ers reduced consumption of all fish in response to the tion. advisory. These consumers did not differentially avoid high mercury fish, nor did they substitute away from high We answer these questions using household-level scan- mercury species into low mercury, high omega-3 species. ner data and cutting-edge statistical techniques. We have Back-of-the-envelope calculations suggest that net advi- every packaged supermarket fish purchase from nearly sory benefits were approximately zero. 15,000 households in the year before the advisory and the

For more information contact: Jay P. Shimshack Department of Economics Tulane University 206 Tilton Hall 6823 St. Charles Avenue New Orleans, LA 70118 Phone: 504.862.8353 Fax: 504.865.5869 E-Mail: [email protected]

54 Marketing and Public Policy Conference Proceedings / 2009 SPECIAL SESSION

RISK, VULNERABILITY, AND IMPROVISATION IN NATURAL DISASTER PLANNING, MITIGATION, AND RECOVERY: IMPLICATIONS FOR MARKETING & PUBLIC POLICY Session Co-Chairs: Stacey Menzel Baker, University of Wyoming, Laramie Ingrid M. Martin, California State University – Long Beach

Session Participants: Stacey Menzel Baker, University of Wyoming, Laramie Leila Borders, Kennesaw State University Deirdre T. Guion, North Carolina Central University, Durham Ronald Paul Hill, Villanova University Ingrid M. Martin, California State University – Long Beach Wade E. Martin, California State University – Long Beach Carol Raish, U.S. Forest Service, Albuquerque Lab Debra L. Scammon, University of Utah, Salt Lake City

Discussant: Ronald Paul Hill, Villanova University

SUMMARY The session begins with a discussion of alternative understandings of natural disaster. Natural disasters have been Today global citizens are told they live in a “risk soci- understood as (1) natural events, (2) “fleeting” events ety” and their risk for experiencing natural disaster is which disrupt routines, and (3) social incidents as disasters increasing (Beck 2004). A variety of explanations for are socially constructed and rooted in the interaction be- this increasing risk have been offered including where tween natural events and social structures (Perry 2007). people choose to live, the structures in which some live, Each understanding leads to different market and policy the protective measures in place or not in place in responses across different phases of disaster (Baker 2009). communities, the material assets humans own, and, more controversial, climate change. Regardless of the Next, three papers are presented which look at disaster from reason, in some cultures at least, there is a strong desire different perspectives and derive insights from disasters to control the environment and mitigate the risk from created by different types of natural hazards (hurricanes, natural disasters. However, the environment cannot firestorms, and tornados). The first paper by Guion, always be controlled and humans cannot always miti- Scammon, and Borders focuses on the tourism industry, gate risk. specifically planning before a disaster and improvisation afterwards. The second paper by Martin, Martin, and Raish From 1997–2006 3,670 disasters caused by natural examines how risk perceptions mitigate perceived vulner- events were reported, over 2.7 billion people were ability and risk-reduction behaviors in areas prone to forest affected, over one million people were killed, and eco- fires. Finally, the third paper by Baker and Hill examines nomic impacts were estimated at U.S. $800 billion how definitions of risk and vulnerability impact how re- (International Federation of Red Cross and Red Cres- sources are distributed in natural disaster recovery. cent Societies 2007). The purpose of this session is to examine the role of consumers, marketers, and public policy makers in times of natural disaster.

Marketing and Public Policy Conference Proceedings / 2009 55 Special Session: Risk, Vulnerability, and Improvisation in Natural Disaster Planning, Mitigation, and Recovery: Implications for Marketing & Public Policy THE INEVITABILITY OF DISASTERS: THE ROLE OF PLANNING AND IMPROVISATION IN SUSTAINABLE TOURISM

Deirdre T. Guion, North Carolina Central University, Durham Debra L. Scammon, University of Utah, Salt Lake City Leila Borders, Kennesaw State University

SUMMARY wildfire, hurricane and avalanche seasons, a deeper understanding of destination specific issues is warranted. This study uses interview data from business owners in Preliminary analysis suggests that expertise and experi- the aftermath of Hurricane Katrina to explore issues ence, strong devotion to the local community, and corpo- surrounding disaster response and recovery. As a major rate commitment to employees all positively contribute to tourist destination New Orleans serves as a rich source planning for and improvising in the wake of disaster. from which central themes around planning and improvi- Although tourism and tourist destinations are vulnerable sation emerge. Since many tourist destinations (domesti- to disasters, data suggest that businesses in the tourism cally and internationally) are located in areas prone to sector are particularly resilient. natural disasters with peak tourist seasons coinciding with

56 Marketing and Public Policy Conference Proceedings / 2009 Special Session: Risk, Vulnerability, and Improvisation in Natural Disaster Planning, Mitigation, and Recovery: Implications for Marketing & Public Policy UNDERSTANDING THE RISK PERCEPTIONS AND ATTITUDES OF DISASTER SUBCULTURES

Ingrid M. Martin, California State University – Long Beach Wade E. Martin, California State University – Long Beach Carol Raish, United State Forest Service, Albuquerque Lab

SUMMARY become imperative. We examine the ways in which these residential groups develop the mentalities or “constructed Over the years, various factors have emerged that have realities” that structure their views and responses con- contributed to an increased risk of catastrophic wildfires cerning risk and disaster preparedness. Our research com- in residential areas that are surrounded by national forests. pares and contrasts the perspectives of these different Some of the factors include years of suppression of natural communities concerning their risk perceptions, feelings fires (lightning strikes, etc.), an increase in populations of vulnerability and control (or lack of) over this natural living in and around the forests, changes in climate includ- risk over a two-year period. ing prolonged droughts. As the incidence of devastating fires increases, managing the risks posed by these fires has

Marketing and Public Policy Conference Proceedings / 2009 57 Special Session: Risk, Vulnerability, and Improvisation in Natural Disaster Planning, Mitigation, and Recovery: Implications for Marketing & Public Policy UNDERSTANDING RISK AND VULNERABILITY: IMPLICATIONS FOR POLICY MAKING AND RESOURCE ALLOCATION

Stacey Menzel Baker, University of Wyoming, Laramie Ronald Paul Hill, Villanova University

SUMMARY dis-empower survivors of natural disasters. For example, regardless of how vulnerability is conceptualized, the Risk is the objective probability of harm in the future; conceptualization has very real consequences on how whereas, vulnerability is the materialization of risk (e.g., survivors view themselves and how resources are allo- Baker 2009). In this presentation we analyze, within the cated. These decisions ultimately impact survivors’ abili- context of an ethnographic study of a rural community ties to work through their vulnerability. impacted by a tornado and numerous studies from the disaster literature, how understandings of disaster, risk, The presentations will intentionally be kept short so that and vulnerability impact in what form and to whom Ron Hill may lead the audience in what promises to be a resources are allocated. Resource allocation decisions are lively discussion. policy decisions which have the potential to empower or

REFERENCES International Federation of Red Cross and Red Crescent Societies (2007), World Disasters Report: Focus on Baker, Stacey Menzel (2009), “Vulnerability and Resil- Discrimination. Bloomfield CT: Kumarian Press. ience in Natural Disasters: A Marketing and Public Perry, Ronald W. (2007), “What is a Disaster?” in Hand- Policy Perspective,” Journal of Public Policy & book of Disaster Research, Havidàn Rodríguez, Marketing, 28 (Spring), in press. Enrico L. Quarantelli, and Russell R. Dynes, eds. Beck, Ulrich (2004), Risk Society: Toward a New Moder- New York: Springer, 1–15. nity. London: Sage.

For further information contact: Stacey Menzel Baker Ingrid M. Martin Department of Management and Marketing Department of Marketing College of Business College of Business Administration 1000 E. University Avenue California State University – Long Beach University of Wyoming 1250 Bellflower Boulevard Laramie, WY 82071 Long Beach, CA 90840 Phone: 307.766.3734 Phone: 949.632.5070 Fax: 307.766.3488 Fax: 949.481.8204 E-Mail: [email protected] E-Mail: [email protected]

58 Marketing and Public Policy Conference Proceedings / 2009 SPECIAL SESSION

ADOLESCENTS AND RISKY BEHAVIORS

Session Chair: Debra M. Desrochers, University of Notre Dame

Discussant: Les Carlson, University of Nebraska – Lincoln

SESSION OVERVIEW our understanding of how anti-smoking advertising campaigns impact adolescents of various ages. Data were A multidisciplinary review of research from neuroscience, collected from over 900 adolescents aged 12–18 as part of psychology, and marketing indicates that adolescents are a multi-million dollar, statewide, anti-tobacco ad cam- fundamentally different than adults, making them poten- paign. In this study comparisons are also made between tially more vulnerable to certain advertising and promo- subsets of this age group to investigate how the ad cam- tions of products that provide immediate gratification, paign interacts with the impact they have on each other. thrills, and/or social status (Pechmann et al. 2005). Given this call to recognize the unique aspects of this age group, The third paper, by Raymond, Tanner, and Mittelstaedt, the objective of this session is to examine three areas of addresses another area of behavioral concern among adolescent vulnerabilities. For each of these topics, the adolescents which is unwanted teenage pregnancies. The authors highlight decisions made during the transition authors report on a study of 1,053 middle and high school from childhood to adulthood that may have long term students who participated in an abstinence education consequences. program. Their findings provide insights on the effectiveness of such programs, overall, as well as other attitudes The Desrochers and Moore paper focuses on the diet and and skills that might favorably impact the potential suc- nutrition behaviors of adolescents. Despite the increased cess of these programs. prevalence of obesity within this age group, they have received less attention than younger children in the quest The discussion leader is Les Carlson of the University of for programs and incentives to reduce the prevalence of Nebraska – Lincoln, a scholar in the areas of consumer overweight. While they are not subject to regulatory or socialization and environmental marketing/advertising. self-regulatory protections like younger children, or as Professor Desrochers will chair the session. The session cognizant of the consequences of their decisions as adults, should appeal to attendees interested in teens, health and they are increasingly responsible for many diet and nutri- healthy lifestyle behaviors, marketing’s impacts on ado- tion decisions that can impact their long-term health. lescent behaviors, and public policy. Overall, this session Using a nationally representative sample (NHANES 2006) fits well within the conference theme of Marketing, health the authors examine several types of behaviors of teens, knowledge, and health outcomes since all three papers ages 16–19. deal with the socialization and transition of adolescents from the protection awarded to the vulnerable population Another risk factor among youth is smoking. In the of children, to the responsibilities of adulthood and the second paper, Andrews, Netemeyer, and Burton increase health and behavioral decisions made during this period.

Marketing and Public Policy Conference Proceedings / 2009 59 Special Session: Adolescents and Risky Behaviors

EXTENDING THE OBESITY DISCUSSION: TEEN DIETS, BEHAVIOR, AND USE OF NUTRITIONAL INFORMATION

Debra M. Desrochers, University of Notre Dame Elizabeth S. Moore, University of Notre Dame

SUMMARY quences of their decisions (see e.g., Pechmann et al. 2005). Unfortunately, almost 18 percent of teens ages 16– The recent rise in childhood obesity across the United 19 are obese and another 15 percent are overweight, States has been well documented. Advocacy groups, totaling more than five million youth in this age group academics, industry and public health officials are work- who are at increased risk for the adverse health conse- ing to inform the problem and develop interventions that quences associated with adult obesity.1 can help young children and their families make better dietary and lifestyle choices. To date, however, the infor- The goal of this research is to examine the diet and mation that has been developed is largely based on studies nutrition-related behaviors of teens (ages 16–19). The of television advertising and its influence on children 2005–2006 National Health and Nutrition Examination under 12 years old. This leaves the teen population rela- Survey data (NHANES) which is collected by the Centers tively understudied in the marketing domain. In fact, the for Disease Control and Prevention (CDC), National Institute of Medicine (IOM) prepared a review of 155 Center for Health Statistics (NCHS) in the U.S. Depart- studies of marketing and its influence on diet and diet ment of Health and Human Services is used for this behaviors and found that only 49 studied youth, ages 12– analysis. Drawing on this nationally representative sample, 18 (IOM 2006). Further, one of the conclusions of the we analyze adolescents’ diets, their beliefs about their review is that there is insufficient evidence on the impact diets and weight status, how often they eat meals away of marketing on the preferences, purchase requests, and from home, their physical activity and TV, video, and beliefs of youth ages 12–18 (IOM 2006). Despite this, computer use, and how older teens use nutritional infor- food marketers do recognize a difference in this age mation, particularly the Nutrition Facts Panel. We also group. In July of 2008 the Federal Trade Commission show how perceptions and dietary behaviors vary as a (FTC) released a report that shows marketers engage in function of their weight status. For example, among those differential spending, promotional activities, and media who are obese (based on BMI), more than 12 percent placements when considering children versus youth (FTC perceive the healthiness of their diet as excellent or very 2008). good, more than 11 percent perceive that they are more active than others their same age, more than 46 percent Adolescents are a unique group because they are not never use the nutrition facts label, and about 18 percent entirely responsible for all their eating and activity choices perceive their own weight as about right. Our analysis will yet, unlike younger children, they do have considerable inform those who are working to address obesity among freedoms and responsibility in these areas of their lives. adolescents and those working on programs to help teens Further, they are too old to be viewed as vulnerable by make better health and lifestyle choices. This research will advocacy groups yet extensive research shows that they also help to develop a research agenda for scholars inter- are more likely to engage in risky behaviors than adults ested in health policy. but without a full understanding of the long term conse-

REFERENCES National Health and Nutrition Examination Survey Data, Hyattsville, MD: Centers for Disease Control Centers for Disease Control and Prevention (CDC) (2005– and Prevention, available at [http://www.cdc.gov/ 2006), National Center for Health Statistics (NCHS) nchs/nhanes.htm], accessed March 5, 2008. in U.S. Department of Health and Human Services, Federal Trade Commission (2008), Marketing Food to

60 Marketing and Public Policy Conference Proceedings / 2009 Special Session: Adolescents and Risky Behaviors

Children and Adolescents: A Review of Industry and Youth: Threat or Opportunity? Washington, Expenditures, Activities, and Self Regulation. Wash- DC: The National Academies Press. ington, DC: Federal Trade Commission Report to Pechmann, Cornelia, Linda Levine, Sandra Loughlin, and Congress, available at [http://www.ftc.gov/os/2008/ Frances Leslie (2005), “Impulsive and Self-Con- 07/P064504foodmktingreport.pdf], accessed Septem- scious: Adolescents’ Vulnerability to Advertising ber 30, 2008. and Promotion,” Journal of Public Policy & Market- Institute of Medicine (2006), Food Marketing to Children ing, 24 (2), 202–21.

For further information contact: Debra M. Desrochers University of Notre Dame 398 Mendoza College of Business Notre Dame, IN 46556 Phone: 574.631.6136 Fax: 574.631.5255 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 61 Special Session: Adolescents and Risky Behaviors

THE EFFECTS OF ADOLESCENT AGE AND ANTI-TOBACCO AD CAMPAIGN ATTITUDES ON ANTI-SMOKING BELIEFS AND INTENTIONS TO SMOKE

J. Craig Andrews, Marquette University, Milwaukee Richard G. Netemeyer, University of Virginia, Charlottesville Scot Burton, University of Arkansas, Fayetteville

SUMMARY tude such that these attitudes have a stronger effect on beliefs and intentions for older than younger adolescents. Smoking continues to be a major risk factor for youth, and The rationale is that, given potential ceiling (for anti- has aptly been described as a “pediatric disease” by smoking beliefs) and floor effects (for smoking inten- former FDA Commissioner Kessler (Hilts 1995). Today, tions) for young adolescents, such campaign effects are tobacco use in the U.S. accounts for approximately 438,000 likely to have greater opportunity to occur for older deaths each year – 20 percent of all mortalities in the U.S. adolescents. (CDC 2008). Currently, 80 percent of adult smokers began smoking their first cigarette by age 18, and more Data on adolescent age, ad campaign attitudes, anti- than 4,000 youth try their first cigarette each day (CDC smoking beliefs, and smoking intentions were collected 2008). from over 900 adolescents aged 12–18 as part of a multi- million dollar, statewide, anti-tobacco ad campaign. Hier- Thus, efforts to increase our understanding of exactly how archical regression modeling reveals that both ad cam- adolescents process and react to anti-tobacco campaign paign attitude (ACA) and age significantly impact anti- advertising are noteworthy (cf., Andrews et al. 2004; smoking beliefs. Yet, the addition of the ACA x age Pechmann et al. 2003; Siegal and Biener 2000; Wakefield interaction indicates that the favorable effects of the et al. 2003). For example, research shows that ad cam- campaign attitude on anti-smoking beliefs become even paign attitudes, prior trial behavior, and social influence stronger for older adolescents. In turn, ACA, anti-smok- all serve to influence anti-smoking beliefs and smoking ing beliefs, and age are all found to impact adolescent intentions (Andrews et al. 2004). Yet, a review of anti- smoking intentions. However, age again moderates the smoking advertising studies concludes that such advertis- campaign attitude effect, indicating that the favorable ing tends to have a more reliable, positive effect on those impact of the campaign on reducing smoking intentions in pre-adolescence or early adolescence stages (Wakefield becomes even more pronounced for older adolescents. et al. 2003). Moreover, there is a tendency for older, Caucasian, and male youth who are not in school to smoke Thus, our results extend prior findings (e.g., Andrews (Jamieson and Romer 2001; see also Mowery et al. 2004). et al. 2004) by showing that the effects of anti-smoking ad Therefore, a strong argument can be made that adolescent campaigns can vary by an important, additional variable – age will have an impact on anti-smoking beliefs and namely the age of the adolescents. Our results imply that smoking intentions such that younger adolescents (e.g., older adolescents’ anti-smoking beliefs and smoking in- aged 12–14) will have a stronger effect on beliefs and tentions will be more favorably influenced when target- intentions than will older adolescents (e.g., aged 15–18). specific adolescent anti-tobacco ad campaigns are used. Yet, adolescent age may interact with ad campaign atti-

REFERENCES derstanding Adolescent Intentions to Smoke: An Examination of Relationships Among Social Influ- Andrews, J. Craig, Richard G. Netemeyer, Scot Burton, ence, Prior Trial Behavior, and Antitobacco Cam- D. Paul Moberg, and Ann Christiansen (2004), “Un- paign Advertising,” Journal of Marketing, 68 (July),

62 Marketing and Public Policy Conference Proceedings / 2009 Special Session: Adolescents and Risky Behaviors

110–23. Youths,” American Journal of Public Health, 94 CDC (2008), “Youth and Tobacco Use: Current Esti- (February), 331–37. mates,” Centers for Disease, Control, and Preven- Pechmann, Cornelia, Guangzhi Zhao, Marvin E. Goldberg, tion, November, available at: [http://wwwtest.cdc.gov/ and Ellen Thomas Reibling (2003), “What to Convey tobacco/data_statistics/fact_sheets/youth_data/ in Antismoking Advertisements for Adolescents: The youth_tobacco.htm]. Use of Protection Motivation Theory to Identify Hilts, Philip (1995), “FDA Head Calls Smoking a Pediat- Effective Message Themes,” Journal of Marketing, ric Disease,” NY Times (online), March 9, available 67 (April), 1–18. at: [http://query.nytimes.com/gst/fullpage.html?res= Siegal, Michael and Lois Biener (2000), “The Impact of 990CE1D81F3AF93AA35750C0A963958260]. an Antismoking Media Campaign on Progression to Jamieson, Patrick and Daniel Romer (2001), “A Profile of Established Smoking: Results of a Longitudinal Youth Smokers and Smoking,” in Smoking: Risk, Percep- Study,” American Journal of Public Health, 90 tion, & Policy, Paul Slovic, ed. Thousand Oaks, CA: (March), 380–86. Sage Publications, 29–47. Wakefield, Melanie, Brian Flay, Mark Nichter, and Gary Mowery, Paul D., Matthew C. Farrelly, M. Lyndon Giovino (2003), “Effects of Anti-Smoking Advertis- Haviland, Julia M. Gable, and Henry E. Wells (2004), ing on Youth Smoking: A Review,” Journal of Health “Progression to Established Smoking Among U.S. Communication, 8, 229–47.

For further information contact: Craig Andrews Department of Marketing Marquette University 606 N. 13th Street Milwaukee, WI 53201–1881 Phone: 414.288.7181 Fax: 414.288.5757 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 63 Special Session: Adolescents and Risky Behaviors

ABSTINENCE EDUCATION AND PUBLIC POLICY: EFFECTIVENESS OF ABSTINENCE-ONLY PROGRAMS AS SOCIAL MARKETING

Mary Anne Raymond, Clemson University John F. Tanner, Jr., Baylor University, Waco John D. Mittelstaedt, Clemson University

SUMMARY nents. Planned Parenthood (2008) says that abstinence- only programs have spent $1.5 billion on activities that According to the Centers for Disease Control and Preven- “promote alarmist misinformation about sexual health tion, in 2005 teenagers between 15–19 years old gave and force-feed students religious ideology that condemns birth to 414,593 children, which accounts for 10 percent homosexuality, masturbation, abortion, and contracep- of the total live births (National Vital Statistics Report tion. In doing so, they endanger students’ sexual health.” 2007). While the rate of teenage pregnancy has fallen over Conversely, proponents of abstinence programs argue the last 20 years, it is a continued source of public policy that programs delay sexual intercourse, compared with concern. Teenage mothers are less likely than their peers other approaches (Jemmott 2006). Missing from this to finish high school, limiting their economic opportuni- discussion is objective research on the effects of absti- ties. Children born to teenage mothers are more likely to nence education on adolescent perceptions of themselves, be raised in poverty than other children, less likely to have their parents and their peers, their attitudes, intentions and health care, and most likely to become teenage parents, behaviors, and their engagement in other risky behaviors. continuing a cycle of poverty and despair. Research This paper reports on a study of 1,053 middle and high indicates the cost of teenage childbearing in the U.S. is school students who participated in an abstinence educa- about $9.1 billion, annually (Hoffman 2007). tion program. Results show that the effects of abstinence- programs are neither as positive as proponents argue, nor Public policy response to teen pregnancy has been of two as dire as detractors claim. Findings indicate the impor- minds. Some advocate comprehensive sex education with tance of strengthening abstinence attitudes and refusal the intent to equip adolescents with knowledge and tools skills. Furthermore, when adolescents understand how to make appropriate choices concerning sexuality, and to teen pregnancy may affect future goals, the program has avoid or delay pregnancy. Alternatively, some advocate a bigger impact. Knowledge about STDs and pregnancy abstinence education, which promotes abstinence as the increased significantly in the post test for all grade levels. safest route to avoid pregnancy and sexually transmitted References are available upon request. diseases (STDs). The latter effort has been facilitated through federal funding, since 1996. In 2008, 26 partici- ENDNOTE pating states shared $50 million allocated for various abstinence education programs. 1 Prevalence rates are based on the analysis of NHANES data, [available at http://www.cdc.gov/nchs/ Answers regarding the effectiveness of abstinence educa- nhanes.htm] and accessed March 5, 2008. tion on teen pregnancy vary between critics and propo-

For further information contact: Mary Anne Raymond College of Business and Behavioral Science 245 Sirrine Hall Clemson University Clemson, SC 29634–1325 Phone: 864.656.2290 ♦ Fax: 864.656.0138 E-Mail: [email protected]

64 Marketing and Public Policy Conference Proceedings / 2009 MISPERCEPTION AND CONFUSION AS UNINTENDED CONSEQUENCES OF GOVERNMENT-MANDATED WARNING LABELS

Deborah J. MacInnis, University of Southern California, Los Angeles Heather Zhong Wan, University of Southern California, Los Angeles

ABSTRACT categories) and an actual warning label potentially mandated by public policy makers (i.e., a proposed California This study shows that the presence of a warning label Proposition 65 warning label) shows considerable sup- (California Proposition 65 warning label on cola describ- port for this idea. ing potential health risks to women and unborn children from caffeine) on one product category coupled with its LITERATURE REVIEW absence on a similar product category (coffee) creates unintended consequences of consumer misperception and Prior Research on Misperception and Confusion confusion. Misperception and confusion are widely recognized as INTRODUCTION outcomes that are inconsistent with consumer welfare. These outcomes can cause negative marketplace effects “There are many potential objectives associated with that undermine “consumers’ rights to safety and informa- information disclosures . . . but a primary objective is a tion” (Mitchell and Papavassiliou 1999; Mitchell et al. better informed consumer” (Stewart and Martin 2004). 2005). They can also induce consumers to make choices that run contrary to their interests. Once formed, Government-mandated warning labels are a type of infor- misperceptions can be difficult to rectify because they mation disclosure that have as their ultimate goal fostering require that consumers have sufficient motivation, ability, better, more informed consumer decision making. Unfor- and opportunity to process information designed to alter tunately, warnings may create effects that run counter to their current beliefs (Johar and Simmons 2000). From the this goal if they result in consumer misperception or standpoint of consumer affairs and public policy, mini- confusion. The potential for misperception and confusion mizing consumer confusion is important, and a number of as unintended outcomes has been recognized by the legal and regulatory decisions have been made to mitigate California Supreme Court, which decided that “a truthful these outcomes (e.g., Wilkie, McNeill, and Mazis 1984). warning can be misleading or fail to communicate the Despite the importance of misperception and confusion to facts necessary for the protection of users” (Stewart and consumer welfare, previous researchers noted that the Martin 2004, p. 188). A misperception is defined as a sources of misperception and confusion have been under- belief about a state of the world that is not true. Consumer investigated (Mitchell and Papavassiliou 1999). Limited confusion is defined as a lack of clarity or certainty in the research has examined warning labels as sources of state of the world – driven by ambiguity. While several misperception and confusion (France and Bone 2005; types of consumer confusion have been identified by Funkhouser 1984). Finding that warning labels produce previous research (Jocoby and Hoyer 1989; Mitchell and misinformation and confusion would be unfortunate – Papavassiliou 1999; Mitchell, Walsh, and Yamin 2005), particularly if these warning labels were mandated by this study examines confusion from ambiguity. public policy makers charged with consumer protection. This research attempts to fill the void in the literature by The objective of this paper is to develop and test a set of identifying a potential situation wherein warning labels previously untested hypotheses about one potential driver create misperception and confusion. of misperception and confusion from warning labels. We hypothesize that these outcomes can result when a warn- Prior Research on Warning Labels ing label appears on one product category but is not present on a related product category for which a common The majority of extent research on warning labels has and central attribute is shared. Results from an experiment focused on factors that impact warning label “effective- using actual target consumers (i.e., expectant mothers ness”; that is, whether and to what extent warning labels living in California who are consumers of the product are (a) attended to, (b) processed, (c) believed, (d) remem-

Marketing and Public Policy Conference Proceedings / 2009 65 bered, and (e) impact behavior. Research thus far demon- central attribute specified in the label could create con- strates that the effectiveness of warning labels is complex, sumer misperception and confusion. In the case of Propo- as it is driven by potentially myriad interactions between sition 65, a caffeine warning label on cola, coupled with characteristics associated with the label, the consumer, its absence on coffee could create misperception and and the situation in which the consumer finds him or confusion about whether the caffeine in cola is stronger herself (see Argo and Main 2004; Stewart, Folkes, and than, different from, or more of a safety concern to Martin 2001 for reviews). pregnant women than the caffeine in coffee. Our expectations are based on prior research on categorization and A smaller body of research recognizes that warning labels consumer inferences. can create unintended consumer consequences (i.e., consequences that run counter to the warning’s purpose). Categorization and Consumer Inferences. Consumers have Two types of such consequences have been documented mental categories in which they store information about in the literature: (1) reactance (e.g., Bushman and Stack product attributes, usage, usage consequences, etc. (Cohen 1996) and (2) fear and excessive risk perceptions (e.g., and Basu 1986; Viswanathan and Childers 1999). Re- Agrawal 1995). In addition to these two unintended search on consumers’ knowledge structure and content outcomes of warning labels, this research underlines indicates that the more two product categories share misperception and confusion as a third type, though attributes, usage contexts, and usage consequences the understudied, unintended consequence of warning labels. more similar they are perceived to be (MacInnis, Nakamoto, As Mitchell and Papavassiliou (1999) speculate, one and Mani 1992). To illustrate, cola and coffee are similar source of misperception and confusion is ambiguous, as both are members of the super-ordinate category of misleading, or incomplete information. We focus on in- “beverages.” Both are dark in color, can be sold in cups at complete information as a driver of misperception and fast food restaurants, and are known to contain caffeine. confusion from warning labels. The overlap in features associated with these two products Incomplete Information as a Driver of Misperception could lead to one of several outcomes when information and Confusion from Warning Labels (e.g., a warning label) is presented for one product (cola) but not the other (coffee). Consumers might generalize – Government rulings sometimes mandate the presence of assuming that information presented about one product is warning labels on an entire product category, such as also true of a similar product for which the warning is cigarettes, alcohol, or drugs (e.g., aspirin). For example, missing (Andrews, Netemeyer, and Burton 1998). For the agency responsible for implementing California’s example, if cola contains a warning indicating that it Proposition 65 is reviewing scientific information that contains caffeine, consumers may infer that coffee could may lead California to mandate the presence of a caffeine be even more of a safety concern for pregnant women warning for cola. The warning would state that the prod- because it contains three times more caffeine per eight oz. uct has caffeine, a chemical known to the state of Califor- serving as does cola. However, we anticipate that con- nia to cause birth defects or reproductive harm.1 Proposi- sumers infer meaning from the presence of a warning tion 65, also called the Safe Drinking Water and Toxic label on one product coupled with its absence on another. Enforcement Act of 1986 was designed to inform con- Hence, we anticipate a different response: we expect sumers of exposures to chemicals known to cause cancer, consumers will sub-type the information presented about birth defects, or other reproductive harm. However, Propo- one product (cola) based on the absence of comparable sition 65 would create incomplete information if caffeine information on a similar product (coffee) (Burke, Milberg, were regulated because it would not mandate a similar and Moe 1997; Dato-on and Dahlstrom 2003). Consum- warning label on other beverages that contain naturally ers might infer that the warning on one product (cola) does occurring caffeine, such as coffee. The reason why is that not apply to the other product without the warning (cof- Proposition 65 is limited to manufactured chemicals (e.g., fee). Specifically, consumers might infer that the caffeine caffeine) that are added to products – not those that are in cola must be different from, stronger than, or more of a natural to the product itself. The caffeine in coffee is health concern than that in coffee because it carries a inherent in the product as it is a natural ingredient in the warning regarding caffeine content and consumer risks, coffee bean. while coffee does not.

We anticipate that incomplete information caused by the Our predictions that sub-typing instead of generalization presence of a warning label on one product category but will occur are supported by research in a different context. its absence on a related product category that shares a Burke, Milberg, and Moe (1997) found that consumers

66 Marketing and Public Policy Conference Proceedings / 2009 made sub-typing inferences when brands advertised a THE STUDY product feature that was actually true of all brands in the category. For example, consumers inferred that a brand of Method canned whipped cream that claimed to have no CFCs was superior to brands in the product category that did not Sample. Hypothesis 1 and 2 were tested in the context of make this claim. In reality, canned whipped creams do not the aforementioned Proposition 65 measure in which a contain CFCs. Hence there was nothing special about the caffeine warning label appears on cola but not on coffee. brand that claimed to be CFC–free. However, consumers Since that warning label describes birth defects or other seemed to infer meaning from the presence of this product reproductive harm from caffeine content, the primary claim on one brand coupled with its absence on other target market for the warning is expectant mothers. Since brands, and inferred that the brand that claimed to have no Proposition 65 is a California law, the study was limited CFCs was different. to respondents living in California.

In the case of the aforementioned Proposition 65 warning Three hundred nine expectant mothers living in California label, sub-typing inferences would lead to conclusions were asked to participate in a study on “beverage con- that are untrue, and hence cause consumer misperception. sumption.” The sample was comprised of respondents Coffee has three times the caffeine content as an equiva- who lived within a 20-mile radius of one of 10 shopping lent amount of cola (not less). One milligram of caffeine malls in or around the San Diego, Los Angeles, and San in cola is just as strong as one milligram of caffeine in Francisco metropolitan areas. Shopping malls were se- coffee (it is not stronger), and the effect of natural and lected so as to maximize the possibility of sampling chemically added caffeine on the body is the same (not respondents who are representative of the educational, different). Both products are generally considered to be income, and ethnic diversity of women of child bearing safe for pregnant women to consume as long as they are age who live in the state. Respondents were prescreened consumed in moderation (e.g., 200–300 mgs; the equiva- to ensure that they met the appropriate criteria for the lent of one to two cups of coffee per day) (OTIS study; that is they were (a) expecting a baby2 and (b) had www.otispregnancy.org). Thus, we predict that: consumed both cola and coffee products within the past Hypothesis 1: The presence of a warning label re- two years. This prescreening ensured that the sample garding a particular attribute on one product category contained only target relevant consumers (e.g., those who (e.g., a warning about caffeine in cola) coupled with would consider buying brands within the cola and coffee the warning’s absence on a product category that product categories). shares that attribute (e.g., coffee) will create consumer misperception that the warning does not apply Respondents were recruited by one of two methods. A to the product without the warning. probability sampling model was used to recruit respondents via phone (N = 95). A convenience sampling model Information presented about one product that is absent for was used to recruit pregnant respondents who were al- a similarly perceived product may also cause consumer ready shopping at the mall (N = 214). The sample ranged confusion. In the case of the Proposition 65 warning label, in age from 18–45, with a mean age of 26.73. Nearly half the warning may foster a lack of clarity about the true state (44%) were pregnant for the first time. The vast majority of the world. In addition, consumers may be confused (92.6%) were born in the U.S. Those who were non-U.S. about whether the caffeine in cola is different from, born had lived in the U.S. for a relatively long period of greater than, or more of a health concern than the caffeine time (M = 20.68 years). The majority of respondents had in coffee. Thus, we anticipate that: health insurance (84.8%).

Hypothesis 2: The presence of a warning label re- Experimental Design. Respondents were randomly as- garding a particular attribute on one product category signed to one of four conditions representing a 2 x 2 (e.g., a warning about caffeine in cola) coupled with between-subjects experimental design. The independent the warning’s absence on a product category that variables included (a) warning label condition (warning shares that attribute (e.g., coffee) will create con- label present vs. absent on cola) and stimulus order sumer confusion about the relationship of the warn- (information about cola presented first, information about ing to the product without the warning. coffee presented second and vice versa).

Marketing and Public Policy Conference Proceedings / 2009 67 The warning label condition manipulated whether re- their ability. The questionnaire took approximately 10 spondents were or were not exposed to a Proposition 65 minutes to complete. After completing the questionnaire, warning label on the back of a package of cola. The cola respondents were thanked and paid for their participation. warning label reads: WARNING: This product contains They were given a set of debriefing instructions and a caffeine, a chemical known to the State of California to handout from the March of Dimes that summarized the cause birth defects or other reproductive harm. The caffeine content of various products and recommenda- wording, font, size, and warning label placement mirror tions for caffeine consumption for pregnant women. These the label that consumers would be likely exposed to if a documents were designed to clarify any misperceptions or warning label for caffeine were to be required. Figure 1 confusion that consumers might have had. illustrates the warning label present condition. This condition mirrors the information consumers would be ex- Dependent Variables. Misperception and confusion were posed to should a warning label on cola be mandated. measured using both closed and open-ended questions. Those in the warning label absent information were not Specifically, misperception was assessed by several closed- exposed to the warning label. The no warning condition ended questions. A set of closed-ended questions mea- represents a control condition, against which the effects of sured respondents’ beliefs about the caffeine in cola vs. a warning can be compared. The warning label absent coffee. They were asked the extent to which they dis- condition was identical to Figure 1 except the caffeine agreed or agreed that the caffeine in cola is (a) different warning label on cola was removed. This condition mir- from the caffeine in coffee, (b) stronger than the caffeine rors reality as it portrays the information consumers are in coffee, or (c) more of a health concern than the caffeine currently exposed to in the marketplace. There was no in coffee (1 = strongly disagree; 7 = strongly agree). Since warning label on coffee in either the experimental or the these statements are not true, the more consumers agreed control condition. The product information shown in with these items, the greater their level of misperception. Figure 1 represents the product marketplace by utilizing Another question tapped respondents’ beliefs about the package information that commonly appears on soft- relative safety of cola compared with coffee. Respondents drink and coffee beverages. were asked to rate whether cola was safer for expectant mothers than, as safe as, or less safe than an equivalent The stimulus order condition manipulated whether the amount of coffee (1 = coffee is much safer than cola; 7 = cola package was presented on the left hand side of the cola is much safer than coffee). Since cola contains less Exhibit with the coffee package on the right-hand side (as caffeine than an equivalent amount of coffee and hence is in Figure 1) or vice versa. There were no order effects; safer than coffee, the lower relative safety of cola to coffee hence order is not discussed further. The results are consumers rate, the greater their level of misperception. aggregated across order conditions, yielding a final sample of 154 respondents in the experimental condition (warn- Two additional open-ended questions measured both ing label present) and 155 respondents in the control misperception and confusion: the first question asked condition (warning label absent). respondents to explain which product is safer – cola or an equivalent amount of coffee; and the second question Procedure. Respondents from both sampling frames were asked respondents why there was a caffeine warning label greeted by a professional interviewer at the designated on the cola product but not coffee product. Responses to shopping mall and led to a quiet location where they could this question were content analyzed to assess the exist- observe an Exhibit and complete a questionnaire. Inter- ence of misperception and confusion. Open-ended ques- viewers (and respondents) were blind to the purpose of the tions were coded by trained coders who were blind to the study and were unaware of any hypotheses to be tested. respondents’ condition or the experimental hypotheses. Respondents were told to imagine that they were shop- Inter-coder agreement for both open-ended questions was ping in a grocery store and that they had picked up and above .90. Disagreements between coders were resolved read the label of the product shown on the left hand side through discussion. of the Exhibit (see Figure 1 for an example). They were instructed to read all of the information provided about Control Variables. The questionnaire assessed a set of that product. When they were finished, they were asked to control variables. Respondents estimated their daily con- then imagine that they went to another section of the store sumption of cola and coffee (number of glasses/cups per and picked up and read the label of the product shown in day) before they were pregnant. These data were collected the right-hand side of their Exhibit. After finishing read- to determine whether misperception/confusion would be ing the information, respondents completed a self-admin- impacted by prior positivity toward these products.3 Re- istered questionnaire at their own pace and to the best of spondents also indicated how many cups of coffee and

68 Marketing and Public Policy Conference Proceedings / 2009 FIGUREFigure 1 1: ExperimentalExperimental Condition Condition Stimuli Stimuli

Panel A: Panel B: Back package of a cola product Back package of a coffee product

VERY LOW SODIUM, 35 mg OR LESS PER 240 mL (8 fl oz) %U HZWKH%HVW(YHU\ 7 LPH Nutrition Facts Standard Serving This Package USE FRESH BEANS. Serving Size 8 fl oz (240 mL) 20 fl oz (591 mL Servings Per Container 2.5 1 For the freshest cup, buy coffee frequently, weekly if possible. Amount Per Serving %DV* %DV* Store in an airtight container Calories 0 0 USE GOOD WATER Total Fat 0g 0g 0% 0g 0% Bottled or filtered water is recommended. Avoid using water from the hot Sodium 30mg 1% 80mg 3% tap or water that has already boiled.

Total Carbohydrate 0g 0% 0g 0% USE THE CORRECT GRIND Each brewing method requires its own Sugars 0g 0% 0g 0% grind- medium for automatic drip, coarse Protein 0g 0% 0g 0% for French press, fine for espresso

MEASURE CAREFULLY * Percent Daily Values (DV) are based on a 2,000 calorie diet We recommend using two level CARBONATED WATER, CARMEL COLOR, ASPARTAME, PHOSPORIC tablespoons (10g) of ground coffee for ACID, POTASSIUM BENZOATE (TO PROTECT TASTE), NATURAL each six ounces (180 ml) of water. FLAVORINGS, CITRIC ACID, C AFFEINE.

PHENYLKETONURICS: CONTAINS PHENYLALANINE CAFFEINE C ONTENT: 34mg/8oz; 86mg/20 fl oz

WARNING: THIS PRODUCT CONTAINS CAFFEINE, A PRODUCT KNOWN TO THE STATE OF CALIFORNIA TO CAUSE BIRTH DEFECTS OR OTHER REPRODUCTIVE HARM

cola they believed were safe for pregnant women to Results consume on a daily basis. This item was designed to determine whether misperception and confusion were Respondents reported being more health conscious now impacted by beliefs about caffeine moderation during that they are pregnant; they reported being more con- pregnancy. Respondents were also asked whether they are cerned about their diets (M = 6.04), are more likely to concerned about their diet now that they are pregnant and search for information on food labels (M = 5.74), and are whether they search for information on product labels and more likely to compare product labels (M = 5.78; 1 = compare product labels (1 = strongly disagree; 7 = strongly strongly disagree, 7 = strongly agree). This data is consis- agree). This information was collected to determine that tent with the assumption that respondents might read the information in Figure 1 would be representative of the information like that found on the packages of products marketplace context in which pregnant women would and compare information from one product (e.g., cola) to find themselves. Finally, respondents were queried on that of another (e.g., coffee). basic demographic variables including education, age, ethnicity, income, number of years in the U.S., health Test of the Hypotheses. The results revealed the follow- insurance status, whether under a doctor’s care for their ing evidence of misperception. Respondents who were current pregnancy, and whether this was their first baby. exposed to the Proposition 65 warning label on cola but Based on a set of t-tests, none of these demographic not on coffee (experimental respondents) were signifi- variables were significant factors of misperception and cantly more likely than those who were not exposed to the confusion and are not discussed further. label on cola (control group respondents) to believe that

Marketing and Public Policy Conference Proceedings / 2009 69 the caffeine in cola is (1) different from the caffeine in respondents in the experimental (versus the control) group coffee (Mexperimental = 4.06 vs. Mcontrol = 3.49; p < .008), (2) were confused about which product is safer ( 2 = 3.90, p = stronger than that in coffee (Mexperimental = 3.75 vs. Mcontrol = .039). 3.25; p < .01), and (3) more of a safety concern given its combination with other cola ingredients (Mexperimental = 4.37 DISCUSSION vs. Mcontrol = 3.80; p < .003). These statements are not true; hence, these results indicate misperception. These results Although limited research has studied when and whether were not contingent on any other factor and support warning labels might induce consumer misperception and hypothesis 1. confusion, these results suggest that such outcomes may indeed be unintended consequences of warning labels. There were no significant main effects for respondents’ We hypothesized and found that misperception and con- perceptions of which is safer-cola or an equivalent amount fusion are likely when a warning label is present on one of coffee. However, a significant interaction effect emerged product category but is absent on a different but related between the experimental condition and respondents’ product category that shares a core common feature understanding of caffeine moderation during pregnancy highlighted by the warning label. Consumers infer mean- (F = 4.98, p < .03). The latter independent variable was ing from the missing information, and infer that the created by a median split on the number of cups of coffee product with the warning label is somehow different from and cola that are recommended for pregnant women. the one without the warning label. Clearly these results Respondents who believed that caffeine consumption should be of concern to public policy makers who are during pregnancy should be limited were significantly charged with mandating warning labels. While govern- more likely to believe that cola is less safe than an ment mandated warning labels (like Proposition 65) are equivalent amount of coffee for pregnant women when designed to inform consumers, to the extent that warning they were (vs. were not) exposed to the Proposition 65 labels create misperception and confusion, consumers’ warning label (Mexperimental = 4.41 vs. Mcontrol = 4.96, p < .05). interests are harmed, not helped. There were no significant differences observed for consumers who did not believe that caffeine consumption Limitations and Future Research should be limited during pregnancy. Since cola is safer than an equivalent amount of coffee, this result reveals The conclusions of this study should be couched within misperception and further supports hypothesis 1. the context of the study’s limitations. First, since we attempted to investigate a real-world government man- Content analyses of the two open-ended questions showed dated warning label (Proposition 65), our hypothesis evidence of both misperception and confusion. The most testing was confined to pertinent product categories (cola common responses to the open-ended question of why and coffee) and sample (expectant mothers living in there was a caffeine warning label on cola but not on California who are consumers of the product categories). coffee represented confusion and misperception. The The generalizability of our research findings need to be prevalence of confusion was manifested by the fact that further tested in other product categories and consumer nearly one third of respondents (32%) were confused by population. Notably though, our results were not contin- the presence of the warning label on cola coupled with its gent on any demographic factors. Second, future research absence on coffee. This result supports hypothesis 2. should investigate not just the presence vs. absence of a Evidence of misperception was identified in two types of warning label as a source of misperception and confusion, responses to the question: 19 percent of respondents but also its content (Rotfeld 2001). For example, the concluded that cola has more caffeine than coffee and an California Proposition 65 warning label would likely additional 15 percent falsely interpreted the label’s pres- utilize the statement shown in Figure 1 (Warning: This ence to mean that the ingredients of cola are less safe for product contains caffeine – a chemical known to the state pregnant consumers than are the ingredients of coffee. of California to cause birth defects or other reproductive Both perceptions are incorrect. Only one percent attrib- harm). This wording itself is ambiguous, as it can lead to uted the presence of a caffeine warning label to the true conclusions that are not supported by facts. That is, the reason – the fact that it may be required by law. These label’s indication that the caffeine it contains can cause results of the open-ended question provide additional birth defects or other reproductive harm is not specific to support for hypothesis 1. Results of the other open-ended consumption during pregnancy. Pregnant consumers might question about which product is safer—cola or an equiva- incorrectly infer that caffeine they consumed before they lent amount of coffee further confirm hypothesis 2. A 2 were pregnant has put them or their unborn child at risk. analysis revealed that a significantly greater percentage of Furthermore, the label is ambiguous as it does not specify

70 Marketing and Public Policy Conference Proceedings / 2009 reproductive harm to whom. It is possible that men who more psychological reactance than did the information read the warning label would infer that the caffeine label. It is possible that an information label can also contained in cola could put them at risk for reproductive produce less misperception and confusion than a warning harm. Such conclusions would be unfounded. label. For example, most cola beverage companies mention on their label the amount of caffeine contained in the Finally, future research should investigate misperception cola (see Figure 1). The mere mention of the presence of and confusion from warning labels on the one hand vs. caffeine in the product might be sufficient to inform information labels on the other. Bushman (1998) identi- consumers of the caffeine content in cola without creating fied two types of labels – a warning label (e.g., “The U.S. the misperceptions and confusion spurred by a warning Surgeon General has determined that eating high fat food label. Hence, additional research on the relative benefits increases your risk of heart disease”) and an information of warning vs. information labels and the conditions label (e.g., “In this product, 90% of the calories come from under which each is most effective warrants additional fat”). Bushman found that the warning label produced research.

ENDNOTES cola, but not other beverages that contain naturally occurring caffeine such as coffee. 1 Whether and to what extent caffeine consumption is 2 Respondents’ pregnancy status was one of a set of associated with birth defects or other reproductive questions that asked about recent major life events. harm is still a matter of scientific debate. If California Other questions included moving, getting married, determines that scientific evidence supports this claim, changing jobs, buying a home, and buying a car. the Proposition 65 warning label will be mandated for 3 They were not.

REFERENCES Fat, and No-Fat Products,” Journal of Applied Psy- chology, 83 (February), 97–101. Agrawal, Madhu (1995), “Warning Labels: The Role of Cohen, Joel B. and Kunal Basu (1986), “Alternative Expertise and Perceived Risk in Pharmaceutical Pur- Models of Categorization: Toward a Contingent Pro- chase Behavior,” Health Marketing Quarterly, 14 cessing Framework,” Journal of Consumer Research, (2), 99–115. 13 (March), 455–72. Andrews, J. Craig, Richard G. Netemeyer, and Scot Dato-On, Mary Conway and Robert Dahlstrom (2003), Burton (1998), “Consumer Generalization of Nutri- “A Meta-Analytic Investigation of Contrast Effects ent Content Claims in Advertising,” Journal of Mar- in Decision Making,” Psychology & Marketing, 20 keting, 62 (October), 62–75. (August), 707–30. Argo, Jennifer and Kelley J. Main (2004), “A Meta- France, Karen R. and Paula F. Bone (2005), “Policy Analysis of the Effectiveness of Warning Labels,” Makers’ Paradigms and Evidence from Consumer Journal of Public Policy & Marketing, 23 (Fall), Interpretations of Dietary Supplement Labels,” The 193–208. Journal of Consumer Affairs, 39 (1), 27–51. Burke, Sandra, Sandra Milberg, and Wendy Moe (1997), Funkhouser, G. Ray (1984), “An Empirical Study of “Displaying Common but Previously Neglected Consumers’ Sensitivity to the Wording of Affirma- Health Claims on Product Labels: Understanding tive Disclosure Messages,” Journal of Public Policy Competitive Advantages, Deception and Education,” & Marketing, 3, 26–37. Journal of Public Policy & Marketing, 16 (2), 242– Johar, Gita and Carolyn J. Simmons (2000), “Use of 55. Concurrent Disclosures to Correct Invalid Inferences,” Bushman, Brad and Angela Stack (1996), “Forbidden Journal of Consumer Research, 26 (March), 307–22. Fruit versus Tainted Fruit: Effect of Warning Labels MacInnis, Deborah J., Kent Nakamoto, and Gayathri on Attraction to Television Violence,” Journal of Mani (1992), “Cognitive Associations and Product Experimental Psychology: Applied, 2, 207–26. Category Comparisons: The Role of Knowledge ______(1998), “Effects of Warning and Informa- Structure and Context,” Advances in Consumer Re- tion Labels on Consumption of Full-Fat, Reduced- search, John Sherry and Brian Sternthal, eds. Provo,

Marketing and Public Policy Conference Proceedings / 2009 71 UT: Association for Consumer Research, 19, 260– gory T. Gundlach, eds. Thousand Oaks, CA: Sage, 67. 335–71. Mitchell, Vincent-Wayne and Vassilios Papavassiliou ______and Ingrid M. Martin (2004), “Advertis- (1999), “Marketing Causes and Implications of Con- ing Disclosures: Clear and Conspicuous or Under- sumer Confusion,” Journal of Product and Brand stood and Used?” Journal of Public Policy & Market- Management, 8 (4), 319–39. ing, 23 (2), 183–92. ______, Gianfranco Walsh, and Mo Yamin (2005), Viswanathan, Madhubalan and Terry L. Childers (1999), “Towards a Conceptual Model of Consumer Confu- “Understanding How Product Attributes Influence sion,” in Advances in Consumer Research, Geeta Categorization: Development and Validation of Fuzzy Menon and Akshay Rao, eds. Duluth, MN: Associa- Set-Based Measures of Gradedness in Product Cat- tion for Consumer Research, 32, 143–50. egories,” Journal of Marketing Research, 36 (Febru- Rotfeld, Herbert Jack (2001), Adventures in Misplaced ary), 75–94. Marketing. Westport, CT: Quorum Books. Wilkie, William L., Dennis L. McNeill, and Michael B. Stewart, David W., Valerie S. Folkes, and Ingrid Martin Mazis (1984), “Marketing’s ‘Scarlet Letter’: The (2001), “Consumer Response to Warnings and Other Theory and Practice of Corrective Advertising,” Jour- Types of Product Hazard Information,” in Handbook nal of Marketing, 48 (2), 11–31. of Marketing and Society, Paul N. Bloom and Gre-

For further information contact: Heather Zhong Wan Marketing Department ACC 306 Marshall School of Business University of Southern California Los Angeles, CA 90089–0443 Phone: 213.740.0285 Fax: 213.740.7828 E-Mail: [email protected]

72 Marketing and Public Policy Conference Proceedings / 2009 CAN CONSUMERS CORRECT FOR SUBTLE PERSUASION? BRAND PLACEMENT EFFECTS AND THE IMPORTANT ROLES OF DISCLOSURE INFORMATION AND TIMING

Margaret C. Campbell, University of Colorado, Boulder Gina S. Mohr, University of Colorado, Boulder Peeter W.J. Verlegh, Erasmus University, The Netherlands

SUMMARY positive attitudes toward the brand (Biehal and Chakravarti 1986; Shapiro 1999). Although brand placement is one of the fastest growing forms of marketing communication (PQ Media 2006), Effects of Placement Disclosure Information there is little consensus about whether or how placement persuades consumers (Russell and Belch 2005). Further- Placement disclosure has been proposed as a method to more, based on the idea that consumers should be pro- allow consumers to “protect” themselves against un- tected from surreptitious persuasion, there have been calls wanted persuasion but little research explores the effects to use some form of disclosure to “make sure consumers of disclosure. We theorize that when disclosure informa- understand and are aware that they are being advertised tion increases the accessibility of consumers’ persuasion to” (Kevin J. Martin, 2008 FCC Chairman). However, knowledge (e.g., Campbell and Kirmani 2000), consum- there has been little work empirically testing the effective- ers “correct” their responses to the brand. In other words, ness of disclosures. Our research examines the extent to information that increases accessibility of persuasive in- which subtle brand placement persuades consumers and tent can lead consumers to modify their responses in order investigates the effects of placement disclosure. We pro- to counter perceived bias. The Flexible Correction Model vide a theoretical analysis of the process under which (FCM) provides some support for the idea that disclosure brand placement and disclosure may affect consumer of brand placement may affect consumers’ responses response and present two experiments that support this (Wegener and Petty 1995, 1997). Specifically, research in process. this stream shows that information to respond in an “unbiased” fashion leads people to modify their responses Understanding Placement Effects to correct for perceived influence.

Drawing from the literatures on persuasion, persuasion The timing of disclosure information, however, may in- knowledge, and flexible correction, we propose that plac- fluence the correction process. While the FCM does not ing a brand in a TV program can increase consumers’ discuss timing, all correction studies that we have found recall of and attitude toward the brand, but that disclosure provide bias information after exposure to the influencing information can reduce these effects. Specifically, expo- stimulus. It is possible that consumers may process the sure to a subtle (e.g., visual only, low plot connection) brand placement differently when they are made aware of brand placement can increase the accessibility of the its persuasive nature before, versus after, the actual brand brand in memory (e.g., Nedungadi 1990; O’Guinn and placement. Since timing issues are under discussion for Shrum 1997) and thereby lead to increased brand salience regulation, we also explore this possibility. and higher “top-of-mind-awareness” (TOMA: recall cued by product category; Keller 1993; Krishnan and Shapiro Research Overview 1996; Nedungadi 1990). Similarly, subtle brand placement can lead to more positive attitude toward the brand. We conducted two studies to gain insight into consumers’ When a brand is subtly included in a scene in a television theories of the impact of placement, followed by two show, consumers are likely to see the brand but are not experiments that demonstrate that subtle brand placement likely to engage in elaborate brand-related processing. increases brand salience (i.e., TOMA) and attitude toward This can be regarded as a peripheral process leading to the brand, but that bias disclosure information reduces higher brand attitudes (e.g., Petty, Cacioppo, and these persuasive effects. In both studies, participants were Schumann 1983). Likewise, subtle placements can in- recruited to participate in a several short studies, including crease feelings of familiarity, which can lead to more one on a TV program. Study participants first watched a

Marketing and Public Policy Conference Proceedings / 2009 73 TV show in which a professional graphic designer digi- (17% brand TOMAcontrolvs. 75% in TOMAplacement; Chi- tally manipulated a cereal box to be either a fictitious square (1) = 18.67, p < .001) as well as on Ab (3.9 vs. 4.9, brand or a real brand of cereal. After answering questions t(124) = 2.12, p < .03). Also, both disclosure at the beginning about the show, they turned to a second “Product study” and the end of the show lead to significantly lower that included measures of TOMA and attitude (Ab). reported TOMA (36% and 37%) compared to placement without disclosure (p’s < .001). As predicted by the Experiment 1. This experiment manipulated brand place- proposed correction process, while disclosure at the end ment (fictitious control vs. placed brand) and instructions similarly reduced the impact of placement on Ab (4.1, to avoid influence by unrelated experiences (present or t(124 = 2, p < .05), disclosure at the beginning did not have absent) in a 2 X 2 design. Results show that brand this effect (4.9, t(124) < 1). placement increased the TOMA of the placed brand (27%) relative to the fictitious brand condition (9%; Chi- Discussion square (3) = 9.9, p < .02). Further, instructions to avoid influence resulted in lower TOMA (7%) compared to the The present research has important practical implications. placement without such instructions (Chi-square (1) = Our results provide strong evidence that subtle brand 5.6, p < .02). A significant interaction showed that brand placement can be persuasive, increasing both brand sa- placement lead to higher Ab (5.5) compared to the ficti- lience, as indicated by top-of-mind-awareness, and atti- tious condition (4.9; F(1, 131) = 3.9, p < .05), and that tude toward the brand. This research also contributes to influence avoidance instructions eliminated the effect the current dialog on placement regulation. In particular,

(4.6; F(1, 131) = 9.92, p < .01). our research lends support to some regulators’ idea that clear disclosure, not embedded in closing credits, but Experiment 2. This experiment manipulated brand place- provided after the brand placement, can be perceived as ment (fictitious control vs. placed brand) and placement useful information for consumers. References are avail- disclosure (none, at the beginning, at the end). This study able upon request. replicated the positive effects of placement on TOMA

For further information contact: Margaret C. Campbell Marketing Leeds School of Business University of Colorado at Boulder 419 UCB Boulder, CO 80309–0419 Phone: 303.735.6305 E-Mail: [email protected]

74 Marketing and Public Policy Conference Proceedings / 2009 VICARIOUS GOAL FULFILLMENT: WHEN THE MERE PRESENCE OF A HEALTHY OPTION LEADS TO AN IRONICALLY INDULGENT DECISION

Keith Wilcox, Baruch College/CUNY, New York Beth Vallen, Loyola College, Baltimore Lauren G. Block, Baruch College/CUNY, New York Gavan J. Fitzsimons, Duke University, Durham

SUMMARY Consistent with Polivy and Herman’s (1985) what the hell effect, which suggests that once individuals decide upon Healthy food options are appearing where many consum- a course of behavior that is inconsistent with a goal (e.g., ers least expect them – restaurants, event venues, retailers, abandoned their diet to eat a cookie) they are likely to and other distributors have added healthier options to continue to indulge to an even greater extent (e.g., eat the supplement their typically unhealthy food offerings. While whole bag of cookies), we suggest that this licensing the proliferation of healthy food alternatives has been effect does not merely result in the selection of a less beneficial for individuals who tend to make healthier food healthy option for those individuals who make their choices when given the opportunity to do so, its effect is decision from the choice set that includes the healthy far from ubiquitous. As a recent article on fast-food chains option, but rather the least healthy option available. Thus, notes, “Fast-food customers had indeed been clamoring we argue that the mere presence of a healthy option in a for healthy alternatives, which prompted an industry wide choice set of relatively unhealthy foods should result in stampede toward salads and orange slices, but just be- less active health goals and, in turn, greater preference for cause customers wanted them on the menu didn’t neces- the most indulgent item compared to instances in which a sarily mean they wanted to eat them” (Keohane 2008, healthy item is not available. p. 95). With the increased availability of nutritious menu options, why have more consumers not swapped their Interestingly, the goal activation processes that we pro- french fries for a side salad? In this research, we present pose as the driver of this behavior suggests the potential evidence that for many consumers, the addition of healthy for an ironic effect of sorts at the individual level; namely alternatives to choice sets can, ironically, increase the that the effect will be accentuated for individuals who are consumption of very indulgent food items. higher in self-control, compared to those who are lower in self-control. Previous research has shown that individuals Recent research suggests that individuals license them- high in self-control have more accessible cognitions asso- selves to indulge in temptations when they have previ- ciated with the achievement of long-term goals compared ously acted in line with a longer-term goal. When indi- to those low in self-control, thus demonstrating a greater viduals focus on their progress toward a focal goal, it focus on achieving important long-term objectives (Giner- allows them to temporarily disengage from that goal to Sorolla 2001). In addition, high self-control individuals pursue tempting alternatives (Fishbach and Dhar 2005; are also likely to rely more heavily on cues that justify Fishbach and Zhang 2008). Although the most common indulgent choices (Kivetz and Zheng 2006). way in which individuals perceive goal progress is by actively pursuing their goal, sometimes simply intending Across four studies in varying food consumption do- to pursue a goal makes it more likely that individuals will mains, we examine how consumers’ food choices differ indulge (Fishbach and Dhar 2005). We extend this rea- when a healthy item is included in a choice set compared soning to suggest that when individuals have the opportu- to when it is not available. In study 1 we presented nity to engage in a course of action that is consistent with respondents with side dish menus consisting of either healthy eating goals, as they do when a healthy item is relatively unhealthy food items (i.e., french fries, chicken included in a food choice set, merely considering this nuggets and baked potato) or the same items in addition to option will satisfy their healthy eating goal – at least a relatively healthy item (i.e., salad). Self-control was temporarily – and, in turn, license them to indulge. measured. We found that, ironically, when a healthy

Marketing and Public Policy Conference Proceedings / 2009 75 alternative was added to a menu, higher levels of self- different levels of healthfulness as less similar to one control increased the likelihood of making a very indul- another, while less accessible health goals result in greater gent decision (i.e., choosing the fries). Interestingly, chang- similarity ratings. Using the menu stimuli used in the prior ing menu offerings altered the choice patterns of those study, we show in study 4 that the presence (vs. absence) who are perhaps least likely to make unhealthy decisions of the healthy item increases the perceived similarity of when not exposed to such contextual cues. the items for individuals with high self-control. In other words, evidence of goal fulfillment was obtained by Study 2 replicated the findings of study 1 in two different showing that high self-control individuals rate the items as food-related contexts, specifically the selection of an (1) less similar in the absence of a healthy item (because entrée and the choice of a within-category packaged snack health goals are more accessible), and (2) more similar in food. the presence of a healthy item (because health goals have been fulfilled). Importantly, we show that once healthy Our third study provided direct evidence of goal activa- eating goals are fulfilled and perceived similarity among tion/fulfillment as the underlying process. Using menu items in the choice set is high, high self-control individu- stimuli identical to those used in study 1, we found that als pay more attention to the most indulgent option in the when the choice set did not include a healthy option, choice set. This provides support for why the most indul- higher levels of self-control corresponded to faster re- gent option, rather than any indulgent option, is chosen. sponse times to health-related words, indicating greater activation of health-related goals relative lower levels of The objective of this research was to understand how self-control. When the choice set included a healthy adding healthy items to food choice sets, a practice grow- option, response times to health-related words were slower ing in popularity among food retailers, impacts consumer for high self-control individuals, demonstrating less ac- choice. As such, this research contributes to research on cessibility compared to when the healthy option was not context dependent decision making that explores cues included. In other words, while high self-control indi- that lead to indulgence by showing that the mere presence viduals are better equipped to activate self-control in of goal-consistent options leads some to indulge by vicari- response to tempting stimuli, they are also highly suscep- ously fulfilling health-related goals and drawing attention tible to cues that reduce the threat imposed by tempting to highly tempting options. In the eating domain, as well stimuli. as in other arenas, inhibiting temptations carries important consequences for consumers as well as society as a whole. Our final study provided additional support for our pro- Thus, an understanding of goal fulfillment processes is of posed goal activation process using a categorization ap- substantial importance for understanding consumer be- proach to demonstrate vicarious goal fulfillment. Prior havior at the individual level, as well as broader issues like research (Ratneshwar et al. 2001) shows that accessible the U.S. obesity epidemic. References are available upon health goals lead individuals to rate food items with request.

For further information contact: Beth Vallen Loyola College 4501 N. Charles Street Baltimore, MD 21224 Phone: 410.617.5466 Fax: 410.617.2117 E-Mail: [email protected]

76 Marketing and Public Policy Conference Proceedings / 2009 SPECIAL SESSION

THE ROLE OF QUALITY INFORMATION IN HEALTHCARE MARKETS: COMPETITION AND CONSUMER PROTECTION ISSUES

Session Chairs: Debra Scammon, University of Utah, Salt Lake City Jack Calfee, American Enterprise Institute, Washington, D.C.

Panelists: Patricia Schultheiss, Federal Trade Commission, Washington, D.C. Kathlene Kendrick, Agency for Healthcare Research and Quality, Washington, D.C. Tom Miller, American Enterprise Institute, Washington, D.C.

SESSION OVERVIEW lines such as those for patients admitted with cardiac problems or treated for pneumonia. Similarly, “report The United States health care system is widely perceived cards” on insurance plans that provide a standard set of as being in a state of crisis. In order to deal with rising National Committee for Quality Assurance (NCQA) data costs, consumers are being encouraged to take a more have been available for some time. More recently infor- active role in decisions about all aspects of their healthcare. mation about the performance of providers has become Providers, increasingly disenchanted due to time and more available. financial pressures, are facing increased pressures for accountability. An HHS initiative began in 2006 established Chartered Value Exchanges (CVEs) in various states (currently The concept of “consumer-directed health care” (CDHC) 20+) with the objective of making value information has received much attention over the last several years as about local health care providers more transparent to both a means for making consumers more sensitive to both employers and consumers. This initiative recognizes that costs and quality of their care (Herzlinger 2002, 2004; healthcare is essentially a “local” industry and that infor- Porter and Tiesberg 2006; Berenson and Cassel 2009). mation about the performance of local providers may be The concept of CDHC is based on the notion that if the most helpful in decision-making. The initiative is consumers are more involved in all aspects of decisions intended to stimulate the availability of quality and cost about their healthcare, consumers and the healthcare information for consumers while at the same time en- system generally will benefit. In order to make decisions abling more informed referral decisions and stimulating wisely, consumers need information about both the cost performance improvement by physicians. Insurance plans and the quality of healthcare services. Providing the could make decisions about plan panel composition based information needed to support choice-making presents upon provider performance. many challenges. CVEs are working to design formats for the dissemination There are a number of initiatives designed to increase the of provider quality information. Before such information availability of quality information in the marketplace to is made available to the public issues related to data spur competition among healthcare providers based on quality (aggregation and attribution), trustworthiness, and quality (Bloche 2006). Hospital quality data are increas- accuracy must be addressed. ingly available on web sites such as HospitalCompare.Com. Hospitals that accept Medicaid and Medicare patients are Research on the impact of quality information on consum- required to report their experience on a number of perfor- ers’ decisions suggests there are likely to be a number of mance measures. Currently these measures focus on the challenges in encouraging consumers to consider quality extent to which hospitals adhere to best-practice guide- information. A recent Kaiser Family Foundation report

Marketing and Public Policy Conference Proceedings / 2009 77 Special Session: The Role of Quality Information In Healthcare Markets: Competition and Consumer Protection Issues indicates that less than 30 percent of Americans say they needs of consumers, physicians, employers and insurers have seen information comparing the quality of different for quality information, barriers to quality-based compe- health insurance plans, hospitals, or doctors during the tition, and policies to facilitate the measurement, collec- last year, despite the wide availability of such information tion, and reporting of quality information. and even fewer report having seen and used such information (Kaiser Family Foundation 2008). The purpose of this session is to explore the implications of some of these emerging issues in the reporting of health Following a 2004 Health Care Report (FTC) suggesting care quality information. Panelists from the FTC, the that price and quality are difficult to measure and that Agency for Healthcare Research on Quality, and the information is not necessarily reliable, the FTC held a American Enterprise Institute will explore the potential workshop seeking input on issues related to quality infor- impacts of quality information on competition, the mation including the extent to which there is demand for challenges of communicating meaningful information on high quality healthcare, the attributes of quality informa- quality, issues of accurately assessing quality, and policies tion that motivate purchasers to select high quality provid- that can facilitate the collection and reporting of quality ers, and the ramifications of quality-based competition on information. the availability of health care. The workshop explored the

For further information contact: Debra L. Scammon University of Utah 1645 Campus Central Drive Salt Lake City, UT 84112 Phone: 801.581.4754 Fax: 801.581.3152 E-Mail: [email protected]

78 Marketing and Public Policy Conference Proceedings / 2009 SPECIAL SESSION

DEMOCRACY, CIVIC PARTICIPATION, AND CONSUMPTION

Session Chair: Canan Corus, St. John’s University, Queens

Discussant: Julie L. Ozanne, Virginia Tech, Blacksburg

SPECIAL SESSION OVERVIEW 2002). A consumer- citizen acts beyond their own immediate, personal interests and considers the impact of their Improving consumer well-being has long been the focus choices on wider society (McGregor 2001). Thus, con- of consumer researchers as well as policy makers. An sumption could be a means for the reflexive individual to integral issue is fostering consumers’ voice and agency in relate to the larger society by responsibly contributing to influencing decisions and strategies that would shape it. In other words, alternative forms of citizen involvement their lives. Our session aims to contribute to the pursuit of for deepening democracy can be formulated through enabling consumers to become integral part of public consumption processes based on consumers’ rights and decision making and action. To this end, we focus on responsibilities. In our session we aim to start conversa- methods and concepts related to citizen engagement, tions on the interaction of consumerism and citizenship; highlighting the effects of media and marketing on civic the Dermody, Hanmer-Lloyd, and Scullion paper tackles participation and citizenship. these issues.

Our inspiration stems from the increasingly evident de- Taking individual responsibility for the impact of one’s cline in civic engagement (Fischer 2004). Consumers’ actions constitutes one crucial part of being an engaged faint sense of connection to their larger communities citizen. Additionally, individuals should come to see how manifests in apathy for societal decision making and broader structural and social change can occur through outcomes (Gaventa 2006). Substantial emphasis is on the involvement at a collective level. As such, societal analy- role of values surrounding consumerism fostered by busi- sis to identify the common good (e.g., individual free- ness and media for consumers’ detachment and thus the doms, equal rights) and collective action is necessary for decline of civic participation. Below we give a brief sustainable democratic societies (Sandlin 2004). Research- overview of relevant points of discussion. ers point out the “democratic deficits” of institutions as obstacles to this end (Gaventa 2006). Businesses, like Researchers highlight the negative influence of consum- governments, dominate social and economic decision erism in cultivating isolation from traditional communi- processes, leaving little room for citizens to engage or ties and the resulting decline of civic participation. For hold executives accountable for their use of power. The example, commercialization of mass media identifies results are often non-inclusive political and economic materialistic values, individual needs and choices as build- decision making and resulting resolute policies. Direct ing stones of a culture (McLeod, Scheufele, and Moy democracy, enacted through transparent dialogue be- 1999). Thus, an individualistic mass consumer society is tween citizens, government and business, is promoted for created as a major source of social influence, at the remedying citizens’ detachment from decisions relevant expense of a civic culture (e.g., Chaplin and Roedder- to their lives (e.g., Gaventa 2006). Corus advocates for John 2007). Accordingly, being a consumer and being a methods of direct democracy, while, in an interesting citizen are conflicting roles and consumption and democ- contrast, Lagos’ research surfaces a mediated form of ratization of society are incompatible goals. governance that places the mass media as the intermediary that shapes citizen-government interaction. An alternative to this view is that of conscientious consumers who expresses their values and beliefs through In sum, the papers in our session contribute to the discus- their purchase and non-purchase choices (e.g., Kozinets sions in civic engagement, consumption and media at

Marketing and Public Policy Conference Proceedings / 2009 79 Special Session: Democracy, Civic Participation, and Consumption multiple points. Corus highlights the importance of demo- cesses and the resulting diminished capacity of people to cratic processes in stakeholder engagement. Drawing on actively hold the government accountable for its decisions the theory and methods of deliberative democracy and at this time of crisis. elaborate on its potential contribution to understanding marketing’s impact on society. Deliberative methods are Our session highlights the intersection of marketing and advocated as instruments to overcome stakeholder en- civic participation using a multitude of methods and gagement obstacles (forums where different views are theories, encompassing different disciplines (marketing, shared in a non-coercive manner) and to engage citizens communication, and political science). Our objective is to in strategic business policies. Dermody, Hanmer-Lloyd, initiate discussions at the crossroads of media, consump- and Scullion critique the incommensurability of consum- tion, and civic life. As such, the session can add to our erism and citizenship in the context of “green” consump- perspectives on the influence of marketing in improving tion. The concept of politicized consumption is utilized to societal welfare and opens horizons to discuss potential reflect a balance between citizens’ civic engagement and partnerships between civic, private, and public organiza- consumers’ autonomous choice. The authors demonstrate tions to serve public needs in different contexts (e.g., the emergence of politicized consumption where pur- environmentalism, disaster management). This session chase behavior is a form of value expression and construc- will appeal to a wide audience including researchers tion of a civic consumer culture. Lastly, Lagos studies interested in socially responsible consumption, corporate citizen involvement in the context of a natural disaster, in social responsibility, public policy makers concerned the birthplace of democracy. Specifically, in the context with improving citizen engagement, and marketing man- of the Great Greek Fires of 2007, the author studies agers seeking to increase stakeholder participation in their people’s dependence on the media for democratic pro- strategies.

80 Marketing and Public Policy Conference Proceedings / 2009 Special Session: Democracy, Civic Participation, and Consumption

EXPLORING THE EMERGENCE OF CIVIC CONSUMER CULTURE THROUGH POLITICIZED SHOPPING

Janine Dermody, University of Gloucestershire, United Kingdom Stuart Hanmer-Lloyd, University of Gloucestershire, United Kingdom Richard Scullion, University of Gloucestershire, United Kingdom

SUMMARY meaning and identity, where consumers are culture-producers. We argue that this alternative view of consump- With consumerism now a dominant ideology and driver tion contends that consumer culture has not replaced civic of much behavior in western society, civic engagement, culture; rather they coalesce to reshape the character of where individuals act as “good citizens,” are purported to citizenship. This is particularly apparent within ethically- be both declining and becoming more individualistic. Yet orientated and sustainable consumption. in evaluating consumerism and its implications for civic society, we unveil a blurring of consumerism and citizen- Thus we are advocating that combining civic and consum- ship, particularly within the realms of politicized con- erist values may hold the key to reinvigorating the health sumption. This metamorphosis indicates that ethical shop- of our planet and our society. We can see evidence of this ping behavior gives consumers a civic voice within the in shopping behavior that has become politicized to fur- marketplace, and, as a result enables the emergence of ther sustainability through human, social, and environ- civic consumer culture. mental capital. Hence “green” consumer-citizens will be pursuing issues surrounding rights, obligations and social Citizenship involves rights balanced with responsibili- exclusion – as part of their economic power of consump- ties, where decision-making involves giving due consid- tion choices for themselves and on behalf of others. eration to justice and to broad societal wishes. Consumer- Typically this will involve boycotts and buycotts in their ism is typically associated with hedonism, instant gratifi- pursuit of sustainability, for example rejecting goods that cation and self indulgence. As a result of these seemingly are not fairly traded, organic, or recyclable. conflicting orientations, critics of consumerism usually perceive citizenship and consumerism to reside at oppo- Consumerism, then, can be empowering, and it has trig- site ends of the spectrum, a contrast between outward- gered a new expression of personal politics for increas- looking, public interest versus private, inward-looking ingly reflexive and politicized consumer groups. We are self-interest, where citizens are “worthy” and consumers witnessing the emergence of a civic consumer culture, are “unworthy.” However this view fails to appreciate rooted in consumerism merged with citizenship, which more contemporary understanding of consumers and their has the potential to enhance the future health of civic “cultures of consumption” (e.g., Arnould and Thompson society, and indeed responsible consumerism in contem- 2005). Within this perspective, consumerism is regarded porary western democracies. as a process of shared social learning, laden with emotion,

For further information contact: Janine Dermody Reader in Consumer Psychology University of Gloucestershire The Park Cheltenham Gloucestershire, UK Phone: +44.01242.714066 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 81 Special Session: Democracy, Civic Participation, and Consumption

DELIBERATIVE DEMOCRACY FOR MULTI-STAKEHOLDER ENGAGEMENT PROCESS

Canan Corus, St. John’s University, Queens

SUMMARY engagement. Allowing citizens to give input into strategic decisions would benefit the corporations directly through Researchers point out the potential benefits of socially consumers’ support and long term demand and indirectly responsible strategies for not only consumer welfare but through gaining the endorsement of the governments for also for long term business. Accordingly, corporations fostering citizen engagement. have confronted increasing expectations of corporate social responsibility (CSR) from civic society in recent Deliberative inclusive processes (DIPs) are proposed as years. Despite the pressing expectations and potential beneficial methods for gaining insight to multifaceted gains of CSR, the relationship between the corporation interests of stakeholders (e.g., Ozanne, Corus, and and stakeholders is still largely bounded by the firm’s Saatcioglu). These methods are grounded in the theory of financial performance and legal accountability. Most of- deliberative democracy, which advocates for “reflective ten, CSR efforts are driven by objectives like abiding with and reasonably open-ended discussion under conditions the law to avoid penalty or gaining consumer endorse- of general equality” (Button and Mattson 1999, p. 609). ment to maintain demand (Garvey and Newell 2005). These methods are geared to build spaces free from Socially responsible marketing should go beyond these domination or deception, where all participants are ca- concerns to encompass the interest of all those who are pable of building and criticizing arguments. Discussions affected by the firm’s decisions. of various methods are included as well as the evaluative criteria for assessing the efficacy of these democratic One driver of the currently bounded corporate societal processes (democratic validity, process validity, dialogi- involvement is that most firms focus solely on the output cal validity, outcome validity). Two case studies regard- of their CSR initiatives (e.g., meeting benchmarks or ing the controversial issue of the genetically modified quotas specified by law, assessments of local economic foods are provided: Canadian Consensus Conference and development). This paper emphasizes the importance of Prajateerpu Citizens’ Jury. These initiatives bring to- the process in CSR initiatives (e.g., democratic, non- gether citizens, policy makers, and business people to coercive discussions) if they are to include and benefit reflect their perspectives and knowledge. The two cases citizens who are affected by corporate decisions. Specifi- are evaluated and critiqued in reference to the aforemen- cally, it is argued that CSR efforts would benefit from tioned evaluative criteria. more inclusive and democratic methods for stakeholder

For further information contact: Canan Corus Department of Marketing St. John’s University 8000 Utopia Parkway Queens, NY 11439 Phone: 718.990.1793 E-Mail: [email protected]

82 Marketing and Public Policy Conference Proceedings / 2009 Special Session: Democracy, Civic Participation, and Consumption

THE CONSUMPTION STRATEGIES OF THE POOR AND THE MEANING OF POVERTY

Bige Saatcioglu, Virginia Tech, Blacksburg

SUMMARY then explore a wide range of consumption strategies and coping mechanisms they employ. Traditionally, poverty is treated as merely an economic problem; the poor suffer from material deprivation. Im- Mobile homes, once a low-cost opportunity for blue- plicit in this approach known as the “absolute poverty” is collar workers to realize the middle-class American dream a one-dimensional deprivation in which individuals lack of home ownership and upward social mobility, have economic capital to meet their primary material and physi- turned into degraded forms of housing since the 1960s. cal needs (Lister 2004). Consistent with this tradition, Their unusual appearance and potential mobility repre- much consumer research on poverty focuses on economic sent a threat to conventional American housing ideals and aspects of resource-constrained consumers’ lives while norms. Moreover, the negative social stereotyping of overlooking social, cultural, and motivational dynamics mobile home residents as dirty, lazy, and criminal-minded (see for exceptions Hill and Stamey 1990; Hill 2001; individuals further contributes to the stigmatization of Chakravarti 2006). Thus, there is a gap in understanding mobile home parks as “white trash icons” (Bérube and how multidimensional nature of poverty is perceived and Bérube 1997). Consequently, my findings suggest that managed by poor consumers. mobile home residents are stigmatized on various dimensions and experience different forms of social devaluation An alternative perspective, namely “relative poverty,” and exclusion. Within such proliferation of co-existing takes into consideration different types of deprivations stigmas as well as material and nonmaterial insecurities, and vulnerabilities experienced by poor consumers. Here, informants develop many strategies to either manage or poverty is not primarily economic and material shortage transform those marketplace dynamics that stigmatize of resources but it also involves a lack of socio-culturally and exclude. For instance, some informants take part in perceived necessities and social exclusion from meaning- the life of their communities (e.g., trailer park community, ful interactions and exchanges (Bauman 2000). As such, the town council, religious communities) via active civic relative poverty becomes a state of multidimensional engagement. Such participation into meaningful social deprivation characterized by both materialized and interactions and civic activities help them cope with nonmaterialized inequality (Sen 2004). My dissertation feelings of isolation, exclusion, and alienation from main- explores these complex forms of insecurities and alien- stream consumer society. Others take a more consump- ation experienced by resource-constrained consumers in tion-oriented approach to poverty management and reject the marketplace. Taking an ethnographic approach within the label of “flawed, blemished, and defective consumer the context of a mobile home park, I investigate how the manquées” (Bauman 2005) through middle-class con- poor negotiate their social identity and manage multiple sumption aspirations and goals. Yet for others who choose deprivations in everyday life. Consistent with the relative to mainly survive, poverty is experienced as a vicious approach to poverty, I first untangle different types of cycle of motivational loss, circles of dependency, and actual and perceived deprivations (both material and chronic capability deprivation. References are available nonmaterial) experienced by mobile home residents. I upon request.

For further information contact: Bige Saatcioglu Virginia Tech 2016 Pamplin Hall Blacksburg, VA 24061 Phone: 540.231.1962 ♦ Fax: 540.231.3076 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 83 SPECIAL SESSION

15 YEARS SINCE NLEA REGULATIONS: EMERGING ISSUES, NEW VARIABLES, AND HOW NUTRITION LABELING REALLY WORKS

Session Chair: Alan S. Levy, U.S. Food and Drug Administration, College Park

Discussant: John Kozup, Villanova University

SUMMARY mation on product labels such as footnotes, disclaimers, dietary guidance messages or “see back panel statements.” The regulations implementing the Nutrition Labeling and Claim proliferation has also highlighted the role of prior Education Act (NLEA) that govern how nutrition infor- knowledge and familiarity in mediating the effects of mation is provided on food product packaging became claims encountered on product packaging as well as the effective 15 years ago. In that time, the promises, limita- role of various kinds of priming, whether due to advertis- tions and problems inherent in communicating to con- ing, news coverage or self-directed information search, sumers about nutrition have become clearer. Nutrition on the likely consumer impacts of nutrition claims. These labeling has not proved to be a panacea for key dietary issues and others are addressed in the papers in the health challenges such as obesity, nor has it proved a session. godsend to effective nutrition education. On the other hand, a good deal of evidence has accumulated on how The range of nutrition claims being made in the market- consumers use nutrition information on food labels, and in place raises the methodological issue of how to evaluate particular on the interplay between front panel claims and when claims can be misleading. Several of the papers in back panel nutrition disclosures. In this session, we present the session look at the information search consequences a set of studies that focus on post-NLEA concerns that of claims when product nutrition profile information is shed considerable light on how nutrition labeling on food freely available but the contents of this information are packaging actually works. manipulated. Information search truncation, the typical effect observed in the presence of front panel claims, The expansion of nutrition claims in the marketplace reduces consumers’ abilities to discriminate between prod- beyond the realm of claims supported by scientific con- ucts of varying nutrition quality. The findings suggest that sensus as originally contemplated by the NLEA is a key whether a nutrition claim is misleading may not be an feature of the post-NLEA marketplace. Claim prolifera- inherent property of the claim, but a property of the tion has spurred consideration of new issues to study with product to which the claim is applied. By reducing dis- respect to nutrition information labeling on food products. crimination when applied to undeserving products, nutri- Because the reliability of claims in the marketplace may tion claims may pose risks to consumer understanding of not meet original NLEA standards, research has tried to the appropriate role of such products in the diet. assess the value of providing additional clarifying infor-

84 Marketing and Public Policy Conference Proceedings / 2009 Special Session: 15 Years Since NLEA Regulations: Emerging Issues, New Variables, and How Nutrition Labeling Really Works

For further information contact: Alan S. Levy Office of Regulations, Policy and Social Sciences Center for Food Safety and Applied Nutrition U.S. Food and Drug Administration 5100 Paint Branch Parkway, HFS 727 College Park, MD 20740 Phone: 301.436.1762 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 85 Special Session: 15 Years Since NLEA Regulations: Emerging Issues, New Variables, and How Nutrition Labeling Really Works THE CONSUMER IMPACT OF CARBOHYDRATE CLAIMS: WORDING EFFECTS AND FRONT LABEL/BACK LABEL INTERACTIONS

Judy Labiner-Wolfe, U.S. Food and Drug Administration, College Park

SUMMARY The data indicate that in the absence of Nutrition Facts label information: (1) respondents perceived that prod- Current FDA regulations make no provision for the use of ucts with claims or statements that implied “low in carbo- nutrient content claims that characterize the level of hydrate” (“Low Carb,” “CarbConscious,” and “Net Carb”) carbohydrate in foods. Consumer and industry groups were lower in carbohydrate than those without such have petitioned FDA for a regulation to define terms for claims or statements; (2) respondents perceived products use in nutrient content claims characterizing the level of with claims or statements that implied “low in carbohy- carbohydrate in foods. This study was designed to evalu- drate” as being more helpful for weight management than ate carbohydrate content claims and disclosures requested those without such claims or statements; (3) respondents in the six petitions, as well as statements used in the often perceived different claims and statements that im- marketplace, in terms of their effects on consumer un- plied “low carbohydrate” to have the same meaning as one derstanding and consumers’ ability to make appropriate another; (4) disclosures on the front panel did not consis- product judgments for healthy dietary practices. tently redress any potentially inappropriate perceptions associated with claims or statements that implied “low in An online experimental study with a sample size of over carbohydrate”; and (5) reaction to the “Good Source of 9,400 was conducted, allowing for a minimum of 180 Carb” claim depended on the product on which the claim respondents per product and label condition. Each re- appeared. When the Nutrition Facts label was also avail- spondent saw a single product/label condition. A label able, (6) respondents made appropriate product judg- condition was a carbohydrate content claim/statement ments, despite the claim or statement on the front panel. alone, with a Nutrition Facts label, or with a disclosure Products with the same Nutrition Facts label were rated and/or a Nutrition Facts label. The products in the study similarly and those with the more healthful Nutrition were a loaf of bread, a 12-ounce juice drink, and an Facts label were rated as such. individual frozen dinner. Data were collected in March 2006.

For further information contact: Judy Labiner-Wolfe Consumer Science Specialist Consumer Studies Staff Division of Social Sciences Office of Regulations, Policy and Social Sciences Center for Food Safety and Applied Nutrition U.S. Food and Drug Administration 5100 Paint Branch Parkway, HFS 727 College Park, MD 20740

86 Marketing and Public Policy Conference Proceedings / 2009 Special Session: 15 Years Since NLEA Regulations: Emerging Issues, New Variables, and How Nutrition Labeling Really Works CONSUMER ABILITY TO INTERPRET TRANS FAT DISCLOSURES ON THE NUTRITION FACTS OF FOOD PRODUCTS

Conrad J. Choiniere, U.S. Food and Drug Administration, College Park Alan S. Levy, U.S. Food and Drug Administration, College Park

SUMMARY and fewer calories than Product A. Subjects chose the product they considered more healthful, gave reasons for Trans fat is one of several nutrition components that their choice, rated five specific product health benefits constitute a food product’s fatty acid profile. Calories, and estimated 6 specific nutrient levels of the chosen total fat, calories from fat, saturated fat, trans fat, and product. The experimental design was a partially crossed cholesterol are required disclosures in Nutrition Facts, factorial where the independent variables were magnitude while polyunsaturated fat and monounsaturated fat may of quantitative differences between the fatty acid profiles be voluntarily disclosed. These components are logically of the two products, the format of the Nutrition Facts and hierarchically related to each other, but how consum- (combined %Daily Value for trans and saturated fat vs. ers integrate this information when they assess an overall %Daily Value for saturated fat only), the presence or product nutrition profile is not obvious. Some research absence of accompanying dietary guidance messages, suggests that consumers may unduly overweight the value and whether or not subjects were primed with a brief of 0 gram trans fat declarations relative to a product’s summary of the scientific evidence about the health ef- overall fatty acid profile. Canada uses a modified nutrition fects of trans fats, saturated fat and cholesterol. facts format that combines the amount of saturated and trans fats in the product and expresses the combined The results indicated that many subjects were more im- amount as a single %Daily Value. The National Academy pressed by zero gm trans fat declarations than by small of Sciences’ Institute of Medicine recommends a similar differences in total fat and calories between products. approach for the U.S. label. FDA asked for comments Relative preference for zero gm declarations became about the possibility that explanatory footnotes might stronger when subjects were primed with a brief scientific help consumers better understand the fatty acid profile of evidence summary. However, subjects were not impervi- a product. The ability of consumers to decipher informa- ous to the magnitude of fatty acid profile differences tion contained in the Nutrition Facts is a key foundational between products and responded appropriately to these assumption that has implications for both the design and differences. Accompanying footnotes had no effect on content of nutrition labeling. how subjects responded to the Nutrition Facts information. The format for declaring %Daily Value had signifi- Subjects were recruited from a large consumer internet cant effects on subject’s judgments, but did not overcome panel for a study of food labeling. Each subject saw the preference for 0gm trans fat declarations. In some Nutrition Facts for a pair of products (A and B) from the instances the combined %DV format actually increased same food type (i.e., margarine, donut, lasagna). The the preference for products with 0gm trans fat. Subjects nutrition profile of the products was manipulated so that responded to large differences between %Daily Value Product A always had zero gm of trans fat, but more levels for the two products, but such differences depended saturated fat than Product B. Product B had less total fat on which format was used to declare a %Daily Value.

For further information contact: Alan S. Levy Office of Regulations, Policy and Social Sciences, Center for Food Safety and Applied Nutrition U.S. Food and Drug Administration 5100 Paint Branch Parkway, HFS 727 College Park, MD 20740 Phone: 301.436.1762 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 87 Special Session: 15 Years Since NLEA Regulations: Emerging Issues, New Variables, and How Nutrition Labeling Really Works CONSUMER IMPACTS OF TRANS FAT CLAIMS IN THE PRESENCE OF STRENGTH OF SCIENCE AND MARKETING PRIMES WHEN PRODUCT NUTRIENT QUALITY IS MANIPULATED

Alan S. Levy, U.S. Food and Drug Administration, College Park Conrad J. Choiniere, U.S. Food and Drug Administration, College Park

SUMMARY varied from below average to above average for the product type. Subjects were primed with short informa- Increased numbers of food products are being marketed tion pieces that preceded the presentation of any product. that completely or partially replace partially hydroge- Priming conditions were comprised of either a brief sum- nated fats (i.e., trans fat) with alternative kinds of fat. This mary of the scientific evidence about the health effects of development has spurred interest in providing for trans fat trans fat, a positive marketing justification for providing claims on food labeling. Currently, manufacturers are health claims on product labels, a more critical marketing allowed to make truthful factual statements such as “con- justification for providing health claims, or a no priming tains zero gm trans fat,” but they are not allowed to make control. Subjects answered questions about the perceived “reduced trans fat” or “trans fat free” claims. The objec- healthfulness of the product, rated specific health benefits tive of the study was to evaluate the likely consumer and estimated the magnitude of specific nutrient levels in behavior effects of allowing trans fat claims on food the product. products. The results indicated that claims led subjects to view a Subjects were recruited from a large consumer internet product as healthier, with greater specific health benefits panel for a study of food labeling. Subjects were given a and more favorable nutrient levels attributed to the prod- front panel for either a margarine, cracker, or pound cake uct. Trans fat free and 0 gm trans fat claims had stronger product which contained one of several possible versions effects in these respects than reduced trans fat claims. In of trans fat claims along with one of several possible kinds the presence of a front panel trans fat claim, subjects were of clarifying statements. Both the front panel and the less likely to consult the Nutrition Facts and were less Nutrition Facts for the product were shown initially, but likely to discriminate between products of varying nutri- the Nutrition Facts was removed from the screen once tion quality. When asked about specific nutrient levels, subjects began to answer questions about the product. For however, subjects were more likely to consult the Nutri- each question, subjects could retrieve the Nutrition Facts tion Facts. Strength of science primes had no effect, but for inspection by clicking on a button on the screen. If and both positive and critical marketing justifications reduced when subjects chose to view the Nutrition Facts, it was the observed impact of trans fat claim relative to a no recorded. The content of the Nutrition Facts was manipu- priming control. lated so that the overall nutrition profile of the product

For further information contact: Alan S. Levy U.S. Food and Drug Administration Office of Regulations, Policy and Social Sciences, Center for Food Safety and Applied Nutrition U.S. Food and Drug Administration 5100 Paint Branch Parkway, HFS 727 College Park, MD 20740 Phone: 301.436.1762 E-Mail: [email protected]

88 Marketing and Public Policy Conference Proceedings / 2009 Special Session: 15 Years Since NLEA Regulations: Emerging Issues, New Variables, and How Nutrition Labeling Really Works BARRIERS TO FOOD LABEL USE AMONG YOUNGER CONSUMERS

Conrad J. Choiniere, U.S. Food and Drug Administration, College Park Anu Mitra, American University, Washington, D.C.

SUMMARY way in which single servings of foods fit into an overall daily diet. A set of focus groups were held to gain a better Recent analyses of FDA’s Health and Diet Surveys understanding of the attitudes and beliefs of younger (Choiniere and Mitra 2006) show a marked increase in the adults concerning nutrition, health and the use of the food percentage of American consumers that report not using label. The focus groups were comprised of consumers the food label when purchasing a new food product. The who report “never” using the food label for a first time percentage of all consumers reporting that they never use purchase, stratified by age (35 years old and younger, the label when purchasing a product for the first time older than 35 years) and gender. The groups were con- increased 43 percent between 1994 and 2002. In 1994, ducted in Washington, D.C. in July 2008. rates of non-use in various subgroups of the population ranged between 10 percent and 15 percent, with few Discussions that emerged from the focus groups reveal significant differences related to age. By 2002, there were that reasons for non-usage may be categorized into four significant increases in consumers in every population predominant themes: (1) low levels of awareness and subgroup that opted not to use the label at all, with nearly knowledge of nutrition and the NF; (2) weak levels of 20 percent of the general population being non-users in belief and confidence in the importance of nutrition in 2002. However, younger consumers, i.e., those born be- reducing the risk of disease; (3) strong beliefs in one’s tween 1973 and 1984, had disproportionately higher rates ability to select healthful foods through the use of simple of decline in label use over that period. Nearly 30 percent heuristics not related to the use of the NF; and (4) high of these younger consumers reported not using the food levels of perceived difficulty in using the NF to make label in 2002 – a 130 percent increase over 1994. healthful decisions. The presentation explores these themes in detail, provides comparisons between younger and This finding is alarming because FDA relies on the Nutri- older non-users of the food label, and draws implications tion Facts (NF) label to provide consumers with objective of the findings for public information and education information about the nutrient content in foods and the initiatives that may help promote label use.

For further information contact: Condrad Choiniere U.S. Food and Drug Administration Senior Scientist Office of Regulations, Policy and Social Sciences Center for Food Safety and Applied Nutrition U.S. Food and Drug Administration 5100 Paint Branch Parkway, HFS 727 College Park, MD 20740

Marketing and Public Policy Conference Proceedings / 2009 89 CHANGES IN U.S. HOME EQUITY LENDING: EVIDENCE FROM THE 1989 THROUGH 2007 SURVEY OF CONSUMER FINANCES

Jeanne M. Hogarth, Federal Reserve Board, Washington, D.C.1 Andrew D. Vlietstra, Federal Reserve Board, Washington, D.C.

SUMMARY acteristics on the ratio for those with home equity loans. The ETV ratio takes on values between zero (if the home Given the recent turmoil in U.S. mortgage markets, and owner has no equity) and one (if the home owner owns the the importance of housing both to the economy in general home outright). and to household net worth in particular, we use the Federal Reserve Board’s Survey of Consumer Finances Preliminary results indicate that risk characteristics, age, from 1989 to 2007 to explore changes in home equity income, and source of financing are significantly associ- lending since 1989.2 Our analysis includes both bivariate ated with the ETV ratio. Furthermore, households in 1989 analysis and multivariate regression to explore these had significantly higher ETV ratios than those in 2004. changes. What Financial Institutions Do Consumers Use for How Has the Incidence of Having a Home Equity Home Equity Loans and Have These Changed Over Loan or Home Equity Line of Credit Changed Over Time? Time? It is still the case that the largest proportion of home equity Overall, the proportion of home owners had any type of borrowers, around 40 percent, get their loans from banks. home equity loan increased from 1989 to 2004, from 13 However, in 2004, higher proportions of borrowers were percent to 18 percent. However, this overall increase in going to credit unions (up 70%) and mortgage companies home equity lending masks the differences between (up nearly 250%). closed-end home equity loans (which decreased 30%, from 9% of home owners to 6% of home owners) and Has the Reason for Using Home Equity Loans and home equity lines of credit (which more than doubled Lines of Credit Changed Over Time? from 5% to 12% of home owners). People use home equity loans for a variety of reasons, from How Has the Level of Home Equity Changed Over home improvement to paying for college to consolidating Time? bills. Under some financing programs, consumers can also use home equity loans to supplement a primary mort- Among all home owners, the home equity-to-home value gage – in this case, the consumer obtains a primary mort- ratio (ETV) has declined from 1989 to 2004, by about 12 gage for 80 percent of the value of their home and a home percent. For those home owners with home equity loans equity line of credit or home equity loan for some or all of or home equity lines of credit, the ETV has declined 16 the remaining 20 percent. Over time, the use of home percent; the mean equity-to-value ratio in 1989 among equity loans for home purchase has increased from 19 home equity borrowers was 0.58 in 1989 compared with percent to 27 percent of those with home equity loans, a 42 0.49 in 2004. percent increase. Use of home equity loans for home improvement has declined 88 percent, while use for other To study the changes over time in a more robust frame- purposes (for example, college expenses) has increased 43 work, we regressed a set of risk characteristics, socio- percent over time. economic and demographic characteristics, and loan char-

90 Marketing and Public Policy Conference Proceedings / 2009 ENDNOTES the presentation by the author does not constitute an endorsement or criticism by the Federal Reserve 1 The analysis and conclusions set forth in this presenta- System and does not imply approval to the exclusion tion represent the work of the authors and do not of other suitable products or firms. indicate concurrence of the Federal Reserve Board, 2 By the abstract deadline we have not finished incorpo- the Federal Reserve Banks, or their staff. Mention or rating the 2007 data into the study; these results are display of a trademark, proprietary product, or firm in only preliminary.

Marketing and Public Policy Conference Proceedings / 2009 91 CONSUMER DESKILLING AND THE CREDIT CARD INDUSTRY: AN INTERGENERATIONAL PERSPECTIVE

Sarita Ray Chaudhary, New Mexico State University, Las Cruces Yam B. Limbu, New Mexico State University, Las Cruces

SUMMARY industry. We explore the parent-child intergenerational transfer of knowledge to determine if such interactions As worldwide history is being written today on the cata- shape credit card acquisition and usage behaviors in strophic repercussions of irresponsible lending practices college students. and misinformed consumer credit decisions, marketing scholars in the last several years have raised the alarm on Our review of extant literature reveals that college stu- a vulnerable consumer group – the college student dents as credit card consumers do not have or are system- (Braunsberger, Lucas, and Roach 2004; Mansfield and atically dispossessed of the information, knowledge, and Pinto 2008; Montgomerie 2006). College students are skills needed to make informed financial decisions (Chen much sought after by credit card companies for several and Volpe 1998; Makela, Punjavat, and Olson 2007). The reasons. Roach (2004) estimates the annual spending notion of consumer deskilling is therefore useful in exam- power of college students in the U.S. to be upwards of ining activities of the credit card industry given the con- $120 billion. Loyalty to the first cards acquired in college cerns regarding its marketing practices. First introduced results in intense competition to capture this consumer in the realm of food consumption by Jaffe and Gertler group (Lyons 2004; Palmer, Pinto, and Parente 2001). (2006), the concept of deskilling originates from labor College students also comprise a steady stream of new theory (Braverman 1974). Deskilling alludes to people customers every year and with efforts to legislate parents who have lost the skill sets to comprehend production in as co-signors have become a more desirable group than its entirety and are micro-managers of individual tasks before (Pinto and Mansfield 2006). Although prior re- (Ritzer 1993). Braverman (1974) traces the earliest his- search has identified several important concerns, this tory of North America when consumption and production study outlines two previously unexplored issues. were comparable activities. However, the balance of power leaned more toward the producers as consumer The first problem is explored by examining marketing self-preservation skills started to depreciate. People be- activities of credit card products targeted at college stu- gan to increasingly rely on goods and services and subse- dents. The intent is to examine if these marketing activi- quently the information provided by large corporations. ties empower consumers to make an educated decision or Jaffe and Gertler (2006) apply the notion of deskilling to if they actively seek to discourage consumers from taking examine the effect of agro-food industry’s marketing the time and effort to understand the implications of activities on the ability of the consumer to make informed purchasing and using credit products. There is evidence decisions on food choices. Often college students are that many college students do not understand the com- enticed by minimum payment options, introductory low plexities of credit card products and are unable to discern interest rate offers and free gifts. In this context, we aim to financial concepts and terminologies (Braunsberger, examine if college students persuaded by such marketing Lucas, and Roach 2005; Warwick and Mansfield 2000). tactics, forego acquiring the necessary skills required to The second question under examination is the understand and use credit card products. intergenerational transfer of financial knowledge between parents and children. Consumer behavior studies con- With its origin in socialization theory, intergenerational ducted over the past twenty-five years concur that parents influence refers to the dissemination of beliefs, knowl- are the most important agents of socialization from the edge and resources from one generation to the next earliest years of children (John 1999; Moschis 1985; (Moore, Wilkie, and Lutz 2002). Since early social skills Moschis and Moore 1979). Students whose parents are are largely acquired through intergenerational transfers, involved in the acquisition of credit cards tend to have we also explore the financial knowledge students imbibe lower credit card balances (Palmer, Pinto, and Parente from their parents and whether there is a prevalent trend 2001). It is possible that today’s college students and their toward lost skills from one generation to the next. Exam- parents have grown up in times of easy credit with an ining intergenerational influence makes consummate sense inherent trust of information provided by credit card as today’s college students grew up observing and learn-

92 Marketing and Public Policy Conference Proceedings / 2009 ing from their parents’ credit card use. A recent study plex financial information. It has emerged as one of the suggests that parents’ credit card use influences their major drawbacks of the consumer of credit. Our prelimi- children’s credit card consumption behavior and attitudes nary findings indicate college students lack the required (Joo, Grabel, and Bagwell 2003). We explore parental proficiency to use and manage credit card products, a trait knowledge transfer and its effects on college students to that may be aided and abetted by the credit card industry. determine if consumer deskilling is taking place across Similarly, the intergenerational transfer of knowledge generations. brings attention to the role of parental influence as there appears to be a discernible disparity between the knowl- This study is exploratory in nature as there is little extant edge of students and their parents. Credit card debts have theory available. A qualitative approach is utilized to far reaching consequences for college students as con- discover meanings and definitions from observed phe- sumers and productive members of society. It may affect nomena (Charmaz 2000). Personal life-history interviews their credit histories, career opportunities, and even higher (Mick and Buhl 1992) of fifteen student-parent dyads education opportunities (Roberts and Jones 2001). Piece- from a Southwestern city in the U.S. are conducted. The meal applications such as parental co-signage or large participants are asked to explain contents of their credit type font mandate in solicitation documents may well card statements such as APR, credit limit, credit balance, prove to be inadequate. Despite bans on some college on- payment cycle, fees structure, disclosure and other terms campus solicitations, students are extensively targeted and conditions. They are also shown actual direct market- through direct marketing solicitations. Clearly the credit ing solicitations of two major credit card companies with card industry and its marketing communication activities brand identifiers removed and asked to explain the con- warrant closer scrutiny. It is a large scale problem where tents. Students are asked how they learn to interpret the all stakeholders, including students, parents, universities, information, specifically whether parental influence plays credit card corporations, public policy makers and legis- a role. Parents are also asked if they discuss credit card lators, should accept the challenge of resolving this issue related issues with their children. together. Since this is a data driven study, we plan to complete the analysis and present the findings using the The present crisis in the finance industry is an indication constant comparative method recommended by Wolcott that consumers are ill-equipped to navigate through com- (1994). References are available upon request.

For further information contact: Sarita Ray Chaudhary New Mexico State University P.O. Box 30001, MSC5280 Las Cruces, NM 88003–8001 Phone: 575.636.2186 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 93 MEASURING CONSUMER PERCEPTIONS OF FAIRNESS IN FINANCIAL SERVICES

James Devlin, University of Nottingham, United Kingdom

SUMMARY ment of fairness is characterized as uni-dimensional and key themes include consideration as to how the pie is In the U.K., ensuring fairness in firms’ dealings with shared out and how the benefits and costs are divided consumers in financial services markets is seen as a between relevant parties. Some authors have refined the crucial strand of Government policy. By encouraging procedural/distributive distinguishing between procedural firms to make sure that treating customers fairly is an fairness and interactional fairness, the latter being the integral element of strategy, policymakers hope that con- inter-personal aspects of an exchange (see for instance sumers will be encouraged to engage with financial ser- Smith et al. 1999; Patterson et al. 2006; Tax et al. 1998). vices to a more satisfactory degree than is the case at Theoretical perspectives on fairness and justice have been present. Although there has been much written on the offered from a general perspective rather than with refer- topic, primarily by regulators, thus far there have been ence to financial services specifically, however, the key few, if any attempts to measure perceptions of fairness principles outlined above contain important insights for from the consumer’s perspective. Therefore, the primary the current project. aim of the project outlined here is to devise a generic tool which can act as a means of determining perceptions of The issue of fairness has also attracted significant atten- fairness at the level of the individual company, industry tion from U.K. policymakers and is viewed as a crucial sector (e.g., banking, life assurance, etc.) and for the strand of consumer protection policy. The primary regu- financial services sector overall. Such a measure, by its lator of retail financial services in the U.K. is the Financial very nature, provides an assessment of perceptions of Services Authority. The FSA has facilitated much discus- fairness, rather than actual fairness, however, it will sion of the meaning of fairness in a financial services provide valuable insights, for instance enabling compari- context and has identified six key Treating Customers sons across sectors, relating perceptions of fairness to Fairly (TCF) outcomes which it believes are critical for certain outcome variables and establishing which ele- the fair treatment of consumers (Financial Services Au- ments of fairness are more important in explaining overall thority 2007): Outcome 1: Consumers can be confident perceptions or fairness. The project is important as, al- that they are dealing with firms where the fair treatment of though under the U.K. regulatory structure firms are customers is central to the corporate culture: Outcome 2: charged with devising metrics to provide evidence that Products and services marketed and sold in the retail they are treating customers’ fairy, thus far little consensus market are designed to meet the needs of identified con- has emerged as to what constitutes fairness in the context sumer groups and targeted accordingly: Outcome 3: of financial services and how to measure whether custom- Consumers are provided with clear information and are ers are being treated fairly. kept appropriately informed before, during and after the point of sale: Outcome 4: Where consumers receive From a theoretical perspective, fairness is made up of two advice, the advice is suitable and takes account of their distinct elements (Kumar 1996). One is the fairness asso- circumstances: Outcome 5: Consumers are provided with ciated with the process used to determine outcomes, or the products that perform as firms have led them to expect and process used to manage exchange relationships. This is the associated service is both of an acceptable standard called procedural fairness. Key themes in discussions of and as they have been led to expect: Outcome 6: Consum- procedural fairness are the importance of due process and ers do not face unreasonable post-sale barriers imposed by the fairness of a party’s policies and procedures. Six firms to change product, switch provider, submit a claim principles of a procedurally fair exchange have been or make a complaint identified: bilateral communication, impartiality, re- futability, explanation, familiarity, and courtesy. The FSA requires firms to show how these core principles have been incorporated into their conduct of business The other main dimension of fairness is concerned with procedures. Firms have also been charged with deriving the fairness of the economic price and outcomes actually metrics to illustrate that they are treating customers fairly. achieved. This is termed distributive fairness. This ele- It is apparent that these six outcomes incorporate elements

94 Marketing and Public Policy Conference Proceedings / 2009 of both procedural and distributive fairness and exhibit fifty randomly selected members of the population. The significant communality with the themes identified above. resultant data were then analyzed to check the reliability As a result, they are a further key source of guidance in our and validity of the research instrument. The vast majority attempt to derive a generic measure of perceptions of of the scales were shown to be highly reliable, with fairness in financial services. Cronbach’s Apha statistics generally of the order of 0.9 or above, and valid in that they were shown to possess uni- Research Approach and Analysis Strategy dimensionality when subject to EFA. Two scales required further attention, with one item being dropped from the The development of the questionnaire was informed by distributive justice scale (as it caused the scale to become both the theoretical perspectives on fairness and the multi-dimensional and introduced issues of cross-load- FSA’s approach to measuring fairness outcomes, as well ing) and one item being re-worded for the behavioral as a consultation exercise carried out amongst the mem- loyalty scale. Subsequently, the final version of the ques- bership of the Financial Services Research Forum (FSRF) tionnaire was employed to collect data on perceptions of (see Cartwight 2008). The FSRF is an industry- fairness in the context of banking (which was used for the policymaker-academic alliance with over fifty organiza- pilot) as well as building societies, general household tional members. Access to FSRF members proved ex- insurers, life assurance companies, investment compa- tremely useful in the initial piloting of the research instru- nies, brokers/advisers and credit card companies. One ment. Efforts to measure fairness in exchange relation- hundred responses were gathered for each context, result- ships in other contexts also provided useful insights. For ing in a total sample size of 700. the project, a questionnaire was derived which measures perceptions of fairness by employing multi-item mea- The previous paragraph summarizes progress at the time sures for the six dimensions of procedural justice and for of writing of the abstract. Subseqently, CFA will be distributive justice, as well as a single-item global assess- employed to check the reliability and validity of the final ment of fairness and some outcome measures, such as measurements. Factors means for the various contexts loyalty and positive word of mouth. Whilst the question- will then be employed to allow comparisons of percep- naire is explicitly structured around theoretical perspec- tions of fairness across contexts. The key drives of overall tives, it also has resonance with, and is significantly perceptions of fairness will also be identified, both for informed by, the policy agenda and the approach of the financial services generally and each of the contexts FSA. mentioned. All data collection and analysis will be completed in time for the conference. References are available A market research agency was then employed to pilot a upon request. draft version of the questionnaire amongst a sample of

For further information contact James F. Devlin Business School Room B 78d, North Building University of Nottingham Jubilee Campus Wollaton Road Nottingham NG8 1BB United Kingdom Phone: +44(0)115.951.5264 Fax: +44(0)115.846.6667 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 95 ONLINE BANKING AND IDENTITY THEFT: UPDATE AND CALL FOR ACTION

Ruxandra Niculescu, Rutgers University School of Business, Camden Mark O’Lexy, Rutgers University School of Business, Camden Ryan Schiffino, Rutgers University School of Business, Camden Carol Kaufman-Scarborough, Rutgers University School of Business, Camden

SUMMARY Unfortunately, studies have found that consumers too readily provide sensitive information posted in various Identity theft is a growing concern for many consumers locations throughout the online environment, especially and has been called the “fastest growing crime in America” in social networking sites such as MySpace and Facebook. (Eisenstein 2008). This paper provides an overview of Various page features include information that is often identity theft statistics, current identity theft methods, and used to create security questions for banking and credit proposed solutions using the banking industry as a case card sites, such as mother’s maiden name, hometown, example. Implications for public policy are also dis- pet’s names, birthdays, and favorite foods. Pritchard (2008) cussed. suggests consumers should be made aware that any type of online data input or posting can “cast a digital shadow” The number of people worldwide using online banking that can remain in cyberspace indefinitely. today is rising at a tremendous rate (Credit Management 2008). “Figures released in late July (2008) by the APACS, Business owners should be proactively responsible for the U.K. payments association, show that the number of protecting the personal data of their customers and em- adults using online banking has increased by 505 percent ployees. Proper safeguards must be in place to minimize in the past seven years, from less than 3.5 million in 2000 the risk of that data falling into the wrong hands. The to just over 21 million last year.” Interestingly, several problem is widespread. The FTC’s 2006 Identity Theft problems are largely created by design, implementation, Survey reveals that over eight million U.S. residents and flow of websites, rather than design flaws or bugs that reported being victims of identity fraud in 2005 (Filipak could be fixed with a patch. 2008). More recent evidence links an increasing likelihood of identity theft to demographics (Anderson 2006). Numerous problems are found. Many banks use social That is, a higher likelihood of identity theft has been security numbers as logins or identifiers to find individu- linked with having a higher income, being younger, and als in their information databases. In addition, customers being female. are frequently asked to give sensitive information on web pages which are not secure, as customers are directed to For instance, the state of Florida requires that businesses third-party unsecured sites without their knowledge. notify individuals if their personal data has been exposed “Phishing” scams involve scam artists sending out emails due to a breach in their securities. Companies must further that resemble those sent by banks, investment account make an attempt to “contain the breach to ensure personal managers, retirement plans or other financial institutions, information isn’t further compromised and to implement attempting to trick naïve consumers into giving sensitive remedial steps to address information security including information to cyber-criminals who pose as a trustworthy issues implicated by the original breach.” Steps should be person or organization, “Dumpster diving” entails search- taken not only to secure data, but to rectify any known ing through the victim’s trash in order to find either utility problems once they are identified. Based on the number of bills or financial records which can help a scammer identity theft cases that transpire each year, it seems that impersonate the victim. In a more recent development, most financial institutions may not be doing all they can cyber-criminals scour social networking sites for infor- to fix the current problems. mation that can help them impersonate victims, mining various databases of personal information, effectively Public advocates and industry experts alike have argued reconstructing enough of an individual’s personal iden- that the federal and state governments should enact laws tity to break into their accounts. that protect consumers against identity theft (Haygood

96 Marketing and Public Policy Conference Proceedings / 2009 and Hensley 2006). Researchers are needed to analyze the tomers. Bank of America uses “the three challenge ques- impact of the Fair and Accurate Credit Transactions Act tion” system also known as Site Key (“Wireless News” (FACTA) and the Identity Theft Penalty Act (ITPEA). 2005). Simply stated, the online system lets you chose There have been a documented 900+ “individual data three questions from a list and input answers that only you breaches by U.S. companies and government agencies would know. Customers can also do all their online since January 2005.” These cases have involved over 200 banking securely, including checking accounts often which million records containing personal information. will allow them to detect fraudulent activity immediately.

FACTA is thought to have led to the more recent “Red Public policy actions are needed to ensure that financial Flags Rule” (Roome 2008). “The Red Flags Rule was institutions have provided adequate protection to ensure developed pursuant to the Fair and Accurate Credit Trans- the safety of their customers’ personal information. Rec- actions (FACTA) Act of 2003. Under the Rule, financial ommendations include: (1) clear guidelines for online institutions and creditors with covered accounts must requests for verification information; (2) policies requir- have identity theft prevention programs to identify, de- ing secure logins throughout secure websites; (3) man- tect, and respond to patterns, practices, or specific activi- dated compliance with the FTC Red Flag Rule; (4) the ties that could indicate identity theft.” The article indi- adoption of stronger authentication systems, eliminating cated that the FTC will be granting a six-month delay on the need to use personal information for security pur- the enforcement of the regulations for this new Rule. poses; and (5) regulation of information required to create online accounts, possibly barring the use of social security Some banks have developed a multi-part security process numbers for this purpose. to ensure a higher likelihood of protection for their cus-

For further information contact: Carol Kaufman-Scarborough Rutgers University School of Business 227 Penn Street Camden, NJ 08102–1656 Phone: 856.225.6592 Fax: 856.225.6231 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 97 A FRAMEWORK FOR SOCIAL CHANGE: INTEGRATING MARKETING AND PUBLIC POLICY

David Brinberg, Virginia Tech, Blacksburg Meghan Pierce, Virginia Tech, Blacksburg Kim Daniloski, Virginia Tech, Blacksburg Julien Stauffer, Università della Svizzera Italiana, Switzerland

SUMMARY These goals are often in conflict. For example, accessibility implies meeting a minimum threshold to make infor- The extensive exchange in the Journal of Marketing and mation available to a wide range of consumers, whereas Public Policy in 2007 regarding the American Marketing full disclosure is often intended to address legal or regu- Association’s 2004 and 2007 definitions of marketing latory requirements and is usually much less accessible to highlights the growing recognition among marketing and consumers (e.g., the risk disclosures in direct to consumer consumer behavior researchers that the study of at-risk ads of prescription drugs). Macro theories of societal and vulnerable populations is an appropriate area for behavior (e.g., Rawls’ Theory of Justice 1971) provide a investigation. context for understanding what goals are valued and the interpretation of these goals. In his 2006 presidential address to the Association for Consumer Research, David Mick called for a shift in The theoretical domain contains concepts and their inter- research focus to topics that enrich consumer welfare. The relationships that identify the mechanisms necessary to great need for social change, or the desire to “make the understand and to create change. The elements in this world a better place,” is calling for research that will make domain are constructs drawn from a wide range of disci- a difference. We believe that a conceptual structure that plines, including marketing, psychology, sociology, pub- outlines the key features of social change will provide a lic policy, and economics. The literature abounds with platform to organize discussion of this emerging research examples of concepts and models that guide social change. area. The aim of this abstract is to present, very briefly, a In psychology, for example, the unified theory developed general framework of the key components of social change by Fishbein and others identified key psychological deter- and to highlight some research opportunities that evolve minants of behavior. Prospect theory, which is at the from considering the components in the framework. junction of economics and psychology, provides an explanation of people’s decisions under uncertainty. In We adapt the research framework developed by Brinberg sociology, social constructionism describes how social and McGrath (1985) to better understand the role of phenomena, including knowledge, are derived from so- marketing and public policy on social change. The Valid- cial interaction. The context in this domain is the episte- ity Network Schema (VNS) identifies three domains that mological assumptions that guide the selection and inter- comprise research: conceptual, methodological, and sub- pretation of the constructs and theories. For example, stantive. Each domain contains three levels that represent feminist epistemology is concerned with the way in which elements, relations between these elements, and a context gender influences our creation and interpretation of knowl- that influences the interpretation of the constructs and edge. Realism, on the other hand, views knowledge as their relations. We adapt the general structure of three independent of external influences such as gender. domains and three levels within each domain, but modify the VNS to examine the interplay between marketing and The substantive domain focuses on social phenomena that public policy as an agent for social change. have been the focal concerns of marketing and public policy research. Sprott and Miyazaki (2002) created four The axiological domain contains the frequently stated categories to summarize marketing and public policy goals in the public policy arena. These goals can be research interests: consumer protection, protection of viewed as the elements within the axiological domain. For marketers, policy and policymaking, and societal issues. example, policy researchers discuss the importance of The elements in the problem domain consist of market accessibility, fair balance, deception, disclosure, segments, and the relations reflect the behaviors of the miscomprehension, competition, and informed choice. various actors. The political, legal, cultural, and regula-

98 Marketing and Public Policy Conference Proceedings / 2009 tory factors provide a context to understand the phenom- determine, using the selected conceptual model, whether ena of interest. The regulatory environment serves two the goal results in the desired behavior. If the behavior distinct functions. From a static view, it provides a context does not reach the desired level, one could subsequently for understanding the phenomena of interest. When change create social change by developing a set of regulations is necessary (i.e., the current state is inadequate), the which would address the problem, such as imposing regulatory environment acts as a force outside the three guidelines for food labels or an educational campaign to domains described above. increase consumer literacy of nutrition labels.

As in the VNS, these three domains result in multiple Social change can be confronted in various ways: through orientations toward research. One orientation first links the goals and values we set as a society, through the the contents of the theoretical and axiological domain, and theories that allow researchers to better understand then applies this goal-model structure to examine a par- behaviors and how to influence them, and finally through ticular marketing-public policy phenomenon. For ex- examining the specific problems society is facing. While ample, suppose we believe full disclosure is a desirable each approach carries various strengths and weaknesses, goal, and we then apply the unified theory to examine the this framework emphasizes the importance of considering impact of full disclosure on making healthy food choices axiological, theoretical, and substantive concerns when given the current regulatory environment. We can exam- examining the role of marketing and public policy to ine the impact of this goal on healthy food choices and create social change. References are available upon request.

For further information contact: David Brinberg Virginia Tech 2019 Pamplin Blacksburg, VA 24061 Phone: 540.231.7639 Fax: 540.231.3076 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 99 KINGDON’S MULTIPLE STREAMS FRAMEWORK: A POLICY ANALYSIS TOOL FOR MARKETING AND SOCIETY SCHOLARSHIP

Jenny Mish, University of Utah, Salt Lake City

SUMMARY Case Illustration

This paper introduces and demonstrates the use of a policy To illustrate the application of MSF in a marketing and analysis tool, the Multiple Streams Framework (MSF) society context, consider the case of the Marine Steward- (Kingdon 2003), for use in marketing and society re- ship Council (MSC) eco-label (Gulbrandsen 2008): search. MSF lies between rational choice and constructivist models of policy analysis, emphasizing the empirical • Policy entrepreneurs. This eco-label was initiated discovery of theoretical insights that are embedded in in 1996 by activist-managers at the World Wide Fund complex qualitative data, much like grounded theory. for Nature (WWF, formerly World Wildlife Fund).

Multiple Streams Framework (MSF) • Problems stream. Indicators showed that over harvesting of oceans had reached disastrous propor- John Kingdon (2003) studied the process by which policy tions. Multilateral governance had failed to protect solutions enter policymakers’ consideration sets. Thus, the ocean’s capacity to renew fish stocks. Retailers rather than emphasizing decision-making, MSF focuses like Wal-Mart were becoming concerned about fu- on the formative stages of the policy process. MSF posits ture supplies. that issues are placed on the table by “policy entrepreneurs” who clarify and create meanings under conditions • Policies stream. Several years earlier, the WWF had of ambiguity, where there are “many ways of thinking been involved in the creation of the Forest Steward- about the same circumstances or phenomena” (Feldman ship Council, a forest products eco-label that was 1989, p. 5, in Zahariadis 2007, p. 66). An increase in gaining market traction. Organic and Fair Trade information does not reduce ambiguity. Instead, policy labels were seeing rapid market growth. entrepreneurs combine and strategically manipulate information from three critical “streams” to promote spe- • Politics stream. International laws of the sea had cific viewpoints and policy solutions (Kingdon 2003). codified compromises among complex stakeholder interests. The Forest Stewardship Council eco-label, The “problems stream” includes contextual information although founded by a multi-stakeholder coalition, such as indicators, events, feedback, and the number of quickly faced competition from an industry-spon- problems occupying policymaker attention. The “poli- sored eco-label. cies stream” consists of a “soup” of competing ideas about possible solutions, from which consideration sets • Coupling the streams. WWF policy entrepreneurs emerge in a process resembling natural selection, priori- teamed up with Unilever, a large fish purchaser, to tizing ideas that are feasible, ideas that fit with dominant jointly launch an eco-label specifying standards for values, and ideas likely to receive political support. The MSC-certified sustainable fisheries, consistent with “politics stream” consists of pressure group campaigns, international law. The WWF-Unilever partnership the national mood, and turnover among political adminis- was expected to bridge stakeholder groups and pro- trations. vide resources for brand building.

Successful policy entrepreneurs often wait for opportune The policy entrepreneurs in this case coupled the increas- moments or “policy windows” to “couple” the three ingly urgent problem of over harvesting with a “policy” streams and garner scarce policymaker attention. Partial solution in the form of an eco-label certification for couplings may be possible sooner, but “the complete sustainable fishery practices. This market-based solution joining of all three streams dramatically enhances the formalized common ground shared by environmentalists, odds that a subject will become firmly fixed on a decision seafood producers, and retailers. It took advantage of an agenda” (Kingdon 2003, p. 202). anti-regulation “mood” and bypassed the delays of inter-

100 Marketing and Public Policy Conference Proceedings / 2009 national policy. This coupling of (1) a clear problem Critics question whether the three streams are truly inde- definition and (2) alignment opportunities in the politics pendent, to which a proponent replied, “stream indepen- stream with (3) a “policy” solution that was both feasible dence is a conceptual device” (Zahariadis 2007, p. 81). He and consistent with dominant values exemplifies the role argued that while many analysis tools leave researchers of policy entrepreneurs. with policies in search of rationale, a strength of MSF is the ability to investigate early stages of the process. MSF Contributions can provide an interim data map from which theoretical propositions or relationships may be discerned. As with MSF provides a robust tool for research questions that other approaches to qualitative data analysis, researchers involve the conditions leading to specific policies. It can must be attentive to assumptions or preconceptions that be especially useful when a policy viewpoint emerges they might bring to the process, using peer checks and from complex, ambiguous societal contexts, such as a multiple coders as needed. macro-behavioral trend or a shift in values. In such situations, this framework offers a focusing structure for The Multiple Streams Framework is one of many policy sifting through and organizing volumes of data. The analysis tools that may be adopted by societal marketing timing of such processes can also be investigated in terms scholars studying complex and ambiguous phenomena. of Kingdon’s policy window concept.

For further information contact: Jenny Mish University of Utah 1645 E. Campus Center Drive Salt Lake City, UT 84112–9304 Phone: 801.520.0878 Fax: 801.581.3152 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 101 SUSTAINABILITY, OVERCONSUMPTION, AND CONSUMER WELL-BEING: IMPLICATIONS FOR MARKETING’S ROLE WITHIN A GLOBAL SOCIETY

Michael Bone, University of Alberta, Edmonton Adam Finn, University of Alberta, Edmonton

SUMMARY of life and the well-being of consumers. This research compiles consumption and quality of life data for a sample With rising attention to environmental and health issues of nations that vary on their level of wealth and develop- such as global warming, air pollution, obesity, and food ment from a variety of government and world organiza- and water safety, marketing practices have come under tion databases. While we believe that the increase in attack for playing an important role in the creation of an aggregate population growth is causing a harmful impact overconsuming society. The way that we, as humans, on the environment, we posit that the acceleration of have been consuming natural resources within developed global warming can be attributed more so to the overcon- nations has been unprecedented with regards to the his- sumption that exists in developed nations, rather than the tory of our civilization. Our pursuit of more goods more consumption patterns within developing nations. Here, conveniently has brought us to a point where carbon we define overconsumption as an excessive level of emissions are out of control and threaten the very place in consumption beyond a human’s necessary level of needs which we live. As a result, there is a question of whether and wants that satisfy their desired level of happiness and we are beginning to see an unwinding social fabric in well-being. In a sense, overconsumption and sustainability developed nations – created and perpetuated by the stress are conflicting paradigms in that a society that lives in placed both on the planet, and on ourselves, from the excess is unsustainable into the future. This is important unquenchable thirst for economic growth and a superior to study as some argue that this push for a higher quality quality of life. of life and the associated increase in consumption of goods and services has in turn led to people in developed Carbon dioxide (Co2), a natural element in the Earth’s nations actually being worse off, as we see declining atmosphere, can be detrimental to the health of the planet levels of health, happiness and well-being (Anielski 2007). and the way in which its systems operate when it is out-of- line with Earth’s natural capacity levels. In the last 50–60 Another function for this research is to expand upon years, we have seen Co2 levels exponentially rise. While extant sustainable marketing literature, which to our knowl- some could argue that increases in emissions are due edge, has not used empirical analysis to bridge the gap mainly to increases in global population, our initial analy- between unsustainable consumption (or overconsump- sis of raw emissions data does not find support for this tion) and the marketing methods currently in practice. We argument. Increases to global population clearly plays a believe that this research will add to the discussion con- role in the current environmental problem; however, what cerning whether or not classical micromarketing objec- marketing impacts is consumer overconsumption and tives (growth) are actually in conflict with macromarketing research could help answer how and why certain seg- goals (quality of life and well-being), which if true, ments of consumers contribute to the Co2 emissions creates a situation where current marketing practices have problem. Global warming is likely occurring because an inherent drive toward unsustainability (Van Dam and there are more humans, but it could be accelerating Apeldoorn 1996). because consumers in developed nations are overconsuming. Critics of high consumption lifestyles suggest To address the issue of overconsumption across nations, this way of life contributes indirectly to environmental we analyze consumption-based variables in order to gain degradation (Lovins, Lovins, and Hawken 1999), de- better insight into the reasons why we see an increase in clines in life systems (Hawken 1993), and lower happi- consumption-based emissions and study what types of ness and well-being (Anielski 2007). nations and societies are the major contributors of these emissions. For this reason, the unit of analysis for this Research Question, Purpose, and Methodologies research are the countries of the world. We have constructed a data set characterizing human consumption, The purpose of this research is to empirically investigate quality of life, and marketing impact variables for a if increases in consumption are, in fact, decreasing quality sample of developed and developing nations. We theorize

102 Marketing and Public Policy Conference Proceedings / 2009 a model of overconsumption at the national level – one in Anielski 2007; Mayer 2008) but not fully developed which the consumption of resources can be attributed to within the marketing literature (exceptions: Burroughs the role in which marketing plays in a country, and at what and Rindfleisch 2002; Sirgy and Lee 2006). We believe level it impacts consumer life through its functional ac- that research in the domain of consumption and consumer tivities (e.g., new product development, aggregate adver- well-being is imperative to the future of marketing, as it tising). We posit that higher levels of marketing activities studies what really matters to consumers. To our knowl- will be correlated to Co2 emissions and decreases to edge, this is the first study to quantitatively model how the consumer well-being over time. This research will be role of marketing and consumer consumption can be conducted using appropriate methods of analysis with linked to our happiness and well-being. regards to the data and the research question. Initial empirical results using principal components analysis It is important that we recognize that marketing is inextri- find support for the argument that the increase in Co2 cably linked to the successes and failures of growing emissions is likely due to increases in consumption rates toward a sustainable and healthy economy. Marketing of consumers in developed nations (stable populations), utilizes its functions (new product development, advertis- not developing nations demonstrating population growth. ing, etc.) to shape our commercial reality and bring goods and services to the marketplace which guide us in our Theoretically, this research is important as it extends work socioeconomic class consumption habits and contribute in sustainable marketing and empirically studies how to our quality of life. As a very powerful tool in a free marketing’s role within society impacts our environment market economy, marketing has the opportunity to rede- and well-being. More importantly, this research contrib- fine how we consume as a society and possesses the utes to the measurement of consumer well-being in rela- opportunity to take us leaps and bounds toward a sustain- tion to consumption factors and quality of life – an area able and healthy future for our planet and ourselves. with important research outside of marketing literature References are available upon request. (Osberg and Sharpe 2002; Nader 2003; Bergheim 2006;

For further information contact: Michael Bone School of Business University of Alberta Edmonton, Alberta, Canada T6G 2R6 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 103 CONSEQUENCES OF LEGISLATING PACKAGING BEHAVIOR: GERMANY’S GREEN DOT PROGRAM

Lukas Jones, University of Oregon, Eugene Lynn R. Kahle, University of Oregon, Eugene Scott Owen, University of Oregon, Eugene

SUMMARY The GD is based on a “polluter pays” principle that many other nations have experimented with in various ways Many companies take environmental issues as one of their (Fuller 1999). The system allows manufacturers to pay primary brand positioning objectives and target a market fees, based on the amount of packaging they produce, into of environmentally conscious consumers. The majority, an industry-led organization, which then collects all of the however, adopt more of a moderate stance, taking those packaging waste on behalf of industry. Packaging from actions that are both environmentally friendly and eco- consumer products constitutes a surprisingly large per- nomically responsible, that is, actions often referred to as centage of overall post-consumer solid waste in many the “low hanging fruit” of this arena. Unless government countries, including Germany. In the GD program, pack- intervenes and defines higher minimum standards, many aging is collected in both household waste containers and companies are reluctant to leap into costly eco-friendly collection bins that are located in high traffic areas (Fishbein programs. 1994).

In Germany, the government has taken an interesting The program’s goals are to encourage both innovative stance by defining corporations’ environmental responsi- packaging design, leading to less overall waste, and a bilities regarding such topics as packaging. Germany was higher recycling rate. If the manufacturers see these the first European nation to have a governmentally con- resource recovery costs realized on the bottom line, then trolled environmental certification label, which was known they will be motivated to find ways to reduce the amount as the Blue Angel seal (Fuller 1999). It comes as no of unnecessary packaging that they produce (DSD Website, surprise then that a program such as Der Grüne Punkt 2008). Whether this happens, at least the level of material (The Green Dot) would come out of the German legal being recycled will increase and attitudes about environ- system. The Green Dot (GD) is a program whereby mental responsibility will perhaps change. The GD pro- manufacturers are required by law to take back the post- gram has had great success since its inception in 1993, and consumer packaging waste from the products that they has, in more recent years, been adopted in some form by produce and recycle it (Fuller 1999; Fishbein 1994). In most other European Union countries. this way, the manufacturers take responsibility for the disposal costs of the packaging waste that they produce. In With the precarious nature of our planet’s ecosystems, the U.S. we require recycling packaging only in some much will need to be done in the coming years to slow, and industries, such as soft drink deposits regulated by “bottle ideally reverse, the damage humanity consumption has bills” (Kahle and Beatty 1987). traditionally caused to the environment. We are at a point in time where “the tragedy of the commons” has become The term packaging waste, in the German use, means reality, and everyone is grabbing their piece of the pie to post-consumer packaging waste that is not contaminated the detriment of collective humanity. The business realm by hazardous components (Fishbein 1994). The German can be a key component in changing this paradigm, program does not include waste involved with the cre- because within business lies the capital and power neces- ation of packaging materials, such as factory pollution, or sary to enact real change. This change is something that other waste from the manufacturing processes. It also needs to be governmentally mandated, as was done in does not deal with packaging that may have contacted any Germany with the GD Program or in the US with topics type of hazardous waste. The packaging that was to be such as leaded gasoline, so that businesses compete with reduced under the program was solely that which con- each other on a level playing field where no business sumer products need for transport, display, and protection obtains an unfair cost advantage by lowering standards. of the product itself. References are available upon request.

104 Marketing and Public Policy Conference Proceedings / 2009 For further information contact: Lynn R. Kahle University of Oregon 1208 University of Oregon Eugene, OR 97403–1208 Phone: 541.346.3373 Fax: 541.346.3341 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 105 SOCIAL VALUES SEGMENTATION IN MARKETING RESIDENTIAL SOLAR POWER INSTALLATIONS

Scott Owen, University of Oregon, Eugene Lynn R. Kahle, University of Oregon, Eugene

SUMMARY make consumers feel like they are making a difference in the world in energy usage. Increased use of residential solar energy could decrease reliance on foreign oil, save money in the long run, and Approximately sixteen percent of consumers value a help address the root causes of global climate change. As sense of accomplishment as their most important value a discipline Marketing could do more to facilitate the (Kahle et al. 1986). A sense of accomplishment could transition of greater usage of solar energy. result from being an innovator in the community, which could increase awareness of solar energy. This sense of Marketing research has shown that most people make accomplishment for alternative energy includes helping purchase decisions based on attributes, and most market- the environment or achieving a desirable goal. Solar ing of residential solar energy systems emphasizes these energy creates a sense of accomplishment for early adopt- attributes. Research has shown, however, that attributes ers, as using alternative energy has the end goal of becom- are meaningful to the extent that they connect to conse- ing a partially self-sufficient energy consumer. Solar quences, and consequences are meaningful to the extent energy could be positioned as a goal or accomplishment that they connect to social values via means-end chains. that is worth achieving to target these consumers. Marketing communication that explicitly links attributes to consequences and consequences to social values is Self fulfillment could be seen as a natural extension of the more effective than marketing communication that does sense of accomplishment phenomenon and is rated as the not provide these links (Kahle and Xie 2008; Limon, most important value by approximately seven percent of Kahle, and Orth 2009). the population (Kahle et al. 1986). These consumers are fulfilled by reducing carbon emissions and reliance on Different consumers have different social values. Thus, foreign energy, and by creating electricity off the grid. one marketing strategy is to segment and reach out to Marketing effectively to these consumers can be accom- consumers with different value-based marketing commu- plished by tapping into the idea that self provided energy nication based on each group’s core social values. For leads to self-fulfillment. example, communications with consumers who value self-respect could focus on the self-reliant independence Being well respected has been shown to be the most associated with a self-contained solar energy system that important value for approximately six percent of the does not depend on large-scale control from others. For population. Being well respected is an important value for consumers who primarily value a sense of accomplish- those who seek a sense of belonging in their peer group. ment, a more effective communication strategy may focus This is contrasted to valuing self-respect, which is an on the technological achievements associated with the intrinsic value. Marketers could easily target this group of system. For consumers who primarily value security, consumers by showing that many people in a given separation from politically risky sources of energy may community utilize solar energy. Additionally, the poten- dominate the communications. tial consumer must adopt solar energy in order to be accepted by the surrounding community. Marketers can Self-respect is the value consumers desire the most, with position solar panels as a status symbol in order to appeal approximately twenty-three percent of the population to these consumers. placing the highest value on self-respect (Kahle and Kennedy 1988; Kahle et al. 1986). The value of self- Approximately seventeen percent of the population re- respect is applicable to solar energy in that consumers gards security as the most important value. Consumers want to respect the items they own as well as the attributes who use solar energy are at least somewhat self-reliant in of those items. Marketers can target individuals who value obtaining their own energy by producing a portion of it. self-respect the most by focusing on the intrinsic attributes By producing energy, consumers become the manufac- of solar energy, such as the ability of alternative energy to turer of the product, ensuring that they are able to control

106 Marketing and Public Policy Conference Proceedings / 2009 the cost and consumption and security of the energy. Fun and enjoyment in life is the most important value to Marketers position solar power as a way to control the about seven percent of the population (Kahle et al. 1986). environment, which will attract consumers who value Marketers can position solar energy as a new, exciting security. product to successfully target the consumers who value fun and enjoyment. The value of security can also be expanded to a macro- level. Current policy shift is occurring where consumers Each of the eight values discussed above provide an are demanding domestically created energy. This policy insight for marketers in segmenting and targeting the shifts from previous energy policy, which obtained en- population. We need more research conducted on the idea ergy from foreign sources. As this policy shift increases, of marketing to consumers in the area of alternative solar power consumers will be seen as innovators in energy and solar panels. The List of Values (LOV) pro- domestic energy production and consumption, which will vides marketers a tool in targeting different segments of assist marketers in targeting this segment. the market to increase sales of residential solar panels. References are available upon request.

For further information contact: Lynn R. Kahle University of Oregon 1208 University of Oregon Eugene, OR 97403–1208 Phone: 541.346.3373 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 107 THE USEFULNESS OF SURVEYS IN ANTITRUST

Carol Miu, Economists Incorporated, Washington, D.C.

SUMMARY of whom ultimately completed the interview. The incidence rate of the Whole Foods survey was 5.9 percent, and Introduction the response rate was 6.4 percent.

Surveys can be useful tools in antitrust to provide valuable Incorrect definition of the population of interest, non- insights on a myriad of merger analysis topics. Consumer random selection of survey locations, non-random sam- surveys can be a source of a crucial data on market pling of consumers, low response rate, and inaccurate definition, market segments, and behavioral intentions interpretation of the data greatly increase the likelihood such as response to price increases. Likewise, surveys of that a survey may be unusable in an antitrust setting. current and potential competitors can yield information on market definition, competitive response, and barriers Sampling Methodology for Surveys Used in Antitrust: to entry. However, a review of the antitrust literature Designing and Conducting a Better Survey reveals that surveys have been used on a very limited basis to study slotting fees and household switching among In contrast to traditional marketing surveys, a question- banking institutions (Wilkie et al. 2002; Kiser 2002). naire for the purpose of antitrust analysis must be designed to specifically address issues such as market defi- Given the power of surveys to shed light on subjects nition and consumer responses to a hypothetical SSNIP. important in antitrust, it is likely that the role of surveys in However, designing survey questions to avoid respondent antitrust can be greatly expanded. However, it is difficult confusion and use of potentially biasing heuristics is a for surveys to pass muster in litigation, due to potential tricky task. It is important to translate regulatory verbiage issues such as sample bias, non-response bias, improper into concepts relevant to consumers, without losing the data analysis, and incorrect deduction of results. The aim link between survey questions. The core question that of this research is to examine past survey errors in antitrust needs to be answered is: “Which products across different litigation, and to create and conduct an improved survey competitors do consumers consider substitutes, and how that will be considered both useful and reliable in merger do the levels of substitutability change with changes in analysis. factors such as product attributes (price, features, size, etc.), customer loyalty, and switching costs?” Sampling Methodology for Surveys Used in Antitrust: An Example of Past Errors We employ proper sampling methodology to develop, pre-test, and pilot test a consumer survey to explore In FTC v. Whole Foods, the U.S. District Court decided relevant issues in merger analysis. Survey questions are against giving “any weight or consideration” to the survey designed to examine the core question of substitutability, conducted by the defendants’ expert because the method- and to gather data useful to geographic and product ological flaws of the survey rendered the data and results market definition. Questions address consumers’ percep- unreliable. A careful examination of the Whole Foods tions of the market and relevant firms, switching inten- survey reveals that a better sampling plan may have tions in the event of product attribute changes (i.e., a eliminated many of the criticisms. In identifying the hypothetical SSNIP), and demographic information. Ret- population of interest, the defendants’ survey expert mis- rospective testing is utilized to determine that consumers takenly defined two populations, “Frequent” and “Cusp” understood the questions and answered them accurately. shoppers at Whole Foods or Wild Oats, which were then quota sampled such that each of the two groups comprised Research has shown that respondents who are asked to 50 percent of the survey response. Results were then make difficult mathematical calculations often resort to erroneously generalized to all current Whole Foods and guestimation instead of doing the math (Converse and Wild Oats customers, although giving “Frequent” and Presser 1990; Tourangeau et al. 2000). This poses a “Cusp” customer groups equal weight is very unlikely a difficulty when trying to gather data about customers’ representative sample of the population of interest. Fur- likely reactions in a 5 percent SSNIP scenario. To reduce ther, a total of 427,397 phone calls were placed to identify the need for calculations the survey can be administered 25,011 Whole Foods or Wild Oats customers, only 1,607 using a letter-paper-sized electronic device similar to a

108 Marketing and Public Policy Conference Proceedings / 2009 tablet PC. The device is able to make calculations and graphics, and presence of the retailer. Days of the week store pricing information on hundreds of products, quickly and times of the day for data collection are varied. Initial asking each customer whether he or she purchases a results indicate that, with proper sampling methodology particular product while displaying a photo and size and survey execution, it is possible to collect data to be information to ensure that the customer has recognized used in antitrust analysis from a representative sample of the correct product. If the answer is “yes,” the device consumers. displays the current price of the product and asks the customer confirm that this is the price that he or she pays. Although the Whole Foods survey was deemed unreli- The device then displays a price that is 5 percent greater able, a properly conducted survey could be valuable in than the current price and asks the customer if he or she antitrust litigation. Before attempting to conduct a survey, would still purchase this product at the higher price. researchers must understand how to how to design surveys that will be defensible and to identify issues such as We define the population of interest as all current custom- biased sampling procedures, problems with questionnaire ers of a particular brick and mortar retailer, and choose a design and administration, omitted control variables, and choice-based store-intercept sampling frame. Stores at improper interpretation of results (Dillman 2007). Based which data are collected are selected in a stratified random on our results, we find that the role of surveys in antitrust fashion based on geographic region, consumer demo- can be greatly expanded.

For further information contact: Carol Miu Economists Incorporated 1200 New Hampshire Avenue, NW Suite 400 Washington, D.C. 20036 Phone: 202.833.5257 Fax: 202.296.7138 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 109 DETERMINANTS OF PROPERTY RIGHTS: MORE EMPIRICAL EVIDENCE

Thomas G. Costello, Eastern Illinois University, Charleston Ayse Olcay Costello, Eastern Illinois University, Charleston

SUMMARY costs under exclusivity; (3) costs of reduced investment created by non-exclusivity (includes reallocation and Property rights related issues have always played an valuation costs); (4) exchange value under exclusivity; important role in determining economic outcomes, whether and (5) social costs of exclusivity. they be issues related to intangible resources (e.g., intellectual property, patents) or tangible resources (e.g., build- In a new study, to resolve some of the puzzles that remain ings, equipment, oil fields). Therefore, how rights are after the first empirical tests of the framework, we tested allocated over resources has always been an important the hypotheses derived from the framework proposed in question, not only for public policy makers, but also Costello and Costello (2005). As a partial test of the managers and consumers alike. For example, a property framework, we hypothesized that the level of definition of rights allocation decision that has been puzzling high property rights over resources depend on (1) social costs level executives is whether companies can enforce their of exclusivity; (2) exchange value under exclusivity; and rights without alienating their consumers, when those (3) costs of reduced investment created by non-exclusiv- consumers, such as the ones that download music or ity (includes reallocation and valuation costs). To test our movies without paying for it, actually believe that they hypotheses we used an experiment in which student have a right to the resources that they are taking; and that subjects were asked to read a short case and make level of they are doing nothing wrong. A good example of execu- definition of property rights decisions. To be specific, tives trying to deal with this difficult predicament is student subjects were asked to read a case about a hypo- Microsoft’s decision to only create relatively minor ser- thetical pharmaceutical company and a drug that the vice disruptions for consumers using pirated software company had developed to cure lung cancer. The case had without actually rendering pirated software completely eight versions/ conditions in which three variables were useless. Nevertheless, Chinese consumers, facing manipulated; (1) the drug’s cost representing the social Microsoft’s relatively minor enforcement action, still costs created for patients and society (two levels, $500; ended up taking offense with the company’s efforts to $50,000 per treatment); (2) the tax rebate that students will discourage consumers from using pirated software de- get as a portion of tax revenue collected from the pharma- spite the company’s undisputed legal rights (Chao and Ye ceutical company to represent personal gain (value) that 2008). the subjects can expect from the company’s profits (two levels, $500; $50,000); and (3) the company’s threatened Given the importance of decisions made regarding levels investment response to the potential usurping of its patent of definition of rights over resources and the allocation of by the government (two levels, stop all future investment, those rights, researchers are beginning to develop theories or do not stop all future investment). The students, ran- that can help both public policy makers and corporate domly assigned to one of the versions of the case, were executives make better decisions regarding those rights asked to read the case and then to decide whether or not (e.g., Costello and Costello 2005). The theoretical frame- they wanted to recommend to their Congressional Repre- work developed by Costello and Costello (2005) had been sentative to revoke the company’s patent. In total, 151 especially useful as it gives decision makers a practical student subjects, in their junior or senior year, participated tool that can be used in making actual decisions, as in the study. Student subjects were deemed acceptable, opposed to the traditional way of looking at property since students routinely make intellectual property rights rights decisions as those where externalities are internal- decisions, such as choosing whether or not to download ized as benefits of internalization exceed its costs (Demsetz movies and music for free. Furthermore, students are 1964). In their framework, Costello and Costello (2005) becoming increasingly more politically active thus im- point out that the level of definition of rights over a pacting property rights legislation. resource depend on five specific cost-benefit components; and these are: (1) capture costs and rent dissipation Hypotheses were tested using a multicategory logit model, created by non-exclusivity; (2) exchange and policing which is a model that refers to all pairs of categories of the

110 Marketing and Public Policy Conference Proceedings / 2009 dependent variable simultaneously and it describes the personal gain from the sales of the drug was high, and the odds of response in one category versus another given the company said that it would stop all future research in the levels of the predictor variables (Agresti 1996). The case of a patent revocation, 43.5 percent of all young, dependent variable y was a categorical variable that rep- female, democratically (i.e., political party wise) inclined resented the level of definition of property rights decision subjects had a desire to revoke the company’s patent. (recommendation of revoke versus don’t revoke). The When social costs were increased, the percentage of the categorical and continuous predictor variables represented same sub-section of subjects that wanted to revoke the the personal value that subjects realized from the sale of company’s patent rose to 72 percent. Finally, if the com- the drug in the form of tax rebates, the company’s poten- pany stopped threatening to cease all future research, 85 tial investment decision in case of a patent revocation, the percent of the same sub-section wanted to take the social costs created by the price of the drug, the age of the company’s property rights away. decision maker, the gender of the decision maker, and the political affiliation of the decision maker. The variables The results of the study suggest that when social costs of that were found to be significant in determining whether assigning property rights are deemed to be high, the or not the study subjects recommended to revoke the instinct of decision makers is to balk at assigning property pharmaceutical company’s patents were SOCIAL COSTS rights regardless of the personal value to them. Further- (Significance: .000) and INVESTMENT (Significance: more, if the decision makers think that the companies .011). The rest of the variables were not deemed to be involved will keep on investing in the creation of the significant by the model; however, the overall fit of the resource in question, their willingness to assign property proposed model was good (Pearson Chi-Square: 61.599; rights in the face of possible social costs goes down Df. 35; Sig.: .004). Furthermore, a perusal of the model dramatically. These results highlight the importance of parameters was informative. For example, if social costs managing a company’s image in its efforts to enforce its of making the lung cancer medicine exclusive were low, property rights. References are available upon request.

For further information contact: Thomas G. Costello School of Business Eastern Illinois University 600 Lincoln Avenue Charleston, IL 61920 Phone: 217.581.6922 Fax: 217.581.6247 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 111 HEADING TOWARD 1984? AN EXPLORATORY STUDY OF BUSINESS MODELS AND REGULATORY FRAMEWORKS FOR INTERACTIVE ADVERTISING IN THE U.S. AND THE E.U.

Olesya Venger, University of Pennsylvania, Philadelphia

SUMMARY popular advertising network within a virtual world of Second Life were analyzed (Shehan 2008; Bach 2008). Until recently, regulations in the field of interactive adver- By critically evaluating mechanisms of interactive advertising have not received much attention in the U.S. and the tising applications in social networks/virtual environ- E.U. However, the latest push in favor of ad networks ments as well as their regulations, this paper points out establishing self-regulatory policies shed light on the various threats they pose to the privacy of users. The practice of interactive advertising and the problems con- regulations and the leading business mechanisms for nected to it (Luft 2008; Lal Bhasin 2008). Given that interactive advertising in both societies differed: strict interactive advertising online and privacy issues associ- regulations for interactive advertising in the E.U. resulted ated with this practice are considered to be sensitive topics in the emergence of behavioral selling as the main busi- within both societies, it is important to compare interac- ness model of interactive advertising in Europe, while the tive advertising regulations in the U.S. and the E.U. lack of stringent controls in the U.S. has increased the prevalence of behavioral targeting. In general, when talk- Insofar as interactive advertising within social networks ing about interactive advertising in general and and online games within virtual environments are heavily advergaming in particular, American consumers are not used by Americans and Europeans alike, the following protected against those who would use their personal data questions were posed by this study (1) What differences without their knowledge and consent. exist between major business mechanisms for interactive advertising employed in the U.S. and the E.U.? (2) What Protection is also lacking in the E.U. against unscrupulous are the business models for interactive advertising in the use of personal data within online video games/social U.S. and the E.U.? (3) What are the differences in the networks; however, the European Internet regulations regulatory frameworks for interactive advertising in the have search engine advertising regulations enforced and U.S. and the E.U.? A sample of online advertising appli- thereby offer greater protection to consumers online than cations for the analysis was drawn from the most popular American regulations. Other than several guidance re- social networks, MySpace and Facebook, which have ports for self-regulatory efforts issued by the Interactive 58.5 million and 18.09 million users respectively (ZDNet Advertising Bureau and the FTC (FTC 2000), there are no Research 2007). noteworthy regulations that would address the privacy issues associated with interactive advertising and The business mechanisms for online games developed by advergaming in the U.S. such companies as Microsoft and Nestle as well as a

REFERENCES (accessed July 21, 2008), [available at http:// www.bloggingstocks.com/2007/05/19/why- Article 29 Data Protection Working Party (2008), “Opin- microsoft-bought-aquantive-broadband-users-up-3- ion on Data Protection Issues Related to Search million-t/]. Engines,” April 4 (accessed June 16, 2008), [avail- COPPA (2000), “Children’s Online Privacy Protection able at ec.europa.eu/justice_home/fsj/privacy/docs/ Act,” April 21 (accessed July 18, 2008), [available at wpdocs/2008/wp148_en.pdf]. http://www.coppa.org/comply.htm]. Barlow, Tom (2007), “Why Microsoft Bought Aquantive: Chester, Jeff (2007), Digital Destiny: New Media and the Broadband Users up 3 Million this Year,” May 19 Future of Democracy. New York: New Press. Dis-

112 Marketing and Public Policy Conference Proceedings / 2009 tributed by W.W. Norton & Co. Internet Marketing and Advertising, 4 (2/3), 213–40. Cho, Michael, Andrew Iskandar, Sanjay Kidambi, Steve Luft, Oliver (2008), “Social Networks Need to Do More Shepherd, and Cuong Do Tuan Vuong (2005), to Raise Awareness about Misuse of Personal Infor- “Search-Based Online Advertising,” October (ac- mation, Says PCC Director,” May 6 (accessed July cessed October 18, 2008), [available at http:// 18, 2008), [available at http://www.journalism.co.uk/ people.ischool.berkeley.edu/~hal/Courses/ staging/2/articles/531706.php]. StratTech07/Tech/Tech05/C-report.doc]. Moore, Elizabeth S. (2006), “Analysis of Online Food eMarketer (2008), “Let the Video Game Ads Advertising Targeting Children,” (accessed July 18, Begin!” eMarketer, March 3 (accessed July 15, 2008), 2007), [available at http://www.kff.org/entmedia/ [available at http://www.emarketer.com/ entmedia071906pkg.cfm]. Article.aspx?id=1005990&src=article1_home]. Nicole, Kristen (2008), “Playfish Creates Addictive Fox Interactive Media (2008), “MySpace Targeting Ca- Facebook Games [The Startup Review],” (accessed pabilities: Reaching the Right Consumer in a 2.0 July 18, 2007), [available at http://mashable.com/ World,” Fox Interactive Media, (February). 2008/06/27/playfish]. FTC (2000), “Advertising and Marketing on the Internet: Schroeder, Stan (2007), “Google’s In-Game Advertising Rules of the Road,” (accessed September 21, 2008), System Coming Soon,” November 7 (accessed July [available at http://www.ftc.gov/bcp/conline/pubs/ 18, 2008), [available at http://mashable.com/2007/ buspubs/ruleroad.shtm]. 11/07/googles-in-game-advertising-system-coming- Hemp, Paul (2006), “Avatar-Based marketing,” Harvard soon]. Business Review, 84 (6), (June), 48–57. Shehan, Michael (2008), “Level One: In-Game Advertis- Internet Advertising Bureau (2008), “Internet Advertis- ing Success,” June 23 (accessed July 18, 2008), ing Revenues Top $21 Billion in ’07, Reaching [available at http://blogs.mediapost.com/ Record High,” May 15 (accessed June 15, 2008), video_insider/?p=185]. [available at http://www.iab.net/about_the_iab/ ZDNet Research (2007), “Top Social Networking Sites in recent_press_releases/press_release_archive/ September 2007,” October 24 (accessed September press_release/299609?gko=1874c]. 18, 2008), [available at http://blogs.zdnet.com/ Internet Advertising Bureau (2008a), “Game Players ITFacts/?p=13360]. Opinions of In-Game Advertising,” (accessed June ______(n.d.), “Re Telecommunication Regulations 16, 2008), [available at http://www.iabuk.net/en/1/ Applicable to Search Engines? Emerging Trends- ingameadvertising.html]. Search Engine Regulations,” (accessed September Lal Bhasin, Madan (2008), “Guarding Privacy on the 18, 2008), [available at http://www.ibls.com/ Internet: Privacy Policy, Government Regulations ebooks.aspx?id=105&s=view]. Other references are and Technology Solutions,” International Journal of available upon request.

For further information contact: Olesya Venger Annenberg School for Communication University of Pennsylvania 3620 Walnut St, Suite 200 Philadelphia, PA 19104 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 113 THE USE OF PUBLIC RELATIONS TO PROMOTE A PLAN FOR REBUILDING A CITY AFTER A DISASTER

Owen Kulemeka, University of Illinois at Urbana-Champaign

SUMMARY The findings reveal that the Rockefeller Foundation’s campaign was what Grunig (2002) labeled as a two-way This paper describes a study that examined how the symmetrical public relations campaign. Grunig, an archi- Rockefeller Foundation used public relations to build tect of public relations theory, argues that when an orga- support for a plan on how to rebuild post-Hurricane nization communicates with publics using a public rela- Katrina New Orleans. Few cities plan on how they will tions campaign, four types of campaigns are possible: rebuild after a disaster. Most rebuilding planning is done press agentry, public information, two-way asymmetri- after a disaster. This often results in conflict as different cal, and two-way symmetrical. groups disagree on how to best rebuild a devastated city (Olshansky 2005). In a press agentry campaign, an organization relies on propaganda to persuade publics (Grunig 1992). An ex- New Orleans exemplifies post-disaster rebuilding con- ample of press agentry is an organization that launches a flict. Following Katrina, the city’s planning department campaign to promote a rebuilding plan that was devel- lacked the staff to draft a rebuilding plan. Hence, the city oped without consulting residents. Although the rebuild- languished for months without a plan. This slowed down ing plan benefits only a few, the organization uses mass recovery because donors and investors were wary of media to spread incomplete information arguing that the investing money in a city lacking a plan (Nossiter 2006). plan will benefit everyone. Frustrated by the delay, various groups embarked on efforts to draft rebuilding plans for the city. The result was In a public information campaign, an organization uses confusion. Six groups produced different rebuilding plans truthful information to persuade publics (Grunig 1992). that addressed similar issues. When released, several of An example is an organization that launches a campaign these plans were criticized as products of backroom deals to promote a rebuilding plan without consulting residents aimed at helping the city’s rich (Quinones 2006). regarding what they want. However, the organization refrains from outright lying regarding the benefits of the By early 2006, it was clear that the delay in adopting an plan. This approach is still problematic because those official rebuilding plan was severely harming the city. To likely to be affected are not consulted. address this confusion, the Rockefeller Foundation convinced the different groups working on rebuilding plans to In a two-way asymmetric campaign, an organization uses consolidate their efforts into one. The result was the research to find better ways of manipulating publics Unified New Orleans Plan (UNOP), an effort aimed at (Grunig 1992). An example is an organization that holds bridging the bitter disputes on how the city should be hearings to discuss how to rebuild. The organization uses rebuilt. the meetings to solicit feedback that it can use to push its own agenda. When faced with criticism, the organization This study examined the public relations tactics that the only makes minor changes but refuses to budge on major foundation used to build support for the UNOP plan. issues that affect publics negatively. Public relations plays an important role in post-disaster rebuilding debates. Groups with competing ideas on how In a two-way symmetric campaign, an organization makes to rebuild use press releases, meetings, and other tactics to a genuine effort to engage in dialogue in order to find a promote their rebuilding agendas. In this paper, the re- balance between the organization’s needs and that of its searcher sought to answer two questions. First, what publics (Grunig 2002). An example is an organization that public relations tactics did the Rockefeller Foundation use holds hearings where input from publics is seriously to promote the UNOP plan? Second, how effective were considered. When faced with criticism, the organization these tactics? Data was collected through participant makes substantive changes that balance its own interests observation and analysis of media coverage. with those of various publics.

114 Marketing and Public Policy Conference Proceedings / 2009 The Rockefeller Foundation’s public relations campaign rebuild their neighborhoods. These plans would be voted was designed to be two-way symmetrical. The foundation on by the whole city hence enabling all residents to have did not simply seek to promote its own plan. Instead, it a say in rebuilding. sought to ensure that all the city’s residents had a voice in shaping a rebuilding plan. Although the foundation’s campaign was designed to be two-way symmetrical, it did not succeed in its goal of To accomplish the goal of gaining input from all residents, ensuring that all residents had a voice. Public meetings the foundation organized a series of public meetings that were dominated by middle income interest groups. Many worked as follows. New Orleans was divided into 13 low-income individuals did not attend because they felt districts. In each district, the foundation provided meeting powerless, lacked transportation, or were busy with work. space where residents could develop plans on how to References are available upon request.

For further information contact: Owen Kulemeka University of Illinois at Urbana-Champaign 228 Gregory Hall 810 S. Wright Street Urbana, IL 61801 Phone: 217.244.6145 Fax: 217.244.7695 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 115 CONTEMPORARY CONSUMERISM: REVEALING 21st CENTURY CONCERNS

Kristen L. Walker, California State University Northridge

SUMMARY Kassarjian 1971; Barksdale et al. 1971) but their findings focus on issues such as personality, socio-economic class, Consumerism is a general concern for the quality of the age, education, and lack a broad understanding of how all environment, rights of individuals, protection of privacy, of these variables may influence a person’s “consumerist” state of the economy and impersonality of big business views. These scales and studies do not examine or include and big government (Bourgeois and Barnes 1979). The current issues such as online consumer purchasing habits U.S. currently faces many of these issues. For instance, or a broad view of consumer concerns and attitudes the notoriety and press about product recalls in 2007 toward manufacturing, government, business and the highlighted the need for attention on the quality of prod- environment. ucts and manufacturing. Global warming and the trend for sustainability and for companies to go green demonstrate Klein’s scale and other measurements (i.e., Consumer a focus on the quality of the environment. The recent Sentiment Index 1986) focus on the view of consumerism government bailouts and stock market plunge(s) in 2009 as business-oriented, or primarily concerned with protect- also indicate that consumers are uneasy about issues of big ing consumers from abuses by business. In reality, con- government and business. In sum, the current economic sumerism has different facets (a) protection against clear- and social climate in the U.S. requires researchers to cut abuse, (b) provision of adequate information to con- revisit concepts and perceptions of consumerism. sumers, and (c) protection of consumers against themselves and other consumers (Day and Aaker 1970). Con- Concern for these issues is not new and Herrmann (1970, sumerism and those who have consumerist attitudes are p. 1) described patterns in these views of consumerism. concerned with both the physical environment and the He discusses three “periods of consumerism,” and how economic environment. The primary outcome of this the consumerist view developed as a result of consumer research is the creation of a consumerism scale that is unrest, or “rising public standards of business conduct and current and one that reflects the trends and issues facing social responsibility.” Herrmann (1970) explained that today’s consumers. the first period was in the early 1900s, the next in the 1930s and third in the late 1960s to early 70s. Each of these Using scaling procedures developed and tested by several periods occurred during “turning points” (Hollander et al. researchers (Netemeyer, Bearden, and Sharma 2003; Flynn 2005) of serious economic and social dislocation. Al- and Goldsmith 1999) this research reports on the process though we should avoid making assumptions about the of a new and revised scale on consumerism, ultimately past and recognize that social phenomena are in constant measuring views on manufacturing, government, busi- flux (Fullerton 1986), the current economic and social ness and the environment. Much of the research on con- climate in the U.S. indicate that we may be on the cusp or sumerism, or other subjects for that matter, are based on perhaps in the midst of another turning point of consum- single item or multi-item ad hoc instruments that were erism. developed solely for use in the purpose of that project and they are not necessarily validated in other studies. On the The last time a scale or tool for measuring consumerism other hand, a scale that is validated is more reliable and views was developed was over 25 years ago, in 1982. In useful in operationalizing latent constructs – in this case his study, Klein (1982) developed a consumerism scale to views on consumerism (Flynn and Goldsmith 1999). measure consumerism attitudes of business entities and Creating a scale will allow determination as to whether a consumers, using U.S. and Swedish subjects. This scale person is a consumerist; someone who supports consumer has been used in consumerism studies (Darley and Johnson issues to protect the “well-being” of consumers (Bour- 1993; Lysonski, Durvasula, and Watson 2003) yet there geois and Barnes 1979). This scale will encompass a has been no formal update to the scale or new scale person’s characteristics, purchasing habits and their views developed in almost a generation. Earlier studies on on manufacturing, government, business and the environ- consumerism also have created survey instruments spe- ment, revealing 21st century concerns. A list of references cifically for their research (Webster 1975; Coulson 1971; is available upon request.

116 Marketing and Public Policy Conference Proceedings / 2009 For further information contact: Kristen L. Walker California State University Northridge 18111 Nordhoff Street Northridge, CA 91330–8377 Phone: 818.677.2040 Fax: 818.677.7669 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 117 WHY DID I EAT THAT? EXPLORING THE EFFECTS OF PERCEIVED GOAL PROGRESS AND RUMINATION ON CONSUMER HEALTH DECISIONS

My Bui, University of Arkansas, Fayetteville Elizabeth Howlett, University of Arkansas, Fayetteville Scot Burton, University of Arkansas, Fayetteville Steven Kopp, University of Arkansas, Fayetteville

SUMMARY 2006; Wansink and Chandon 2006; Chandon and Wansink 2007b). However, limited research has examined the Despite numerous efforts made among public policy internal evaluative motivations and mechanisms of the groups, researchers, and consumers to address the social self, which may lead to over-consumption from a con- issue of obesity, the percentage of Americans who are sumer psychological standpoint. This research examines overweight or obese has not diminished. As evidenced by motives that lead to differences in consumers’ consump- myriad of diet advertisements and the proliferation of tion decisions from the perspective of the self. More smaller-sized food packages, it is evident that many specifically, we address effects related to the consumer’s consumers are trying to lose weight. However Americans, level of goal progress toward his or her ideal weight and in general, have not been successful in their weight loss an individual’s ruminative propensity. attempts; instead, many appear to be gaining weight despite current market initiatives. A total of 95 subjects were surveyed initially for this study from a southern university in the U.S. Of the 95 students Diseases such as diabetes, stroke, and cancer are only a surveyed, 89 of the subjects were interested in losing few of the major health-related complications associated weight and only data collected from those interested in with obesity according to the Centers for Disease Control losing weight were used in this study and analysis. A 2 and Prevention (CDC 2007). Obesity has reached epi- (perceived goal progress: near ideal weight goal vs. far demic proportions and continues to increase; according to from ideal weight goal) X 2 (ruminative propensity: high the American Obesity Association, about 60 million ruminator vs. low ruminator) between-subjects experi- Americans are obese and 127 million are overweight. mental design was used to examine the effects on health- These increases in overweight and obesity are associated related consumption behavior. Perceived goal progress with a 20 percent increase in the annual expenditures on was manipulated within the study while ruminative pro- food purchases for consumption outside the home over pensity was a measured variable that was dichotomized the past three decades (Lin et al. 1999). using a median-split. This study tests predictions concern- To help combat this problem, legislation has recently been ing the effects of perceived goal progress and ruminative proposed at the state, local and national levels that would propensity on consumers’ willingness to pay more for mandate the provision of nutrition information on both healthier meals, attitudes toward engaging in a more menu boards and menus (i.e., Menu Education and Label- active lifestyle and mood. ing Act (MEAL) and Labeling Education and Nutrition Act (LEAN). Additionally, fast food restaurants includ- Willingness to Pay More for Healthier Meals ing KFC, Taco Bell, Pizza Hut, Long John Silver, and A&W All American Food announced in October 2008 There was a perceived goal progress by rumination inter- that they would voluntarily place product calorie info on action (F(1, 85) = 5.05, p < .05). Means indicate that low menu boards at its US-based restaurants. The provision of ruminators report greater willingness to pay more for calorie info should be completely phased in by January of healthier meals than high ruminators (M = 5.57 vs. M = 2011. 4.78) in the near goal condition. There was no overall main effect found for perceived goal progress (F < 1) nor There has been some recent research exploring the effects was a significant main effect found for rumination (F(1, of nutrition labeling of restaurant items (Burton et al. 85) = 2.10, p < .05).

118 Marketing and Public Policy Conference Proceedings / 2009 Attitude Toward Engaging in a More Active Lifestyle more positive (p < .05) mood in the near ideal weight condition than high ruminators (M = 5.81 vs. M = 4.71). The analyses yielded a perceived goal progress by rumi- The analyses also revealed a main effect of rumination nation interaction (F(1, 84) = 5.01, p < .05). The manipu- (F(1, 85) = 4.27, p < .05). Means indicate that low lation of perceived goal progress resulted in differences in ruminators reported significantly more positive (p < .05) attitude toward engaging in a more active lifestyle be- mood in comparison to high ruminators (M = 5.61 vs. M = tween high (low) ruminators in the near goal condition. 4.71). No overall main effect for perceived goal progress Follow-up contrasts indicate a marginally significant dif- was found (F < 1). ference (p = .09) between low and high ruminators’ attitudes toward engaging in a more active lifestyle, with These findings suggest that rumination is a key individual low ruminators reporting more favorable attitudes than difference factor that plays an important role in self- high ruminators (M = 6.36 vs. M = 5.77) in the near goal regulatory behavior. Although low ruminators generally condition. There were no overall main effects for rumina- have a more positive outlook than high ruminators, it is tion and perceived goal progress (F’s < 1) on attitude not clear how one’s tendency to ruminate influences the toward engaging in a more active lifestyle. outcomes of self-regulatory attempts. Since both consumers’ progress toward their ideal weight goal and rumina- Mood tive propensity appear to influence consumption-related attitudes, findings should be of interest to fast food restau- Analyses revealed a perceived goal progress by rumina- rants marketers and public policy makers. References are tion interaction (F(1, 85) = 5.56, p > .05). Follow-up available upon request. contrasts show that low ruminators have significantly

For further information contact: My Bui Sam M. Walton College of Business 302 Business Building University of Arkansas Fayetteville, AR 72701 Phone: 479.575.4005 Fax: 479.575.8407 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 119 THE RELATIONSHIP BETWEEN HOMESICKNESS AND COMPENSATORY FOOD CONSUMPTION AMONG COLLEGE STUDENTS

Elizabeth Crosby, University of Illinois at Urbana-Champaign Cele C. Otnes, University of Illinois at Urbana-Champaign

SUMMARY In particular, researchers find that consumers may engage in eating unhealthy foods when they experience negative The concept of home is “both a concrete physical object emotions (Beckwith 1987; Garg, Wansink, and Inman and an abstract term reflecting an individual’s relation- 2007; Kim et al 2006; Woodruffe 1997), in order to ship to that object” (Allen 2008, p. 85). Home is an provide short-term gratification and relieve negative psy- important component of identity and can reinforce ties to chological states. a consumer’s family, community, and culture (Hauge and Klstad 2007). Consumers have become increasingly mo- As mentioned, one cohort prone to homesickness is the bile, with Americans moving an average of 10.5 times in college student population. Given that scholars have es- their lives (Long 1988). Although moving often results in tablished the link between negative moods and unhealthy positive outcomes, it is also linked to temporary condi- compensatory eating, and that compensatory eating tions such as “moving day blues” as well as more perma- emerged as a widespread coping strategy in our study, we nent states of depression, longing, and mourning for believe it is imperative to examine the links between weakened social and cultural ties (Marcoux 2004). In college students’ eating behaviors and experiences of short, these effects can result in homesickness, which has homesickness. Understanding these linkages is important been generally described as a “complex cognitive-moti- because the types of eating habits established during the vational-emotional state . . . associated with ruminations college years often endure throughout adulthood. Fur- about home, an intense desire to return home, depressed thermore, U.S. college students are experiencing increas- mood and somatic symptoms” (Fisher and Hood 1987, ing levels of obesity and overweight (Cousineau, Goldstein, p. 426). Research confirms that the effects of homesick- and Franko 2004). Thus, important public policy issues ness are overwhelmingly negative – leading to loneliness, emerge with respect to educating college students about sadness, and difficulty adjusting to the new location the downsides of compensatory consumption habits. We (Constantine et al. 2005) as well as depression (Beck, explore the following research questions: (1) How do Taylor, and Robbins 2003) and anxiety (Van Tilburg, students use compensatory food consumption to cope Vingerhoets, and Van Heck 1999). with homesickness? (2) How can food consumption affect homesickness? and (3) How does compensatory food Homesickness is relatively widespread in the U.S., with consumption that is associated with homesickness impact 50–75 percent of the population experiencing it to varying consumer welfare? We demonstrate that homesickness degrees (Fisher 1989). It has been observed in a variety of can influence compensatory food consumption directly populations including children at boarding school (Fisher, and indirectly, find that food consumption can trigger or Frazer, and Murray 1986) and summer camps (Thurber increase consumers’ feelings of homesickness and show 1999), university students (Van Tilburg, Vingerhoets, that homesickness and compensatory food consumption and Heck 1999; Constantine et al. 2005), immigrants can negatively affect consumer welfare, through such (Matt 2007) and migrants (Salih 2001). While many of effects as weight gain. these studies explore how consumers cope with homesickness, scholars have not examined how consumers Method. engage in compensatory consumption to cope with this condition. Compensatory consumption occurs “when an As part of a larger study on how homesickness influences individual feels a need, lack or desire which they cannot consumption, we offered students in two undergraduate satisfy with a primary fulfillment so they use purchasing marketing courses at a Midwest university extra credit in behavior as an alternative means of fulfillment” exchange for written narratives that described times when (Woodruffe-Burton 1998, p. 301). Consumers are more homesickness affected their consumption behavior. We likely to engage in compensatory consumption when they collected 209 narratives (138 from females; 71 from experience negative emotional states (Woodruffe 1997). males). We eliminated 12 of these due to poor quality,

120 Marketing and Public Policy Conference Proceedings / 2009 leaving us 197 narratives that yielded over 700 pages of tent with the findings of Garg and her colleagues, we find data. Although we did not anticipate such an overwhelm- the foods our college students choose for recreational ing discussion of food consumption in these narratives, eating are often high in fat and low in nutrition (e.g., chips, 164 of 197 narratives (83%) mentioned food consumption candy, cookies, and pizza). Although these foods do not as either a trigger and/or coping mechanism for homesick- remind our informants of some aspect of their home, they ness. do help boost their moods. Furthermore, our informants engage in recreational eating both in private and with In conducting our analysis, we read all narratives multiple others . . . times to identify salient and emergent themes. Once we chose to focus on the links between food consumption and We also find that food consumption can exacerbate – and homesickness, we practiced dialectical tacking (Strauss even trigger – homesickness. That is, foods that remind and Corbin 1998), immersing ourselves in the interdisci- our informants of their home or geographic region can plinary literature on homesickness, compensatory food trigger homesickness, even if our informants were not consumption and product symbolism to help illuminate experiencing this condition prior to consuming these our informants’ experiences. foods. Furthermore, if informants are already homesick and attempt to cope by consuming “home” foods, doing so Four main themes that pertain to the linkage between can increase feelings of homesickness. Bronwyn writes compensatory food consumption and homesickness that when she feels homesick, she drinks hot chocolate emerged in our analysis. That is, informants attempt to because it reminds her of home. However, she then feels cope with homesickness by consuming foods that (1) “even more homesick than before.” In consuming foods remind them of their home, (2) are associated with the that remind them of home, informants often dwell on their geographic region that they perceive as home; (3) are absence from these places, reinforcing the separation in embedded in home rituals, with informants attempting to their own minds. This finding echoes Belk’s (1988) asser- recreate these food rituals in their new settings, and (4) are tion that possessions can serve as powerful extensions of associated with recreational eating. selves – including places associated with selves.

First, consumers often attempt to cope with homesickness Eating food that is not prepared as well as the food at home by buying and consuming foods that remind them of can also trigger or exacerbate homesickness. Harry notes home. For example, Amanda buys the same garlic spread that consuming subpar dorm food made him miss his as her mother to cope with her feelings of homesickness. home, where his mother cooked delicious meals. Further- Consumers also cope by consuming foods that remind more, recreational eating can also increase consumers’ them of their geographic region of origin. Melissa (F, 20s) homesickness because it only provides short-term gratifi- notes that when she is homesick, she eats hash browns cation, often leaving the consumer more homesick than because they remind her of “the South.” before once the effects wear off.

Other consumers cope by attempting to recreate food Finally, we find that the relationship between homesick- rituals, substituting artifacts and ritual performance roles ness and compensatory food consumption can have both (Rook 1985) when these cannot be reproduced. For ex- physical and emotional negative consequences on con- ample, many informants replace the family members who sumer welfare. Several informants report that they gained would be present at meal ritual with their friends. Infor- weight from using food as a coping mechanism. Others mants also alter family recipes when certain ingredients or report a lack of energy and increased feelings of sluggish- brands are not available, or go to a restaurant that offers ness. As noted above, using food consumption as a coping foods integral to a ritual. Nathan recreates his family pasta mechanism for homesickness can be counterproductive, night by going out for pasta with friends. Even though he as it can increase homesickness and along with it the cannot enjoy his mother’s pasta, Nathan notes that his feelings of loneliness, stress, anxiety, and depression. We recreation of pasta night at a restaurant helps alleviate his believe that public policy makers could use the findings in homesickness. this study to help develop nutritional information and messages that are aimed at helping college students real- The final way consumers use compensatory food con- ize the deleterious effects of compensatory eating. For sumption as a coping mechanism for homesickness is by example, health messages that remind students to be “self- engaging in recreational eating. When consumers do so, compassionate” (e.g., to forgive themselves for overin- they eat to enjoy short-term gratification and relieve dulging a little bit, but to still be aware of the health effects negative psychological states (Beckwith 1987). Consis- of overcompensation) might improve their ability to con-

Marketing and Public Policy Conference Proceedings / 2009 121 sider and revise their eating habits before these habits is this cohort highly susceptible to homesickness, but cause too much harm (Adams and Leary 2007). students tend to practice poor eating habits even in the absence of homesickness (Cousineau, Goldstein, Franko In conclusion, we believe the links between homesickness 2004), exacerbating the negative physical effects of com- and compensatory food consumption on the welfare of pensatory eating on their well-being. college students deserve to be studied more fully. Not only

REFERENCES tive Study Among Residents in High-Priced and Low-Priced Neighbourhoods in Norway,” Housing, Adams, Claire E. and Mark R. Leary (2007), “Promoting Theory, and Society, 24 (4), 272–92. Self-Compassionate Attitudes Toward Eating among Kim, Junyong, Brian Wansink, Meryl P. Gardner, Se- Restrictive and Guilty Eaters,” Journal of Social and Bum Park, and Collin Payne (2006), “The Effects of Clinical Psychology, 26 (10), 1120–44. Mood on Comfort Food Consumption,” Advances in Allen, Sarah (2008), “Finding Home: Challenges Faced Consumer Research, 33 (1), 248. by Geographically Mobile Families,” Family Rela- Long, Larry (1988), Migration and Residential Mobility tions, 57 (January), 84–99. in the United States. New York: Russell, Sage Foun- Beck, Richard, Cathy Taylor, and Marla Robbins (2003), dation. “Missing Home: Sociotropy and Autonomy and their Marcoux, Jean-Sebastien (2004), “Moving on to Some- Relationship to Psychological Distress and Home- thing Else: The Social Relations of Women During Sickness in College Freshmen,” Anxiety, Stress, and Separation,” in Contemporary Consumption Rituals: Coping, 16 (June), 155–66. A Research Anthology, Cele C. Otnes and Tina M. Beckwith, Joan B. (1987), “Psychological Functions of Lowrey, eds. Mahwah, NJ: Lawrence Erlbaum Asso- Eating, Drinking and Smoking in Adult Women,” ciates, Publishers. Social Behavior and Personality, 15 (2), 185–206. Matt, Susan J. (2007), “A Hunger for Home: Homesick- Belk, Russell W. (1988), “Possessions and the Extended ness and Food in a Global Consumer Society,” Jour- Self,” Journal of Consumer Research, 15 (Septem- nal of American Culture, 30 (1), 6–17. ber), 139–69. Rook, Dennis (1985), “The Ritual Dimension of Con- Constantine, Madonna G., Gregory M. Anderson, LaVerne sumer Behavior,” Journal of Consumer Research, 12 A. Berkel, Leon D. Caldwell, and Shawn O. Utsey (December), 251–64. (2005), “Examining the Cultural Adjustment Experi- Salih, Ruba (2001), “Moroccan Migrant Women: ences of African International College Students: A Transnationalism, Nation-States and Gender,” Jour- Qualitative Analysis,” Journal of Counseling Psy- nal of Ethnic and Migration Studies, 27 (4), 655–71. chology, 52 (January), 57–66. Strauss, Anselm and Juliet Corbin (1998), Basics of Cousineau, Tara M., Marion Goldstein, and Debra L. Qualitative Research: Techniques and Procedures Franko (2004), “A Collaborative Approach to Nutri- for Developing Grounded Theory, 2nd ed. Thousand tion Education for College Students,” Journal of Oaks: Sage. American College Health, 53 (2), 79–84. Thurber, Christopher A. (1999), “The Phenomenology of Fisher, Shirley and Bruce Hood (1987), “The Stress of the Homesickness in Boys,” Journal of Abnormal Child Transition to University: A Longitudinal Study of Psychology, 27 (2), 125–39. Psychological Disturbance, Absent-Mindedness and Van Tilburg, Miranda A.L., Ad J.J.M. Vingerhoets, and Vulnerability to Homesickness,” British Journal of Guus L.Van Heck (1997), “Coping with Homesick- Psychology, 78 (4), 425–41. ness: The Construction of the Adult Homesickness ______(1989), Homesickness, Cognition, and Coping Questionnaire,” Personality and Individual Health, London: Erlbaum. Differences, 22 (6), 901–07. Garg, Nitika, Brian Wansink, and J. Jeffrey Inman (2007), ______(1999), “Determinants of Homesickness “The Influence of Incidental Affect on Consumers’ Chronicity: Coping and Personality,” Personality Food Intake,” Journal of Marketing, 71 (January), and Individual Differences, 27 (3), 531–39. 194–206. Woodruffe, Helen (1997), “Compensatory Consumption Hauge, Ashild Lappergard and Arnulf Kolstad (2007), (Or: Why do Women Go Shopping When They’re “Dwelling as an Expression of Identity. A Compara- Red Up? And Other Stories),” Marketing Intelli-

122 Marketing and Public Policy Conference Proceedings / 2009 gence and Planning, 15 (7), 325–34. Theory,” Advances in Consumer Research, 32 (1), Woodruffe-Burton, Helen and Richard Elliott (1998), 461–65. “Compensatory Consumption and Narrative Identity

For further information contact: Elizabeth Crosby or Cele C. Otnes University of Illinois at Urbana–Champaign 1205 South Sixth Street Champaign, IL 61820–6980 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 123 WHEN FOOD IS MORE THAN NUTRITION: REGULATING EMOTIONS WITH FOOD

Elyria Kemp, Texas State University, San Marcos Steven W. Kopp, University of Arkansas, Fayetteville

SUMMARY individuals may turn to in order to mitigate negative emotions, is food, particularly fatty and high carbohy- A number of factors have been cited as contributors to the drate-content foods. obesity epidemic, including increased consumption of meals outside the home, larger portion sizes, the prolifera- As an exploratory assessment of how individuals engage tion of fast food, and sedentary lifestyle behavior in consumption behavior to mitigate negative emotions, (Lackdawalla and Philipson 2002; Cutler, Glaeser, and we asked seventy-six college undergraduate students to Shapiro; Seider and Petty 2004). Many initiatives and recall a time when they felt bad about something and what campaigns have been launched to change not only eating they did to make themselves feel better. Fifty-four percent habits, but also the way that food is marketed (Eldrige of the students indicated that they engaged in some form et al. 1998; Gogi 2006; Gibson 2007). However, policy of eating to down-regulate negative emotion. Other ac- initiatives and social marketing programs might also tivities reported included shopping (24%), drinking (14%), consider the role that individual emotions play in food and participating in some form of entertainment (21%). consumption. The purpose of this research was to exam- Subsequently, emotional eating is a common consump- ine which negative emotions were more likely to prompt tion activity used by a segment of the population in order impulsive hedonic food consumption and to suggest a to manage emotion. This research examined the relation- broader understanding of what may be influencing eating ship between discrete negative emotions and impulse habits. eating propensities. Specifically, the relationship between individuals’ intentions to engage in an impulsive hedonic Emotions have a profound effect on our consumption and food purchase and emotional eating behavior when expe- purchase behavior. Evidence from the consumer research riencing sadness, anxiety and anger were investigated. literature suggests that individuals may engage in consumption to “down-regulate” or decrease the negative Methodology and Results emotions they experience (Garg, Wansink, and Inman 2007; Andrade 2005). Individuals may especially use In order to access intentions to purchase a hedonic food eating to regulate their emotions. Emotional eating is a product, one hundred thirty-three participants were pre- common phenomenon in response to stress or sadness sented with a shopping scenario adapted from Rook and (D’Arrigo 2007). Research suggests that individuals may Fisher (1995) which asked them to image that they had receive positive chemical reinforcement from consuming gone to the grocery store to pick up some needed items. some foods, particularly those high in carbohydrates While in the store, they saw a decadent bakery-style (Martin 2001). After continuous reinforcement, these cheesecake. Even though cheesecake was not one of the behaviors become habits. Each time the behavior is rein- items they had gone to the store with the intention of forced, the habit is strengthened (Thayer 1996; Hirschman buying, participants were asked to indicate their desire for 1992). the cheesecake as well as the probability that they would purchase the cheesecake. Next, participants were then The “broaden-and-build theory” of positive emotions given the Emotional Eating Scale (Arnow and Agras (Fredrickson 1998, 2000) from the positive psychology 1995). The EES assessed eating in response to three literature suggests that positive emotions have the ability discrete negative emotions: anger, sadness and anxiety. A to “down-regulate” or even “undo” the effects of negative five-point scale was used to rate the degree to which an emotions. Because hedonic products often generate posi- individual would eat based on an emotional feeling. tive affective responses from the consumer (Hirshman and Holbrook 1982; Voss, Spangenberg, and Grohmann Results indicated that there was a significant and positive 2003), individuals experiencing negative emotion may relationship between the sadness subscale and desire, (r = consume or purchase hedonic products in an effort to .22, p < .05). Those individuals that had greater tendencies invoke positive emotion. Among hedonic products that to eat when feeling sad, expressed more desire for the

124 Marketing and Public Policy Conference Proceedings / 2009 cheesecake. Additionally, results showed that these “sad from this study suggest situational and contextual factors emotional eaters” demonstrated a higher probability of in unhealthy eating. Studying the nuances of emotional purchasing the cheesecake, (r = .18, p < .05). However, eating has significant consumer welfare implications due there were no significant relationships between desire and to high obesity rates in the United States and the failure of probability of purchasing the cheesecake for both the most diets (Felgal et al. 2002). Regulating emotional anger and anxiety subscales (p > .05). distress often takes precedence over impulse-control and individuals may engage in consumption activities to Discussion achieve short-term gratification from negative feelings. Social marketing programs might be developed to educate Although past and current efforts to persuade consumers and make individuals aware of why they might be engag- to eat healthier have tended to focus on the provision of ing in such destructive behavioral patterns and encourage objective information (e.g., nutrient content), the results more healthy alternatives to manage emotions.

For further information contact: Elyria Kemp McCoy College of Business Administration Texas State University San Marcos, TX 78666 Phone: 512.245.7882 Fax: 512.245.7475 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 125 INDIVIDUALISM, COLLECTIVISM, AND CORPORATE SOCIAL RESPONSIBILITY: INFLUENCE ON OBESITY BELIEFS AND ATTITUDES TOWARD PUBLIC POLICY

Jung-Sook Lee, Towson University

SUMMARY measurements of the key variables included in the conceptual framework of the study. Items that measured Since the U.S. General Surgeon’s report declared obesity importance of obesity, personal and social causes of an epidemic in 2001, the issue has generated active obesity, personal and social responsibilities to solve the research that has included studies on the public’s view on obesity problem, and attitudes toward obesity-related obesity as a public health issue and its attitudes toward public polices were adopted partly from previous public possible regulatory policies on food marketing. Support opinion surveys on the topic (Oliver and Lee 2005; San from the public is vital in a successful implementation of Jose Mercury News/Kaiser Family Foundation 2004). any public policy related to obesity and public’s view on Additional items for these measures were created and obesity may change over time (e.g., Brownell 2005). tested by the author. Individualism and collectivism were While previous research helps us understand where the measured using part of a scale developed by Singelis and public stands on these two issues separately (e.g., Oliver his colleagues (Singelis et al. 1995) as reported in Triandis and Lee 2005), no research, to the author’s knowledge, (1996). Beliefs in Corporate Social Responsibility (CSR) has directly tested the relationship between the two. Also, were measured with scale items used in previous research no previous research has investigated the factors that may (Cone Corporate Citizenship Study 2004; Mohr et al. influence consumers’ beliefs on obesity, which, in turn, 2001; Webb et al 2007). may affect their attitudes toward public policy. To the extent that college students’ opinion also reflects This study first examines any changes since the last opinion of the general public, the study findings support survey data were reported in 2004 in terms of consumers’ an increasing trend reported in previous research that beliefs on obesity (importance, personal cause, social obesity is viewed as a serious public health issue, it is cause, personal responsibility, social responsibility) and caused largely by the social environment we live in, and their attitudes toward public policy on food marketing. solutions to the problem need to be addressed at a societal Next, it investigates the relationship between consumers’ level, noted in a descending order of substantial changes beliefs on obesity and their attitudes toward different since the last survey results were reported in 2004. Con- types of policies on food marketing. Borrowing the frame- sumers’ attitudes toward moderate policy measures, con- work from previous research (Moorman and Price 1989; sumer education and nutritional information disclosure, Seider and Petty 2004), four types of policy are included are more favorable than are their attitudes toward strong in the study: consumer education, disclosure of nutritional policy measures, imposing tax on junk food or restricting information, imposing a special tax on junk food as food advertising. These results are consistent with those financial disincentives, and restriction of food advertis- from earlier surveys. ing. Finally, the study explores the roles of individualism, collectivism, and beliefs in Corporate Social Responsibil- Results from stepwise regression analyses indicate that ity (CSR) on obesity-related beliefs and attitudes toward consumers’ belief that solving obesity is a social respon- public policy on food marketing. Individualism, collec- sibility is the most significant and robust predictor of their tivism, and CSR represent consumers’ value orientations attitudes toward all four types of public policies. Beliefs that are thought to be inherently related to their beliefs on social cause of obesity and importance of obesity, in about the causes of obesity and responsibilities to solve this order, are the next significant predictors. Beliefs on the obesity problem, which, in turn, may affect their personal responsibility to and personal cause of obesity attitudes toward obesity-related public policy. become significant predictors of attitudes toward public policy only when the policies involve specific financial The data were collected from a survey of 316 college disincentives or restriction, and only after most of the students from a medium-size university in the Mid-Atlan- variance is already explained by social responsibility to tic region in December 2008. The questionnaire included and social cause of obesity.

126 Marketing and Public Policy Conference Proceedings / 2009 Both collectivism and beliefs on Corporate Social Re- likely to receive the public’s support. Consumers may see sponsibility (CSR) demonstrate significant positive rela- these policies as the government’s effort to provide indi- tionships with the importance of obesity, social cause of viduals with necessary information so that the individuals obesity, and social responsibility to obesity. Their rela- can solve the obesity problem on their own. Findings from tionships with personal cause of or personal responsibility the current study suggest that support for these policies to obesity are mixed. Findings from hierarchical regres- can be further strengthened by focusing on social causes sion analyses indicate that collectivism and CSR are of obesity, social solutions to obesity, collectivistic val- directly related to the attitudes toward public policies, in ues, and Corporate Social Responsibility. addition to their indirect relationships through obesity- related beliefs, when the policies are about consumer On the other hand, polices implementing a special tax on education or information disclosure. Their relationships junk food or restriction of food advertising pose a greater with public policies involving financial disincentives and challenge and are less likely to be supported by the public. restriction of food advertising are observed only indi- Consumers may see these policies as unnecessary govern- rectly through obesity-related beliefs. Individualism shows mental intervention in a private matter of one’s food significant positive relationships with the importance of choices. However, past research has shown that public obesity, personal cause of obesity, and personal responsi- policies with restrictive enforcement were more effective bility to obesity. It does not hold any significant relation- in creating desired behavioral changes than were policies ships with social cause of or social responsibility to focusing on education and information. Findings from the obesity. Hierarchical regression analyses suggest that the current study suggest that support for these policies may relationships between individualism and the four public be increased by bringing public’s attention to the impor- policies occur only indirectly through obesity-related tance of obesity as a serious public health issue and the beliefs. No direct relationship is observed. limitations in treating the problem as an individual responsibility, while generally emphasizing the social causes Public policies on consumer education and nutritional of and social responsibility to obesity. References are information disclosure are safe to implement and are more available upon request.

For further information contact: Jung-Sook Lee Towson University 8000 York Road Towson, MD 21252 Phone: 410.704.2601 Fax: 410.704.3656 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 127 IMPROVING EFFECTIVENESS OF RESTAURANT MENU LABELING: AN EXPLORATORY RESEARCH

Saroja Subrahmanyan, St. Mary’s College of California, Moraga

SUMMARY well as on restaurant menus. All FG and IDI participants were California residents. Number of comments in each The Nutritional Labeling and Education Act (NLEA) of blog varied from five to four hundred. Several themes 1990 was expected to assist consumers in making healthier were identified from all these sources. These have been food choices in the U.S. National studies indicate that grouped into three categories and discussed below in the three quarters or more of American adults report using Results section as: how menu labeling is perceived to be food labels. There are also several studies that link use of helpful, insufficient and unhelpful. food labels on packaged items with healthier diets (Kim et al. 2000; Kreuter et al. 1997; Neuhouser et al. 1999; Results Hawkes 2004). Yet, the Center for Disease Control and Prevention reported that in a survey done between years The most frequently mentioned reason for how menu 2003–2006, there was no significant reduction in obesity labeling can be helpful is how it can point out hidden rates and that over 60 percent of adult Americans continue calories. A typical sentiment was: “It is so easy to be fat in to be overweight, and one-third are obese. While research the U.S.! I think that many people do not realize where has shown that many factors could contribute to the their extra weight is coming from. I was eating Ceasar continued prevalence of obesity (Sturm 2008), increased salads every day so where did all this extra weight come consumption of food-away-from-home and increasing from?!!?” This is consistent with studies that reported that portion sizes have been implicated as major causes of consumers significantly underestimate the amount of calo- obesity (Nielsen and Popkin 2003; Young and Nestle ries, fat, saturated fat in less healthy restaurant items 2002; Thomson et al. 2004; Wootan and Osborn 2006; (Burton et al. 2006). However, such reactions seem to be The Keystone Forum 2006). This has led health and from those who already pay attention to labels on pack- nutrition advocates to spearhead menu labeling laws. aged foods and are nutrition or calorie conscious or have Such laws have already been passed in several places. In health issues like diabetes. Total calories in a menu item most cases, the laws require that chain restaurants (10–20 can also alert consumers to larger portion sizes as noted by outlets depending on the city or state) post the calorie a woman who ate occasionally at Burger King, “I switched content of standard items listed on the menu. In some out a Whopper for a Whopper Junior when I saw the cases, the law also requires that brochures be made avail- numbers.” Some participants also mentioned that being able that list in addition the levels of sodium, saturated fat required to report on calorie and fat content will push and trans-fat. However, research on effectiveness of res- restaurants to avoid “bad fats,” offer healthier choices and taurant menu labeling is still in its early stages and there smaller portion sizes. Thus, the current format of menu is still no conclusive evidence on how it would affect labeling appears to be most helpful to those who are consumer behavior. The purpose of this exploratory study already pay attention to such information. is to investigate consumers’ perceptions of menu labeling and how they would act based on that information. This The main point of criticism on why current form of menu can help uncover factors that would improve effective- labeling is insufficient was that the current focus is on ness of menu labeling. calories, not overall nutrition and lifestyle. Many participants felt that additional information on protein, mineral Methodology and fiber content were needed to make more informed decisions. Some pointed out that having mainly calorie In this study, I used multiple exploratory research meth- and fat information places more emphasis on weight ods. I conducted two focus groups (FGs) each with seven rather than overall health. One person mentioned that, participants, five in-depth interviews (IDIs) and content “An eight-ounce glass of skim milk is more caloric than a analysis of twenty blog-sites that had consumers’ opin- similarly sized serving of diet soda, but the milk is more ions on menu labeling. The FGs and IDIs were structured nutritious.” Another participant pointed out that “The to elicit participants’ knowledge, perceptions, and usage emphasis on weight can be emotionally triggering for (or likely usage) of nutritional labels on packaged food as those struggling with eating disorders.” The negative

128 Marketing and Public Policy Conference Proceedings / 2009 effects of emphasizing calories has been documented by There were some who thought that menu labeling is not a study (Food Affairs 1996) that showed that teenage helpful, because choosing good foods is a matter of girls’ diets lacked the daily requirements for calcium and commonsense and everyone already knows that eating iron and overall energy intake was 82 percent lower than fast food is unhealthy. “I still eat there because it is required. Another related issue brought up was that if you convenient, cheap and tasty,” said a FG participant who were not full after eating the lower calorie meal, you frequently ate at fast-food outlets. Some participants also compensated for it later. As a blogger noted, “Anyone mentioned that the joy of eating is reduced by the constant who is actually counting calories at mealtime is either calorie reminder, especially when dining out was a treat. eating healthily, or “cheating” on their diet with snacks.” Others who were critical of menu labeling mentioned that Many participants also expressed interest in having infor- having food regulated to a large extent reduces the role of mation about ingredients such as: freshness, where they self-responsibility and increases the chances that we con- are sourced, whether they are organic, whether they have stantly look for scapegoats for our weight problems. preservatives and so on. Some participants felt that having labels that mentioned health benefits or risks would be In conclusion, although menu labeling is perceived by easier to interpret. Such participants referred to the Berke- many consumers as a good first step, more research is ley Eat Well program, the traffic light format in the U.K. needed to improve its effectiveness in achieving the or the recent NuVal system that gives nutrient density overall goal of promoting healthier diets. As a follow-up scores. Showing health benefits or risks could be effective research to this exploratory study, I plan to develop a scale in restaurant settings as shown by the Albright et al. to measure effectiveness of menu labeling and then exam- (1990) in the Stanford Five City projects. Many partici- ine how it is related to various factors such as nutrition pants in the FGs, especially those who exercised very knowledge, wellness consciousness, label format, pres- regularly, mentioned that the role of exercise in maintain- ence of health claims and attitude toward menu labeling. ing health is not highlighted enough. References are available upon request.

For further information contact: Saroja Subrahmanyan St. Mary’s College of California 380 Rheem Center Moraga, CA 94575 Phone: 925.631.8675 Fax: 925.376.6521 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 129 CONSUMER USE OF NUTRITION SYMBOLS

Monique Mitchell Turner, University of Maryland, College Park Lucinda Austin, University of Maryland, College Park Meryl Lubran, University of Maryland, College Park Sejal Patel, University of Maryland, College Park Christine Skubisz, University of Maryland, College Park

SUMMARY wanted to know how nutrition symbols interact with ability and motivation to affect attention to nutrition Over the last several years, food and beverage manufac- information. turers, as well as trade groups and nutrition and health organizations, have developed symbols to be placed on To explore this question, a 3 (Task motivation: health food packages to guide consumers toward “better-for- goal, taste goal, no goal) X 2 (Ability: high vs. low) X 12 you” choices. These nutrition symbols, which are usually (Food Products: 6 with nutrition symbols, 1 with a health found on the front of packages, vary in their format from claim, 1 with a nutrient claim, 1 with mixed symbols, and simply presenting a single icon indicating a food is health- 3 with no symbols) mixed design was conducted. Partici- ful to presenting detailed graphics indicating the content pants (n = 90) were undergraduate students from a large of selected nutrients (Hentzepeter 2007). Problemati- Mid-Atlantic university (Mean age = 19.89, SD = 1.99) cally, depending on the format, nutrition symbols may be who received course credit for participation. Motivation regulated differently. was experimentally manipulated by giving participants a task to shop for another person who either wanted to Under the Nutrition Labeling and Education Act of 1990 maintain a healthy diet (health goal), or who was inter- (NLEA), the Food and Drug Administration (FDA) has ested in food based on the taste (taste goal). Remaining authority over claims, both expressed and implied, that are participants were assigned to the control condition and made on food packages. The NLEA was enacted to ensure were not given a goal. Ability was measured by asking that information on food packages is truthful, not mislead- participants if they had ever been formally taught to read ing, and helps consumers make informed and healthier a nutrition label. Symbol type was varied by having all food choices (FDA 1994). Claims that can be used on food participants view images of 12 real food products. fall into three categories: nutrient content claims, health claims, and structure/function claims. Nutrition symbols Participants arrived at the laboratory and were seated at a may or may not fall into one of these three categories, computer terminal that had a non-intrusive near-infrared causing several implications. Heart-shaped logos, for eye-tracker in a 17” monitor (MyTobii D10 Eye Tracker, example, are treated as implied health claims because they Tobii AB). The experimenter read motivation instructions suggest a relationship between the product and a health- and asked the participant to begin shopping. For all 12 related condition (i.e., coronary heart disease). As a result, products, participants viewed an image, which had the all labels bearing this type of symbol must also include the front of the package side-by-side with the product’s full health claim and the food must meet certain scientific nutrition label, and indicated their purchase intent. Images criteria (e.g., not too much fat). were presented in random order. Eye movements were recorded while participants viewed each product. Depen- In order to understand how consumers use different nutri- dent variables were participants’ gaze duration (in milli- tion symbols to make food choices, it is critical that we seconds) on three separate portions of the nutrition label: explore how much cognitive processing of these symbols the top portion of the label which includes calorie and takes place versus traditional nutritional information, and serving size (NF1), the middle portion of the label which to what end. Dual process models, such as the Heuristic- includes fat, cholesterol, sodium, carbohydrates, and pro- Systematic Model (Chaiken 1980), predict that people tein (NF2) and the bottom portion of the label which will engage in “rule of thumb” processing (e.g., over-use includes vitamins, minerals, and the Daily Value footnote symbols and other cues instead of processing relevant (NF3). To measure purchase intent, we asked “How likely information) when their ability and motivation to process are you to purchase this item?” (1 = “very likely,” 5 = is low. Using dual process models as a framework, we “very unlikely”).

130 Marketing and Public Policy Conference Proceedings / 2009 Mixed model repeated-measures Analysis of Variance label while individuals’ who were less motivated (taste (ANOVA) was employed to examine main effects and condition) relied on cues, such as nutrition symbols, to interaction effects between conditions and dependent make their decision. Of interest to policy makers, the variables. Products were treated as a within subject factor; results also suggest that we cannot treat symbols equally ability and motivation were treated as between subjects in terms of the effects they exert on attention. So, the factors. Within-subjects analyses indicated a significant question is not whether the presence of a symbol affects main effect for product for NF1 (F(11,178) = 3.50, p < the amount of time people spend looking at the nutrition .001, partial η2 = .04) and NF3 (F(6.54,178) = 3.25, p < label; rather, the question is what kind of symbol the .05, partial η2 =.04). A significant interaction for product product has and how does that particular symbol influence by motivation was found for NF1 (F(22,178) = 1.83, p < gaze duration. .05, partial η2 = .04), NF2 (F(11,178) = 1.80, p = .05, partial η2 = .02), and NF3 (F(13.07,178) = 2.54, p < .05, This study provides an exploratory analysis of the use of partial η2 = .06). Between subjects analyses indicated a nutrition symbols by consumers. There are a few limita- significant main effect for motivation for NF1 (F(2,178) = tions to these results. This was a highly educated sample 21.68, p = < .001, partial η2 = .35), NF2 (F(2,178) = 6.03, and ability was self reported, limiting generalizability and p = < .05, partial η2 = .13), and NF3 (F(2,178) = 15.82, p < validity. In addition, because of the eye-tracking soft- .001, partial η2 = .28). No other interactions or main ware, images of the front of the packages had to be effects emerged. displayed next to the nutrition labels limiting external validity. Future research should focus on controlling These results suggest that depending on an individual’s potential confounding factors such as brand, involve- motivation, nutrition symbols affect gaze duration, re- ment, familiarity, and flavor, as well as manipulating the gardless of ability. So, in the case of nutrition labels, types of symbols. In order to know if relying on a symbol ability is not a strong factor; but, motivation is critical. has a positive or negative effect on purchasing decisions, Importantly, individuals’ who were highly motivated more research is needed to understand if consumers make (health condition) spent time scrutinizing the nutrition inferences beyond the literal meaning of the claims.

For further information contact: Meryl Lubran Department of Communication Center for Risk Communication Research 2130 Skinner Building University of Maryland College Park, MD 20742 Fax: 301.314.9471 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 131 THE RELATIONSHIP BETWEEN ADVERTISING VULNERABILITY AND WEIGHT-LOSS STRATEGY USAGE AMONG OBESE CONSUMERS

Sheri L. Worthy, Mississippi State University, Starkville Kenneth Pilcher, Mississippi State University, Starkville Kristine Lokken, University of Alabama – Birmingham Abbe Boeke, University of Alabama, Tuscaloosa

SUMMARY among vulnerable obese consumers and the nature of weight-loss strategies they use. Several researchers have suggested a link between demographic characteristics such as low levels of education, Pre-surgical candidates for bariatric surgery (193 female, lower income, minority status, older age, being single, or 61 male) were surveyed and all 254 met criteria for being female to consumer vulnerability. Smith and Coo- seeking bariatric surgery (i.e., BMI of 40 or greater, or per-Martin (1997) expanded the definition of consumer BMI of 35 or greater with comorbid conditions such as vulnerability to include variables such as low levels of uncontrolled hypertension, diabetes, obstructive sleep cognitive ability and restricted mobility and defined vul- apnea, etc.). BMI was calculated by weight in pounds nerable consumers as, “those who are more susceptible to divided by height in inches squared multiplied by 703. economic, physical, or psychological harm in, or as a Weight and height were self-reported. A pre-surgical result of economic transactions because of characteristics mailing packet included questions related to demograph- that limit their ability to maximize their utility and well- ics, personal health and medical information, family and being” (p. 4, 1997). Although no research exists, obesity personal weight history, current eating behaviors, physi- could also be linked to consumer vulnerability. cal activity, dieting attempts, and consumer vulnerability.

Currently, there is no formal system of regulating the For this study, the authors developed a three-by-three weight-loss industry in the U.S. For example, dietary matrix with the X-axis representing product harmfulness. supplements do not need approval from the FDA before The three levels of beneficial, harmful due to abuse/ they are marketed, nor do manufacturers need to register misuse, and inherently harmful were adapted from Cui their dietary supplements with the FDA before producing and Choudhury (2003) and Rittenburg and Parthasarathy or marketing them (U.S. FDA 2001). Therefore, it is (1997). The Y-axis for the three-by-three matrix repre- essentially the consumer’s responsibility to determine sents consumer advertising vulnerability (sophisticated, whether or not a weight-loss product or service is safe and/ at-risk, vulnerable; Rittenburg and Parthasarathy 1997). or effective. This can be particularly difficult considering Other researchers have also used similar matrices and the prevalence of deceptive advertising in the weight-loss their work was used in this model (Jones and Middleton industry. 2006; Smith and Cooper-Martin 1997).

The aim of this research was to develop a three-by-three An Analysis of Variance (ANOVA) was performed with matrix of consumer advertising vulnerability (sophisti- gender as the independent variable and weight-loss strat- cated, at-risk, vulnerable) by nature of weight-loss strategy usage as the dependent variable. Results indicated egy usage (beneficial, harmful due to abuse/misuse, in- women had a significantly higher mean score than men for herently harmful) for extremely obese consumers opting all three weight-loss strategy categories. A similar ANOVA for bariatric surgery. Extremely obese is defined as having using race/ethnicity as the independent variable indicated a BMI (body mass index) of 40 or higher, which in and of no mean differences between Caucasians and African itself may increase a consumer’s vulnerability to weight Americans on type of weight-loss strategy used most loss strategies. This paper was an initial step in defining frequently. Another ANOVA with gender as the indepen- that matrix. Thus, the purpose of this research was to dent variable and consumer advertising vulnerability as determine characteristics of advertising vulnerability in the dependent variable showed no significant differences bariatric surgery candidates and investigate relationships between men and women. However, African Americans

132 Marketing and Public Policy Conference Proceedings / 2009 in this sample had a higher mean vulnerability score (m = diet use (m = 5.56, F = 4.96, p = .008) than at-risk 4.04) compared to Caucasians (m = 3.28, F = 7.31, p = consumers (inherently harmful m = 1.82; total diet m = .008). Pearson’s Correlations indicated significant posi- 5.00) and sophisticated consumers (inherently harmful tive relationships between annual income and use of m = 1.15; total diet m = 3.93). beneficial (r = .177, p = .008) and potential harmful weight-loss strategies (r = .178, p = .007), but no relation- In summary, results from this study indicate that obese ship between annual income and inherently harmful women with higher levels of education attempt more of all weight-loss strategies. Educational level was positively categories of weight-loss strategies. Obese consumers related to all three weight-loss strategy types and BMI was with higher income levels are more likely to try weight- negatively related to use of beneficial and inherently loss strategies that are considered beneficial or at least harmful weight-loss strategies. There was an inverse only harmful if misused. Obese African Americans in this relationship between educational level and consumer sample had higher mean consumer advertising vulner- advertising vulnerability (r = -195, p = .011). A final ability scores and obese consumers with higher levels of ANOVA revealed a mean difference between sophisti- education had lower consumer advertising vulnerability cated and at-risk/vulnerable consumers in their use of scores. Advertising vulnerability was related to increased inherently harmful weight-loss strategies. Vulnerable usage of inherently harmful weight-loss strategies. Refer- consumers reported more inherently harmful weight-loss ences are available upon request. strategy use (m = 1.97, F = 5.23, p = .006) and more total

For further information contact: Sheri L. Worthy Mississippi State University P.O. Box 9746 Mississippi State, MS 39762 Phone: 662.325.0918 Fax: 662.325.8188 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 133 THE MAGICAL POWERS OF SUPPLEMENTS: DO SUPPLEMENTS ALWAYS LEAD TO A HEALTHY LIFESTYLE?

Sayantani Mukherjee, California State University, Long Beach Ingrid M. Martin, California State University, Long Beach

SUMMARY anything when I am at work. Since these things I put in my body are so unhealthy, I have no choice but to lean toward Dietary supplements are ubiquitous in consumers’ healthy vitamins.” Parya’s pursuit of a healthy lifestyle is oriented lifestyle with the U.S. spending $20.3 billion annually toward avoiding losses indicating a prevention focus. She (Picciano 2004). However, there is mixed evidence that pursues this through a vigilant strategy of using supple- supplements improve health with concerns over the safety ments (as substitutes) to minimize the impact of an un- of certain supplements (Hellerman 2007). A better under- healthy lifestyle. standing of the underlying dynamics of supplement consumption is needed to design corrective interventions The second theme that emerged was that magical beliefs (Mason, Scammon, and Fang 2007). In this research, we guide consumers’ decision making (Kramer and Block address the question of whether supplements “always” 2008) for supplements. For example, although partici- lead to a healthy lifestyle. This is important from a public pants acknowledged that positive effects of supplements policy perspective given that consumers “should” use could be in their mind, they still cling to the use of dietary supplements in a safe and healthy manner consis- supplements for peace of mind. Again, this was after tent with the FDA. Additionally, we examine magical consumers recognized that supplements probably did not thinking underlying supplement consumption. This is help but it couldn’t hurt to take them. On the other hand, critical because magical beliefs have been attributed as a there is an underlying sense of unease if supplements are pervasive reason for some unsafe consumer practices not taken. For example, Jordan believes that: “My lifestyle regarding supplements (Stevens 2001). right now isn’t very healthy. I get really stressed out. My eating habits aren’t good. The reason that I take my We conducted sixteen depth interviews and three focus supplements is because it is a more natural way of doing groups in a pilot study. Content analysis revealed two it rather than taking medications.” important themes. First, two distinct consumer groups emerged, who differed in their supplement usage styles Consistent with prior research (Bolton et al. 2008), our (complement with vs. substitute for “other healthy prac- qualitative data suggests that participants associate supple- tices”). This theme expands on prior research in remedy ments with being natural and safe. It was as if supplements marketing (Bolton et al. 2008) that treated supplement by virtue of being natural possessed magical powers that takers as a homogeneous group. Further, our analysis made them safe and effective leading participants to revealed that supplement usage varied according to con- believe that supplements can’t hurt. Further, the unease sumers’ regulatory orientations (Higgins 1997). When related to not taking supplements could be attributed to probed on the role of supplements in “healthy lifestyle,” magical beliefs concerning “tempting fate” which make the complement group associated them with an eager adverse events such as future health-related problems strategy oriented toward advancement, betterment and more vivid and accessible (Rise and Gilovich 2008). This growth of a healthy lifestyle (promotion orientation). For aspect of “tempting fate” could explain why consumers example, consider Ryne’s narrative: “Supplements work rationally believe that supplements may not produce the hand in hand with eating healthy. They remind me to eat expected benefits but they continue taking them-for con- healthy and when I am eating healthy I feel better, so I feel sumers, supplements are the insurance to attain a healthy more like going to the gym and it all ties together. I think lifestyle (Tykocinski 2008). you can’t have one without the other.” Now consider Parya’s account: “I take supplements because these are The next step is to examine magical beliefs about supple- something that I need. I don’t do it to improve something ments within an experimental setting. We hypothesize about myself. I look at my lifestyle, I go to work, I go to that “natural” properties of dietary supplements is a unique school. So it is really hard for me to have my vegetables underlying essence, often associated with “natural kind” because I am always on the go. The best thing that I get categories (Gelman 1988). Previous research suggests over here is a hot dog because they don’t really offer that natural kind categories have a rich inductive potential

134 Marketing and Public Policy Conference Proceedings / 2009 that help individuals in predicting other properties based information when evaluating the effectiveness of supple- on this essence (McGill 1998). This may explain our ments (vs. OTCs). This would imply that consumers may qualitative findings that consumers draw inferences (of- judge supplements according to lower standards of evalu- ten incorrectly) that natural properties of supplements ations than OTCs. We will test this hypothesis in a second implies that they are safe and thus can’t hurt. We are study with a third study will test whether regulatory focus currently conducting a study to examine whether supple- acts as a moderator for these effects. ment users associate this group of products with magical powers – an underlying essence that is unique to this We contribute to the growing literature on remedy mar- category. Using over-the-counter medications (OTCs) as keting by identifying consumer segments based on style a control group, we expect that there is a greater likelihood of supplement use. Further, we contribute by integrating of associating special powers with dietary supplements literature on categorical thinking and regulatory focus (vs. OTCs). We also examine whether consumers predict theory with magical beliefs about dietary supplements. a greater number of properties related to safety based on Our research also has public policy implications. It pro- the fact that dietary supplements (vs. OTCs) are natural. vides insights into conditions under which magical beliefs Further, extant literature suggests that individuals are less may be maladaptive and accordingly may help regulators likely to consider necessity information for explanations in developing countermeasures. References are available based on natural categories. Drawing upon this, we argue upon request. that consumers should place less emphasis on necessity

For further information contact: Sayantani Mukherjee Marketing Department College of Business Administration California State University, Long Beach 1250 Bellflower Boulevard Long Beach, CA 90840–8503 Phone: 562.985.4256 Fax: 562.985.1588 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 135 A TEST OF OBJECTIFICATION THEORY ON FEMALE CONSUMERS’ PRODUCT EVALUATIONS

Chrissy Mitakakis, Baruch College, New York Stephen J. Gould, Baruch College, New York

SUMMARY like experiences for females due to strong associations with the common, accepted societal standard and ideal of It has been found that one of the most prominently the female body. Therefore, it was proposed that objecti- imposed social beliefs and values by the media has been fying advertisements would result in females expressing the gender stereotype, and more specifically, the objecti- a need to compensate for their lack of ability to meet the fied view of the female (Monro and Huon 2005). Western cultural standard and ideal. Thus, within the consumer media undoubtedly provides the ideal standard to which context, this may translate into females evaluating certain females compare themselves (McKinley and Hyde 1996). products that are considered for the public domain as more One particular theory, termed objectification theory, ar- self- and appearance-enhancing, hence exhibiting better gues that in Western culture and society, the female body evaluations and a higher willingness to pay (WTP) for is regarded as a sexual object, or simply “a body (or these “public” products. collection of body parts) valued predominantly for its use (or consumption by) others,” (Fredrickson and Roberts We examined several manipulations of objectification in 1997, p. 174). As a result of this “third person” point of a pretest, and assessed the extent to which different types view of their bodies, emphasis is placed upon the physical, of stimuli were considered objectifying. Based on our observable bodily characteristics instead of non-physical data, we found that advertisements with a generally sexual attributes that are part of the first person point of view. tone that included thin, attractive individuals were re- Consequently, these views come to be internalized by garded as the most objectifying stimuli, while stimuli those who are exposed to them, leading to a process depicting inanimate objects (i.e., a book advertisement) known as self-objectification. Accordingly, much evi- were considered the least objectifying. An objectification dence has been found to suggest that self-objectification focus exercise was considered the next most objectifying produces a host of negative consequences among women, manipulation, while a self-reflection exercise was consid- including shame, guilt, anxiety, body dissatisfaction, and ered the next most non-objectifying. In our first study, we a lack of awareness of internal bodily states (Moradi et al. utilized a sample of 81 students from a Northeastern 2005). Additionally, research has also suggested that university who were paid $5 for their participation. Par- these findings are limited to the female gender, and has ticipants were randomly assigned to one of four experi- failed to find any effects of objectification among males. mental conditions: two objectifying (objectifying adver- Given the fact that such objectifying media is proliferated tisement or objectification focus exercise) and two neutral in outlets such as product advertising, it is arguable that (book advertisement or self-reflection exercise). The prime females’ constant exposure to objectifying images and was followed by a “hypothetical shopping experience” in messages can produce unique effects on their consumer which a series of four products were individually evalu- behavior. In this research, we explored some of the effects ated in terms of quality, popularity, and WTP. Two of of advertisements on females’ consumer behavior, and these products were relatively private (stereo system and placed a particular emphasis on the role of objectification pillow) and two were considered more public, appear- in the process. In addition, we posited that objectifying ance-related products (jeans and tooth-whitening tooth- advertisements, defined as those in which a female body brush). Following the product evaluations, participants is depicted as merely an object or set of body parts, will answered the Self-Objectification Questionnaire (SOQ; increase females’ levels of self-objectification. As such, Noll and Fredrickson 1998), which measured their state of we also proposed that these increased levels of objectifi- self-objectification. The SOQ asks participants to rate the cation will lead females to pay more for items that serve importance of five appearance-related bodily attributes some type of public, appearance-related function. We and five competence-based bodily attributes. The higher expected such effects to be based on the negative state that the score, the more the individual is said to be self- self-objectification creates. Specifically, exposure to ob- objectifying. Finally, participants answered a series of jectifying stimuli has been shown to result in dissonance- demographic questions.

136 Marketing and Public Policy Conference Proceedings / 2009 The results confirmed our hypotheses regarding the ef- In conclusion, while our first study demonstrated signifi- fects of objectification on females’ consumer behavior. cant results, further research is needed to better assess the Firstly, the test of the four manipulations confirms the effects of objectifying advertisements on females’ con- notion that the objectifying advertisement led to a higher sumer behavior. Based on our findings, however, it is state of self-objectification than either the self-reflection evident that objectifying advertisements can elevate levels of self-objectification, which in turn can affect fe- exercise or the book advertisement (M = -1.30 vs. objad males’ consumer behavior. Therefore, it is imperative that M = -8.40 vs. M = -10.81; p < .05). Interestingly, bookad self-ref public policymakers take this issue into account when we did not find any differences of objectification scores making decisions about advertisement regulation. While between males and females (M = -5.41 vs. M = male female such regulation may be difficult to impose, however, -5.62, t(79) = .077, p > .60). However, we found a policymakers may consider educational programs ad- significant three-way interaction for condition, gender, dressing the negative effects of consumers’ exposure to and WTP for the “public” products (p < .04), suggesting objectification. In addition, a next logical step might be to that self-objectification operates within a complex matrix further examine the role of objectification on the extent to of effects, including gender differences. While our results which females engage in other types of consumption, provide important support for the notion that self-objecti- including high-risk behaviors like gambling, alcohol and fication can influence females’ consumer behavior, it is substance use, and reckless sexual behavior. References only a first step which warrants further research. are available upon request.

For further information contact: Chrissy Mitakakis Marketing Department Baruch College 1 Bernard Baruch Way, Box 12–240 B New York, NY 10010 Phone: 646.312.3418 Fax: 646.312.3271 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 137 ARE YOU ASHAMED? THE ROLE OF SHAME IN HEALTH COMMUNICATIONS

Kate Pounders, Louisiana State University, Baton Rouge

SUMMARY unhealthy behaviors (e.g., drug, tobacco, and alcohol use). Understanding when shame promotes healthy be- The author examines the role of shame in a health commu- haviors and prevents unhealthy behaviors will provide nications context. Society is faced with an array of daunt- consumer welfare advocates with insight into how to ing and severe health issues, many associated with shame. more effectively develop and implement health cam- It has been suggested that shame may prevent consumers paigns. This research outlines a conceptual model of from seeking diagnosis and treatment of particular health shame in a health communications context and identifies conditions (e.g., mental illness and sexually transmitted aspects of self (public vs. private) and regulatory focus diseases). Paradoxically, it has also been suggested that (prevention vs. promotion) as potential moderators. Im- shame may be an effective deterrent for engaging in plications for consumer welfare advocates are discussed.

For further information contact: Kate Pounders Department of Marketing E.J. Ourso College of Business 3119 A Patrick Taylor Hall Louisiana State University Baton Rouge, LA 70803 Phone: 225.578.8779 Fax: 225.578.8616 E-Mail: [email protected]

138 Marketing and Public Policy Conference Proceedings / 2009 TOWARD A DRUG FREE AMERICA: AN EVALUATION OF ANTI- DRUG PSA EFFECTIVENESS ACROSS SENSATION SEEKING AND GUILT PRONENESS

Karen Becker, The College of New Jersey, Ewing Rowena Briones, The College of New Jersey, Ewing Brian Keefe, The College of New Jersey, Ewing

SUMMARY Study

In 2002, the University of Michigan’s Institute for Social Subjects. One hundred and four high school and college Research announced the first drop in illicit drug use by aged men and women from central New Jersey partici- teens since 1994; while in 2007, it was reported that the pated in the study without compensation; 61 percent were number of eighth graders using illicit drugs had been cut female. in half since 1996 (Serwatch 2008). These statistics are mirrored by the Department of Health and Human Ser- Independent Variables. Three independent variables: vices (NIDA 2008). Although many factors interact to message style, sensation seeking and guilt proneness were cause this level of change, one of the drivers is likely to be tested. Message style varied from a positive emotional the proliferation of anti-drug public service announce- appeal, to a negative emotional appeal and a rational ments issued by the Partnership for a Drug Free America. appeal. Sensation seeking was measured via Zuckerman’s Sensation Seeking Scale, and guilt proneness was mea- Understanding that public service announcements pro- sured via the Test of Self Conscious Affect (TOSCA), vide information that potentially persuades consumers to developed by Tangney (1988). make more informed socially desirable choices, we examined three different for anti-drug PSAs created by the Procedure and Stimuli. Participants were first asked to Partnership for a Drug Free America. Specifically we complete the sensation seeking and guilt proneness scales. evaluate the impact of these messages on two individual They were then shown a series of PSAs, including the difference measures: sensation seeking and guilt prone- three of interest. Next, they were asked to respond to a ness. Additionally we are interested in understanding the series of questions about the announcements of interest, interaction between message-type and level of sensation including: clarity of message, emotional intensity, factual seeking and of guilt proneness – specifically, what mes- information, production quality, motivational impact (will- sage appeals are most effective with these populations ingness to act as advocated for by the ad), personal possessing these character traits. relevance to the issue, and ability to alter perception of issue (perceived effectiveness). Hypotheses Results The following hypotheses are proposed: Sensation Seeking. In contrast to predictions for sensa- H1A: High sensation-seeking individuals will be more tion seeking, we find that only the positive emotional motivated to act by PSA messages than low sensa- appeal has a differential effect in relation to level of tion-seeking individuals. sensation seeking. Although H1A is not supported we do H1B: High sensation-seeking individuals will be more find partial support for H1B in that the positive emotional motivated to act by emotional appeals than low appeal has a greater impact on high sensation seekers than sensation-seeking individuals. low sensation seekers, with high sensation seekers indicating a greater willingness to act in a manner advocated H2: High guilt prone individuals will be more moti- for in the PSA (which in this case was not to use drugs) vated by PSA messages than low guilt prone indi- (F = 3.639, p < .05), and a stronger sense of perceived viduals. effectiveness for the PSA (F = 5.856, p < .05).

Marketing and Public Policy Conference Proceedings / 2009 139 Guilt Proneness. As predicted by H2, we find that high of message effectiveness (F = 1.257, p > .1). We did find guilt prone individuals are more likely to be motivated to that in contrast to the high guilt prone individuals, low act in the manner advocated for by the PSA across all guilt prone individuals are more likely to perceive a message styles (F = 12.687, p < .001) and perceive the negative emotional message or factual PSA as having messages as more effective (F = 18.356, p < .001). Across greater clarity than a positive emotional message (F = high guilt prone individuals, we find the negative emo- 11.487, p < .001) tional appeals to be least effective in terms of motivation (F = 6.845, p < .05). When the message is a positive This study points to several directions for future research emotional appeal or a factually based message, high guilt and managerial applications. First, replication of the guilt prone individuals are more likely to perceive the message findings across other PSA contexts such as smoking or as having greater clarity (F = 8.374, p < .001; F = 6.40, p < gambling is warranted. Second, refining our understand- .01), are more motivated to act (F = 6.436, p < .01; 3.91, ing of the interactions between individual difference p < .05), and perceived the message to be more effective variables and message style is important. Managerially, (F = 6.436, p < .01; F = 4.637, p < .04). In contrast, with we demonstrate the importance of understanding the low guilt prone individuals, we find no significant differ- target audience and tailoring messages accordingly. Ref- ence across message style with regard to the PSA’s ability erences are available upon request. to motivate behavior (F = .579, p > .1), or alter perceptions

For further information contact: Karen Becker The College of New Jersey 2000 Pennington Road Ewing, NJ 08628–0718 Phone: 610.866.0901 E-Mail: [email protected]

140 Marketing and Public Policy Conference Proceedings / 2009 SHIFTING RISK-TAKING PREFERENCES TO RISK-AVOIDANCE PREFERENCES: THE EFFECTS OF MARKETING MESSAGES ON CHRONIC REGULATORY FOCUS

Jane McKay-Nesbitt, Bryant University, Smithfield Malcolm C. Smith, University of Manitoba, Winnipeg

SUMMARY the chronic regulatory focus of the individual, it was expected that chronic focus would be strengthened. Con- Introduction and Conceptual Development versely, it was expected that incongruence between the message and individual focus would weaken chronic Effective communication of risk information is particu- focus. larly important when marketing messages are designed to persuade consumers to avoid high risk behaviors such as Method and Results smoking or overeating (Block and Keller 1995). Regula- tory Focus Theory (RFT) has identified differences in the One hundred and twenty undergraduate university stu- extent to which individuals exhibit a chronic propensity to dents (61% males, Mage = 20.38) participated in the study. adopt risky behaviors in pursuit of their goals (Crowe and Message focus was a manipulated variable and chronic Higgins 1997). Promotion-focused individuals are will- focus was measured with the Regulatory Focus Question- ing to take risks and prefer an eager means of achieving naire (RFQ) (Higgins et al. 2001) as part of a battery of goals, whereas prevention-focused individuals are less other tests (Time 1). The RFQ provides a promotion and willing to adopt risky goal pursuit strategies. Thus, having a prevention score for each individual. Following Cesario, a chronic promotion focus may be an impediment to the Grant, and Higgins (2004), subjects were assigned to a avoidance of some harmful, risky behaviors. Conversely, chronic prevention or a chronic promotion group based on a prevention focus may serve an individual better when it a median split of the difference between their promotion comes to avoiding risky health behaviors. and prevention scores. Gender was included as a covariate because of evidence that females are more risk-averse There is evidence that an individual’s regulatory focus than males (Hersch 1996) and that a prevention focus is can be temporarily induced by external influences (Crowe associated with risk aversion. Three weeks later (Time 2), and Higgins 1997), such as marketing messages (Zhao subjects read either a promotion-focused or prevention- and Pechmann 2007). However, although much of the focused marketing message that pertained to exercise. current research on regulatory focus is based on indi- The promotion-focused message contained words such as vidual regulatory focus that has been primed, it is not “achieve benefits” and the prevention-focused message known how these situationally induced regulatory atti- contained words such as “avoid problems.” After reading tudes interact with chronic regulatory attitudes. For ex- the message, subjects completed the RFQ again to assess ample, it is not known if an individual who prefers risky the changes in chronic focus that had occurred as a result (risk-free) goal pursuit strategies will become less promo- of reading messages that were either congruent or incon- tion (prevention)-focused and more (less) risk averse gruent with their dominant focus. For those who were when exposed to marketing messages with a prevention promotion-focused, the variable of interest was only the (promotion) focus. Thus, this paper addresses the ques- promotion score and for those who were prevention- tion of how marketing messages that are framed with a focused, the variable of interest was only the prevention particular regulatory focus affect chronic regulatory atti- score. tudes. Separate repeated measures ANOVAs were run for each This study explored changes in chronic regulatory atti- group of subjects. Chronic promotion (prevention) scores tudes following exposure to marketing messages intended were treated as a within-subjects variable. The covariate to induce a particular regulatory focus. It was expected analysis revealed that females were significantly more that marketing messages framed with a particular regula- prevention-focused than males. Consequently, gender tory focus would shift chronic regulatory attitudes. When was included as a between subjects factor. However, there the regulatory focus of the message was congruent with were no effects for gender, regardless of whether mes-

Marketing and Public Policy Conference Proceedings / 2009 141 sages were congruent or incongruent (p’s > .10). More These findings offer assistance in designing and targeting importantly, the analysis revealed that when message marketing messages aimed at encouraging individuals to focus was incongruent with an individual’s chronic focus, avoid risky behaviors. They suggest that prevention- chronic regulatory focus was weakened. Paired-samples focused messages may be more effective than promotion- t tests showed a marginally significant reduction in pro- focused messages when marketing messages encourage motion-focused individuals’ promotion scores (MTime1 = risk-avoidance behaviors. When a message encourages

3.97, SDTime 1 = .39, MTime2 = 3.84, SDTime 2 = .49, t (24) = the avoidance of risky behaviors, promotion-focused 1.673, one-tailed p = .053) when they read a prevention messages that reduce an individual’s preference to adopt message, and prevention-focused individuals’ prevention a prevention focus may not be as effective as prevention- scores when they read a promotion message (MTime1 = focused messages that do not alter chronic prevention

3.70, SDTime 1 = .36, MTime2 = 3.53, SDTime 2 = .62, t (22) = attitudes. Similarly, prevention-focused messages that 1.71, one-tailed p = .051). Conversely, the analysis re- dampen a chronic promotion focus which is associated vealed no differences in chronic promotion or chronic with risk taking behavior may be effective when the goal prevention scores when message focus was congruent is encourage individuals to engage in less risky behavior. with chronic focus (p’s > .10). An individual’s chronic regulatory focus has been shown Conclusion to influence preferences for risk taking or risk avoidance behavior. Thus, understanding how chronic regulatory The finding that incongruence between the regulatory focus shifts in response to marketing messages is an focus of the message and the chronic focus of the message important development in our quest to understand the recipient weakened chronic focus is important as it points factors that contribute to the effective communication of to a possible mechanism by which an individual’s chronic risk information to consumers. References are available risk preferences can be shifted. It also points to a possible upon request. mechanism underlying regulatory fit effects on persuasion (Aaker and Lee 2006).

For further information contact: Jane McKay-Nesbitt Bryant University 1150 Douglas Pike Smithfield, RI 02917 Phone: 401.232.6940 Fax: 401.232.6068 E-Mail: [email protected]

142 Marketing and Public Policy Conference Proceedings / 2009 MOTHERS’ PERCEPTIONS OF WORD OF MOUTH MARKETING WHEN TEENS ARE USED AS BUZZ AGENTS

Bob D. Ahuja, Xavier University, Cincinnati Tara A. Michels, Xavier University, Cincinnati Clint S. Schertzer, Xavier University, Cincinnati Mary M. Walker, Xavier University, Cincinnati

SUMMARY However, through the use of WOMM communications, marketers have found a new way to actually involve the Background consumer in the marketing process for everything from assisting in the design of new products, to providing The concept of word of mouth marketing (WOMM1) relevant and timely marketing research, to generating communications and the spread of information about consumer driven advertisements that have high source products and services among customers have been of credibility. According to Christopher Van den Bulte interest to marketing professionals since the early 1960’s (Knowledge@Wharton 2007) “Marketers right now are (Rogers 1995) and continues up to present time.2 WOMM talking about becoming a meaningful part of customers’ occurs when a person who is knowledgeable or has an lives and less of an in-your-face presence.” WOMM’s opinion about a product or service, completely of her or “effectiveness is a function of creating advocacy and his own accord (i.e., unprompted by a third party) share amplifying the ability of advocates to spread the word” beliefs, attitudes, and experiences about a product or (Woehrle and Marks 2004). A recent global Nielsen service. Even in this computer age, we still like to talk in Internet study found “78 percent of respondents consid- person. Eighty percent of word of mouth conversations ered recommendations from customers as the form of take place in live dialogue compared to 20 percent online advertising they trust most” (Stores Magazine 2007). (Balter and Butman 2005). Most of these conversations, WOMM is here to stay. This relatively new marketing whether online or face-to-face, are assumed to be sincere, tactic is changing the way marketers target, listen and honest (Kaikati and Kaikati 2004), and spontaneous. involve consumers.

In the past decade there “. . . has been a meteoric rise for Objective of this Study word of mouth marketing” (McConnell 2007). WOMM spending increased from $76 million in 2001 to $150 In 2007 the authors of this study published an article in the million in 2005 (Horovitz 2005) to nearly $1 billion in marketing press providing the results of several profes- 2006; and, it is expected to grow to about $3.7 billion by sionally recruited and conducted focus groups of teen 2011 (McConnell 2007). In addition to hiring an ad boys and girls regarding their opinions and uses of WOMM agency to develop and deliver a WOMM campaign, a efforts aimed at teens. Building on this research, the marketer in 2008 has many additional choices of vendors authors of this article sought to uncover the opinions of that specialize in WOMM. From BzzAgent.com that mothers4 of teens regarding (WOMM). obtains WOMM feedback from all types of consumers, and Tremor.com that serves as a WOMM vehicle to target Methodology teens, to Vocalpoint.com that involves women, the WOMM industry continues its rapid growth. The data discussed in this paper comes from a national probability sample of 416 mothers with at least one teen Many in the marketing industry believe there is a trend at living full time at home5 conducted in the spring of away from traditional media (tv, radio, print) to the ever 2008. It should be noted that before moms were asked growing forms of nontraditional media3 (Kovacic 2008). about buzz marketing, they first were asked “How famil- Historically, traditional media has engaged the consumer iar are you with Buzz Marketing?”6 Immediately follow- in a passive way, i.e., the consumer really didn’t have to ing this question moms were given two definitions, one do anything other than minimal behavior such as turn a for WOMM7 and one for Buzz Marketing.8 page, or turn the channel on tv, or turn to a favorite radio station. The consumer has been only marginally involved Possibly indicating a level of high interest in this topic, we in these traditional advertising media. received a 90 percent response rate to the open-ended

Marketing and Public Policy Conference Proceedings / 2009 143 question “Please type in the space below what thoughts ents, and customers always have the right to say no to a you have about buzz marketing”(this was question 10 of buzz campaign,” which is correct as buzz agencies do give 20 in the questionnaire). The responses to this question agents the choice to buzz products they have interest in were given to a local full service marketing research firm and turn down products they don’t feel like doing. for analyses. The questions were uploaded into a proprietary vendor’s server for analyses. The vendor’s software In the unfavorable category the most prevalent responses established a hierarchy of phrases and words found in the were “not a good idea,” “I don’t like it.” Thirty respon- responses. dents thought this type of marketing was “deceptive” or “dishonest,” “exploitive of teens,” may “brain wash or Nine patterns emerged from this process and these nine corrupt teens” and that it is “not appropriate for teenag- areas were used in the coding of every response to ques- ers.”12 Others thought that buzz marketing places “too tion ten. The nine areas of analyses included three general much peer pressure” on teens and that there “already is categories, Favorable, Unfavorable, and Neutral and within enough marketing to teens” and “no additional market- each of these categories, three sub-groups were delineated ing” to this age group is “necessary.” Some mentioned (1) respondent thoughts on the concept itself, (2) respon- that buzzing was “predatory,” and relied on friends “us- dent comments regarding the efficacy and outcome of ing” friends. Other respondents noted that buzz marketing such marketing tactics, and (3) general miscellaneous has the potential to be “difficult for parents to monitor,” comments. Responses were coded with as many specific and that there were “no guidelines to follow to make sure codes as were necessary to capture all the opinions being it is not being misused.” Also, those in this category said offered. Responses had the possibility of having numer- buzz marketing may “cause teens to make decisions or do ous code combinations, e.g., favorable/unfavorable, un- things they should not do or really don’t want to do.” Some favorable/neutral and so on. Specific codes within the with unfavorable comments thought this marketing tactic nine coding categories were created when two or more was “too secretive, too subtle, and not upfront enough.” respondents offered the same information. For example, three respondents indicated that buzz marketing was new/ Unexpectedly, most responses were in the neutral cat- different/unique. Therefore “new/different/unique” be- egory. Some moms said that “parental consent” should be came a code in the favorable concept category. required so parents could “monitor” their children. Those who responded in the neutral area mentioned that they Results were “Not familiar enough with the idea of buzz marketing,” or that they “needed more information to form an In general, 1159 or 31 percent of the responses were in opinion.” Others whose comments were coded as neutral whole or part favorable of the concept of buzz marketing, made statements such as “it depends on if it is moral,” if with 94 or 24 percent of the responses containing in whole it is “honest,” and not “deceptive” or “misleading.” Some or part unfavorable comments, and 209 or 56 percent of noted that “young people need to be taught to be smart the responses neutral. consumers and use information wisely,” while others felt “confident that the child will make the right decision The most prevalent comments offered by respondents in regarding buzz marketing.” the favorable category contained statements such as “sounds good,” “a good idea,”10 “I like it,” “sounds cool,” Conclusions and “seems like fun.” Others mentioned that this was a “good marketing tool,” that is was “beneficial to those Buzz marketing may be the perfect way to market a involved,” and that promoting products through buzz product, i.e., consumer-to-consumer recommendations. marketing offers “first hands on experience in using a When this process is enhanced and commercialized by product to see if you like it.” In some cases buzz agents are hiring and training teen buzz agents, the teen takes on given product samples or coupons to obtain a free product many marketing roles, she/he is an endorser, a distributor, and several other similar coupons for their friends, par- a researcher, an influencer, and, of course a consumer ents, or themselves.11 Some thought that providing free too.13 Those who liked buzz marketing saw it as educa- product to the buzz agent would give the agent a “good tional tool and a way to expose teens to the real world. incentive to buzz,” and therefore would serve as “pay or However, given that the majority of responses contained compensation for the job.” Some made statements that neutral and unfavorable comments, it may be concluded reflected the real purpose behind buzz marketing, i.e., it that the efficacy of buzz marketing using teens as buzz “allows people to get an honest opinion from a real agents remains a marketing tactic that is of questionable person.” Some respondents recognized that “teens, par- moral and practical value to many mothers of teens.

144 Marketing and Public Policy Conference Proceedings / 2009 ENDNOTES 6 A five point scale was used with 1 representing “Not at all,” 3 anchored to “Somewhat Familiar” and 5 1 WOMM has been labeled many things. For example, representing “Very Familiar.” WOMM has been called buzz marketing, viral mar- 7 “Word of Mouth Marketing” is the act of consumers keting (the term used to refer to WOMM in the virtual sharing opinions and information about products and or computer world), social networking marketing, services. advocacy marketing, to name a few. 8 “Buzz Marketing” is the practice of gathering volun- 2 The Journal of Public Policy and Marketing, Spring teers to try products, then sending them out into the 2008, Volume 27, Number 1, contains an a special world to talk up their experiences with the people section on covert marketing including four articles them meet in their daily lives.” and three essays and commentaries. 9 Some double counting occurred as responses could 3 Such as company sponsored websites, internet adver- be in multiple classifications. tising, packaging and labeling, movie theater/video 10 Quotes around comments in this section mean the gaming advertising, digital advertising such as email comment was the exact word responses to Q10. and text messaging, in-store marketing, public enter- 11 These coupons could be handed out to friends and tainment events, premium distribution, product place- associates of the buzz agent, and/or could be used by ment, character licensing, cross-promotions and co- the buzz agent to acquire free merchandise. branding, sponsorship of sport teams or athletes, 12 Disclosure is NOT required by all buzz agencies. word-of-mouth marketing, viral marketing, in- Many buzz agencies, e.g., Tremor.com that uses school marketing, advertising with philanthropic teens, leave the choice to disclose up to the buzz agent purposes. themselves in each buzz experience or situation. 4 These mothers were not related to the teens men- 13 These roles were discussed more fully in the article tioned in the 2007 piece on WOMM among teens. Ahuja, Michels, Walker. and Weisbuch 2007, “Teen 5 The data was collected using an online questionnaire Perceptions of Disclosure in Buzz Marketing,” Jour- sent out to mothers in the Zoomerang.com database. nal of Consumer Marketing, 24 (3), 151–59.

REFERENCES (Summer), 6–22. McConnell, Ryan (2007), “Word-of-Mouth Worth $1 Balter, Dave and John Butman (2005), The Grapevine. Billion,” Advertising Age, 78 (46), 4. London, England: Penguin Group, 57. Knowledge@Wharton (2007), “Social Marketing: How Kovacic, William E. et al. (2008), “Marketing Food to Companies Are Generating Value from Customer Children and Adolescents: A Review of Industry Input,” December 12, retrieved November 16, 2008 Expenditures, Activities, and Self-Regulation,” A from [http://knowldedge.wharton.upenn.edu/ Report to Congress, Federal Trade Commission, article.ffm?articleid=1864, page 2]. Washington, D.C. Rogers, Everett M. (1995), The Diffusion of Innovations, Horovitz, Bruce (2005), “P&G ‘Buzz Marketing’ Unit 4th ed. New York: The Free Press. Hit with Complaint,” USA Today, October 18, re- Stores Magazine (2007), “It’s a Matter of Trust,” 89 trieved January 4, 2006 from [http:// (November), (11), 18. www.usatoday.com/money/industries/retail/2005- Woehrle, Ted and Brian Marks, (2004), “Building Word- 10-18-buzz-usat_x.htm?csp=34 page 3]. of-Mouth Advocacy for Your Brand,” Presentation Kaikati, Andrew M. and Jack G. Kaikati (2004), “Stealth at a conference held in Cincinnati, Ohio, (January) Marketing: How to Reach Consumers Surrepti- 16. tiously,” California Management Review, 46 (4),

For further information contact: Bob D. Ahuja Xavier University 3800 Victory Parkway Cincinnati, OH 45277 Phone: 513.745.3000

Marketing and Public Policy Conference Proceedings / 2009 145 THE ROLE OF VALUES IN CAUSE-RELATED MARKETING: A THAI PHARMACEUTICAL INDUSTRY STUDY

Apichai Apirattanapimolchai, Southern Cross University, Australia Peter Vitartas, University of Lethbridge, Alberta

SUMMARY part of their business strategy (Pfizer Thailand Founda- tion 2007; Raks Thai Foundation 2008). This paper reports on a study being undertaken in Thai- land that explores the effect of values on the purchase Cause-related marketing programs have been found to intention and corporate image of a pharmaceutical com- positively influence consumers’ attitudes to purchase pany undertaking a cause related marketing program. intention, purchase decision, brand choice, and brand switching behavior (Barone, Miyazaki, and Taylor 2000; In recent years, with people becoming more concerned Chaney and Dolli 2001; Rose III, Stutts, and Patterson about world social issues, there has been a growing 1991; Sen and Bhattacharya 2001). Furthermore, cause- interest from corporations, academics and the community related marketing programs positively influence custom- in the use of cause-related marketing. Responding to the ers to try new brands (Rose III et al. 1991) and stimulate public interest has led many companies to express their more purchases than cash rebates (Strahilevitz and Myers social responsibilities via different causes which aim to 1998). provide benefit to the overall community (Kotler and Lee 2005; Van den Brink, Odekerken-Schröder, and Pauwels Demographic and psychological factors also have an 2006; Webb and Mohr 1998). important role in shaping the consumers’ purchase intentions. Two important psychological factors identified Of particular interest in the field has been the relationship from the literature, of relevance to cause-related market- between the three entities; the customer, the cause and the ing, are self-image and values (Hawkins, Best, and Coney company. Findings from previous studies reveal that most 2001; Widing et al. 2003). By matching customers with a consumers express a favorable attitude toward cause- particular cause, a company can build cognitive and related marketing campaigns because they are a good way affective components through the corporate identity in of fund raising (Chaney and Dolli 2001) and as a means to their cause-related marketing programs. This allows for a help their society (Barnes 1992; Kropp et al. 1999). At the congruency between themselves and the cause (Gupta corporate level cause-related marketing helps firms en- and Pirsch 2006). hance their reputation by communicating to customers the causes they have initiated and provided support and Values have been defined as enduring beliefs that a resources to and as a result, a higher corporate image is specific mode of behavior or end-state of existence is expected (Barone, Norman, and Miyazaki 2007; Chaney preferable to opposite modes of behavior or end-states and Dolli 2001; Docherty and Hibbert 2003; Van den (Bearden and Netemeyer 1999; Kropp, Holden, and Lavack Brink et al. 2006). Having a cause that fits neatly with the 1999). Values are factors that shape individual self-im- company and one that relates to customers, means stron- ages and provide underlying motives that have an impact ger positive relationships can be developed (Gupta and on shaping consumers’ attitudes and behaviors (Kropp Pirsch 2006). et al. 1999; Widing et al. 2003). Consequently, proposing values consistent to ones’ self images can fill the gaps One area where there is a close fit between companies and between consumers’ self image and product or brand their causes is in the health area. Companies in the personality/image (Heath and Scott 1998; Renolds 1965; pharmaceutical industry, for example, have adopted cause Widing et al. 2003) facilitating the purchase decision. related marketing into their marketing practice. Cross promotion between pharmaceutical companies and car- Customers attitudes toward cause-related marketing pro- ing causes provide a natural alignment and is evidenced grams have been found to differ depending on their by the practices of many companies such as Bristol- values, in particular, the personal values of warm relation- Mayers Squibb and Prizer (ABPI 2008; Envision Solu- ships, self-fulfilment and security have more positive tions LLC 2007; PR Week 2004). In Thailand, many drug attitudes toward cause-related marketing (Kropp et al. companies have also adopted cause-related marketing as 1999). Furthermore, customers who are more predisposed

146 Marketing and Public Policy Conference Proceedings / 2009 to charitable giving are more likely to think highly about kets 2008). It depends largely on the import of pharma- cause-related marketing campaigns and be more support- ceutical products from multinational companies (Drug ive of these programs as part of payments going to charity. Control 2003) who command a major share of the market To date no studies have examined the effect of values on and receive significant financial and marketing support purchase intention. from their parent companies (Chan and Ching 2005). As a result, it is argued, they have the ability to utilize more The present study will examine the effect of values, as sophisticated marketing activities than local companies measured using the List of Values (LOV) (Bearden and (Mansfield 2007). As consumers purchase many of their Netemeyer 1999) on purchase intention for a cause related medicines over the counter they have greater choice than marketing campaign. Control will be maintained for the consumers in countries where prescriptions are used. company-cause fit, customer-cause fit and corporate im- Pharmaceutical companies are adopting cause-related age. It is expected that values will impact on purchase marketing to build their corporate image and encourage intention regardless of the controlled variables. switching to their brand. The study that is currently being administered in Thailand. The study is being undertaken in Thailand as it is an important pharmaceutical market in Southeast Asia be- Results from an initial pilot test of the questionnaire cause of its relatively high level of development, educa- confirm the importance of company-cause fit on corpo- tion levels and acceptance of western pharmaceutical rate image and purchase intention, and that values are an products (Business Wire 2008; IMS Health Incorporated important moderator of consumers’ purchase intention. 2006; PricewaterhouseCoopers 2008; Research and Mar- References are available upon request.

For further information contact: Peter Vitartas University of Lethbridge 4401 University Drive Lethbridge, Alberta T1K 3M4 Phone: 403.329.2015 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 147 A COMPREHENSIVE ANALYSIS OF CONSUMER TARGETING

Thomas A. Klein, The University of Toledo and TUI University

SUMMARY practice as competitors focus their attentions on the same most attractive segments in terms of access and buying Targeting has long been recommended as a mutually power. As a result, for example, rural or lower income beneficial marketing strategy for both consumers and consumers tend to be ignored, deprived of goods and sellers. This paper examines that recommendation, chal- services they might prefer or forced to travel farther or pay lenges it under some circumstances, and considers some more to obtain what is more in line with the preferences of broader social implications of this strategy. consumers in metropolitan areas or with higher incomes.

Whether pegged to benefits sought, psychographic or life The exploitation aspect of targeting is too often carried to style characteristics, behavioral patterns, or demograph- vulnerable populations – children, elderly, socially imma- ics, consumers are more likely to obtain the products and ture, or suffering from disease – in promising benefits that services they prefer in terms of price levels, qualities, and cannot be realized (except, perhaps, psychologically), but convenience through marketing programs that focus on that have special appeals that correspond to their vulner- those preferences, which tend to be segmented, than from abilities. programs that approach potential buyers as if they were homogeneous in their wants and needs. The emergence of unobtrusive data gathering technolo- From a seller’s perspective, targeting reduces buyer resis- gies during the past decade – e.g., spyware, CRM technol- tance, promotes customer loyalty, and allows for the ogy, and surveillance devices – raises important new optimum exploitation of value as consumers are likely to questions, particularly for those in the research and com- pay more for the products and services that most closely munications industries as well as for marketers that sub- fit their needs. Targeting also serves as the basis for scribe to their services. Even when the application of competitive strategy as marketers seek out underserved findings is to large groups, gathering personal informa- and emerging segments with a potential interest in at- tion in the absence of informed consent may be viewed as tributes that differ from or are superior to those available a violation of privacy. As a platform for unwanted promo- from alternative sources. While targeting sometimes im- tional messages via spam, pop-up ads, and direct mail, poses offsetting costs in production and logistics, depend- these technologies pose a consumer cost in nuisance and ing on available technology and scale economies, these inconvenience that can be quite significant. When those may be less than the benefits noted above and, in any applications are individualized in so-called “one-on-one event, have provided incentives for cost-reducing opera- marketing,” these effects can be especially troubling. In tional developments such as flexible manufacturing and extreme examples, those costs can be quite tangible, e.g., containerization. price discrimination or limited choice.

Finally, when examined at the industry level, the presence These “clouds” amount to externalities, spillover or social of multiple targets is assumed to encourage specialization costs that are difficult or impossible to control by consum- across industry members such that all meaningful seg- ers who are disadvantaged or violated as a result. Accord- ments enjoy the benefits of this strategy. ingly they raise both ethical and, particularly when the populations affected are large and defenseless, public Altogether, this appears to be a “win-win” development. policy questions.

A more extensive examination of this phenomenon, how- A standard response to these issues is a call for greater ever, particularly in the context of contemporary commu- vigilance by consumers and, it seems fair to observe, the nications technology, discloses some troubling clouds in public disclosure of these practices has probably encour- this otherwise sunny appraisal. aged that. The ethical perspective calls for marketers to abstain from practices that exploit consumer vulnerabili- The assumption that various industry members will target ties and/or to be more proactive in their information the range of meaningful segments is regularly violated in policies and product and distribution strategies.

148 Marketing and Public Policy Conference Proceedings / 2009 Nonetheless, the continued presence both of consumers regulation, boycotts, information services or mandates, who are unable to exercise adequate vigilance and of rules requiring or prohibiting certain practices, subsidies predatory marketers that pursue their economic interests that encourage improved access or affordability, manda- at the expense of these vulnerable consumers raises the tory redress provisions, and civil or criminal penalties question of whether additional measures are called for. calibrated to respond to the severity or duration of the cost Such actions might include – depending on the particular experienced. practice, population affected, or injury – industry self

For further information contact: Thomas A. Klein The University of Toledo MS # 103 Toledo, OH 43606–3390 Phone: 419.530.4368 Fax: 419.530.4610 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 149 TARGETING DIFFERENT SEGMENTS OF COLLEGE STUDENT SMOKERS: PRELIMINARY RESULTS AND PROPOSED EVALUATION STRATEGIES

William D. Diamond, University of Massachusetts, Amherst

SUMMARY Main Study

Bandura (1977) predicted that success raises self-efficacy Data were collected online from 766 students. Of these, and failure lowers it. Self-efficacy is very similar to 161 (21%) had smoked at least one cigarette during the perceived behavioral control (Schifter and Ajzen 1985). past two weeks. When asked to try a short cessation The present research designed a trial cessation programs attempt, 112 (70%) agreed. Of 106 (66% of all the so that smokers in different segments tried to quit for long smokers in the sample) completed both the first data enough to raise their perceived behavioral control, but not collection session and a second session approximately a so long that the attempt failed. These quit lengths were week later. different for different segments of college student smokers. Based on the pretest results, the website randomly requested that student smokers try to stop smoking either for Pretest two days or for a week. The incentive for trying (not necessarily succeeding) was entry into a $100 lottery. The study was conducted online. Of the 389 students in the original sample, 96 (25%) had smoked at least one Results cigarette in the previous two weeks. These student smokers were offered an opportunity to try a limited cessation. Of the 29 lighter smokers who tried a two-day cessation, 27 reported success. In contrast, only 13 of 26 heavier Reflecting the relatively low amounts that most college smokers reported being able to quit for two days. This students smoke, the 42 participants who had smoked difference was statistically significant ( 2(1) = 10.76, p < between one and ten cigarettes in the past two weeks were .001). Of the 26 lighter smokers who tried a one-week classified as lighter smokers. The 54 participants who had cessation, 12 were successful. Only one of the 27 heavier smoked more than ten cigarettes in the past two weeks smokers who tried a one-week cessation reported being were classified as heavier smokers. successful.

A Motivated Trial Cessation Attempt A significant triple interaction effect (supported by analyses of simple interactions and simple main effects) indi- Hypothesizing that quitting for a limited time might allow cated that a two-day quit attempt increased perceived smokers to feel successful without achieving permanent control for heavier smokers and a seven-day quit attempt cessation, the researchers used a web page to ask smokers increased perceived control for lighter smokers (F(3,100) = to consider a one-week cessation attempt in return for 4.39, p < .01). entry in a $100 lottery. Approximately a week later, participants reported their success and answered other Conclusion questions during a second online session. The results indicate that proper short cessation requests A total of 69 student smokers participated and subse- can increase the beliefs of college students that they will quently completed the second session of the research be able to stop smoking in the future. If the cessation study a week later. Of these, 31 were lighter smokers and request is too long or too short, it will not increase 38 were heavier smokers. perceived control over long-term cessation.

The median quit length for the heavier smokers was three Dependent Variables with Greater Policy Implications days. The median quit length for heavier smokers was one One of the common criticisms of “laboratory” smoking week. The results were used to design the main study. studies is whether they really can cause behavioral change.

150 Marketing and Public Policy Conference Proceedings / 2009 A reasonable way of evaluating future studies of trial antismoking counseling, or other programs that have been cessation attempts would be to evaluate whether they can shown to increase cessation. lead participants to “self-qualify” for active quitlines,

For further information contact: William D. Diamond Marketing Department Isenberg School of Management University of Massachusetts, Amherst Amherst, MA 01003 Phone: 413.545.5671 Fax: 413.545.3858 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 151 ABERRANT SERVICE CONSUMER BEHAVIOR AND COPING

Sally A. Hibbert, Nottingham University, United Kingdom Maria G. Piacentini, Lancaster University, United Kingdom Margaret K. Hogg, Lancaster University, United Kingdom

SUMMARY ters are often unsuccessful due to a combination of resource constraints, restricted consumer socialization and Introduction and Purpose the insensitivity of service providers to the resources and limitations of particular groups (Penaloza 1995; Lee et al. Lack of access to goods and services has long been 1999; Kaufman-Scarborough and Baker 2005; Adkins recognized as a major cause of consumer vulnerability and Ozanne 2005; Viswanathan et al. 2005). because it restricts consumption (Andreasen 1975; Baker et al. 2005). However, existing literature has largely The purpose of this study is to explore how vulnerable concentrated on access to services and the barriers to the young consumers’ navigate the consumption environ- take-up of services (Scammon et al. 1995; Franzak et al. ment and maintain access to services as they adapt to the 1995; Baker and Kaufman-Scarborough 2001). Rather roles and responsibilities of adulthood. less attention has been paid to the consumption of extended services and particularly to the restrictions that Research Design consumers face when trying to maintain service access. The objective of this research is to explore the restrictions Qualitative research was undertaken with 12 young moth- faced by vulnerable young mothers as they try to navigate ers who have faced early transitions to adulthood. Infor- the consumption environment, forging and maintaining mants were recruited through charities that support disad- access to services in their consumer lives in the transition vantaged young people. All informants had moved out of to adulthood. their family or foster home during their teens; most had lived in several different types of accommodation since Smooth Service Encounters – Enduring Service Ac- leaving home. All were aged between 17 and 22 years. cess Two informants were pregnant with their first child at the time of the fieldwork. Ten informants had either one or The service quality and customer satisfaction literature two children below the age of three years. has paid considerable attention to the issue of service failure (Krepapa et al. 2003). Aside from the research into Our research design combined focus groups and long the potential threats to service provision represented by interviews. Of the 12 informants, four participated in a dysfunctional consumer behavior, little attention has been focus group and an interview; four women were only devoted to how customers’ performance can threaten interviewed; and four only participated in a focus group. service encounters (e.g., Reynolds and Harris 2005; Ful- Informants were encouraged to recount their personal lerton and Punj 2004). Effective consumer participation histories and to reflect on the restrictions and challenges in service encounters is vital in realizing service quality that they had encountered in household production and and satisfaction (Bitner et al. 1997), and there is growing consumption roles, concentrating particularly on the peri- interest in the potential of consumer education to develop ods since they had left home. The analysis sought to customer capabilities (Eisingerich and Bell 2006, 2008; identify key themes, according to the methods outlined by Ozanne and Murray 1991) in order to anticipate and Corbin and Strauss (1990). manage disruptions to on-going service relationships. Key Findings Vulnerable consumers are those that experience “a state of powerlessness that arises from an imbalance in market- A common situation for our informants was to manage to place interactions . . . when control is not in an individual’s successfully secure access to a service initially, but to later hands, creating a dependence on external factors to create face the service being withdrawn because they had not fairness in the marketplace . . . the experience affects adhered to the conditions/contractual agreement of the social and personal perceptions of the self” (Baker et al. service. Services where loss of access had occurred were 2005). For vulnerable consumer groups, service encoun- typically those extending over the medium to long-term

152 Marketing and Public Policy Conference Proceedings / 2009 including financial services, utilities, telecoms, and ac- ers (e.g., emotional outbursts) tended to be ineffective in commodation; but also transaction-based services such as resolving problem. grocery retailing. Our analysis focuses on the key themes relating to why access to these services had suffered Coping. Informants primarily adopted problem-focused disruption (perceived lack of control and lack of “know- strategies to cope with threats to maintaining service how”); and the ways that the young people responded to access, with social ties having an important (positive) their difficulties (coping and the impact of coping). impact on coping. Many discussed their past and present social networks, with family and professional support Perceived Lack of Control Over Circumstances. Lack of workers being the most important sources of social sup- control over circumstances was described in terms of port. All informants were in contact with at least one financial restrictions and the informants’ social environ- member of their birth family, usually female (e.g., mother, ment. Financial hardship was particularly severe when grandmother, sister), who provided valuable support. they first moved to independent living, but also during Examples of overt support included sisters lodging to- volatile periods of their transition to independence, e.g., gether and sharing housekeeping costs; a grandmother when suffering from drug addiction, abuse, relationship advising and ensuring attendance at debt counseling ses- break-up, becoming unemployed or having a baby. Many sions; and a father accompanying his daughter to negoti- reported that they had defaulted on payments for rent, ate her payments arrears at the housing office. Families utilities and debt during such periods. These examples were also an important source of information and advice illustrate the importance of change and transition as ante- about entitlements and about how to approach service cedent of consumer vulnerability. recovery.

Two features of the informant’s social environments All informants had access to a professional support worker made it difficult for them to take control in maintaining from a charity as well as a state social worker, and these access to service: firstly, their peers and secondly, crimi- people featured prominently in most stories, signaling a nality in their neighborhoods and the wider environment. heavy reliance on this source of social support. Those who One informant was banned from her local supermarket for drew on social support moved from a position of relying shoplifting for a friend’s goods “X asked me to get some on others to one of seeking guidance. Some informants dog food. I knew I wouldn’t get the money back, so I took coped with service consumption independently, shifting it” (Kelly). Consequently, she was banned from this retail over time from seeking empathy from service providers to outlet. Several informants told stories about people be- more self-reliant solution seeking. There was substantial friending them when they moved into a neighborhood, variance in the rate at which individuals moved toward subsequently using their apartment as a meeting place, greater consumer empowerment. Noticeably, when social often causing criminal damage, leading to the informants support was provided by trained adults (rather than fam- being in contravention of their tenancy agreements. ily/friends), the informants’ consumer education was achieved through coaching rather than modeling, acceler- Lack of “Know-How.” Lack of consumer expertise threat- ating their empowerment. ened continued access to many services. The informants related stories in which they had faced the threat or reality of losing their accommodation due to their lack of skills The informants’ capacity to cope with the demands of and expertise in the main consumer and household pro- service consumption typically reflected how (dis)orderly duction roles necessary for securing and maintaining their or complicated their lives were at particular periods of accommodation (including budgeting; paying bills and their transition to adulthood (as discussed earlier). During rent; cleaning, maintenance, and repair). such periods the informants faced more problems in relation to services and coped with these less effectively. Consumer competence and good interpersonal skills were important when the service had suffered some disruption Impact of Coping. Effective coping enabled the infor- and the two parties were involved in service recovery. mants to move, albeit slowly, toward greater stability in Informants possessed varying degrees of mastery over their lives and to feel more empowered to cope with adult consumer skills, sometimes enabling them to navigate the roles. This improved their potential to address their so- marketplace effectively (e.g., negotiating an up-grade on cially and economically marginalized circumstances and a SKY TV package). Successful episodes were important that of their children. Examples of such improvements in in terms of the practical consequences and the psychologi- their lives included: moving to better neighborhoods; cal benefits. In contrast, those who were less knowledge- accessing nursery places, enrolling for college courses; able and skilled in their interactions with service provid- and obtaining and keeping a part time job.

Marketing and Public Policy Conference Proceedings / 2009 153 Where episodes that threatened access to services were hood more chaotic and stressful with important public not successfully resolved severe disruptions to lifestyles policy consequences for these consumers’ financial and resulted. These disruptions often made the transition to social marginalization and also for their children. Earlier adulthood more disorderly and stressful, with similar studies of consumer vulnerability have provided valuable consequences for their children, and negative impacts on insights into coping and resistance strategies (Cornwell the family’s financial and social marginalization. and Gabel 1996; Ozanne et al. 1998; Adkins and Ozanne 2005; Viswanathan et al. 2005). This study adds to this Conclusion work by illustrating a developmental approach to learning coping strategies that characterizes consumer transitions. Our study highlights the importance of being able to This research thereby adds to the literature on the limita- navigate the consumption environments and maintain tions that vulnerable consumer segments face in the mar- access to services in order to meet basic needs and avoid ketplace (Hill and Stephens 1997). References are avail- grave disruptions to lifestyle and livelihoods. Threats to able upon request. these service relationships make the transition to adult-

For further information contact: Sally Hibbert Head of the Marketing Division Business School Nottingham University Jubilee Campus Wollaton Road Nottingham NG8 1BB United Kingdom Phone: 0115.846.7653 E-Mail: [email protected]

154 Marketing and Public Policy Conference Proceedings / 2009 COULD WE HAVE SEEN IT COMING?: ON THE PROSPECT OF MINING ONLINE CONSUMER REVIEW DATA IN ORDER TO FORECAST PRODUCT SAFETY ISSUES

Gavin Huntley-Fenner, Exponent Inc., Irvine Miles Munro, University of California, Irvine, Exponent Inc., Irvine

SUMMARY useful by other customers? (3) To what degree could such safety information, if available, provide an “early warn- Introduction ing” of magnet ingestion hazards?

For toy manufacturers in United States, 2007 was the Methods “year of the recall” culminating in a new law (The Con- sumer Product Safety Improvement Act of 2008, “CPSIA”) Our investigation focused on safety related reviews in the that substantially overhauled the Consumer Product Safety 1999 to 2008 time-frame and we focused on toys that had Commission (CPSC). Among the reforms, was the estab- been recalled between March 2006 and April 2008 (CPSC lishment of a CPSC incident reporting database to support “Magnetix” recall on 3/31/06 to “magnetic dart boards” a consumer alert early warning system. Data could come on 4/8/08). Specifically we were concerned with toys that from consumers, government agencies, health care pro- were recalled due to concerns of oral ingestion of power- fessionals, child service providers, etc. and would be ful magnets. Data were collected from a well-established, publicly accessible, searchable, and available on the retail website which markets a wide range of products. internet. There are to be processes in place for verifying This site features a robust customer review system that is and amending data and for including manufacturer re- used on a regular basis. Other researchers have shown that sponses. Legislators’ and policy makers’ expectations for the frequency, length, and quality of reviews on this site the database are that it should “rapidly provide consumers have a positive impact on book sales relative to competi- with ‘early warning’ information about specific products tors. that could pose serious safety hazards.” To conduct the analyses, we programmed a “bot” using This database component of the CPSIA is modeled after Matlab. A bot is automated software designed to quickly United States government databases such as those con- perform a large set of actions over a computer network (in trolled by the Food and Drug Administration and the this case the internet) which would otherwise take a Department of Transportation. Those databases collect human operator much longer to perform. The bot was information from a variety of sources including individual devised to exhaustively query the retailer’s website from consumers and professionals. However, since consumer an inventory of recalled products and retrieve all customer awareness of health and safety information sometimes reviews for those products. The search inventory was follows mass-media publicity trends, there is a question of composed of recalled magnet toys (e.g., “Magnetix”) and whether consumer-sourced data regarding toy safety could general product categories (e.g., “magnetic building set”). serve as input to an early warning system or whether they After every product or review retrieval, the bot parses the are likely to be lagging indicators of hazards. To try to hypertext to extract the data and then cross-indexed and answer this question we examined customer reviews on a stored that data. All queries were submitted within the major online retailer website for information about oral retailer’s “Toys & Games” category of the website. ingestion hazards relating to toys containing small magnets. Online retailers collect data about products from customers on a voluntary basis. While most reviews are Analysis and Results focused on usability and product effectiveness, reviewers are also free to post product safety-related issues. Our To prepare for the analyses, the data were pruned for specific questions were the following: (1) Would online duplicate entries. Next, the review title and text data were consumer reviews on a site that does not specifically ask automatically searched and flagged for oral ingestion for safety information, nevertheless, contain such infor- terms (e.g., “swallow”) and general safety terms (e.g., mation? (2) Would the safety information be perceived as “danger”) from a compiled inventory of 40 terms. These

Marketing and Public Policy Conference Proceedings / 2009 155 “flagged” review texts or titles were then individually significant deviations from the expected probability dis- read and screened to eliminate any false alarms that might tributions. occur due to alternate usage of the hazard words (e.g., “die” as in death versus dice) and also to identify flagged The predictive foreshadowing of the recall(s), and thus reviews that pertained specifically to concerns about oral early warning ability, of the review system is best assessed ingestion hazards. The data included the review text, plus by examining the antecedence of the flagged reviews and overall product rating as well as product durability, edu- their characteristic density with respect to time. The cational, and “fun” as well as review text string length, analysis showed that several generic non-MAG safety- how often reviews were read (i.e., reported as specifically related reviews preceded the first recall (“magnetix”) with helpful or unhelpful). the earliest coming 896 (10/17/03) and 865 (11/17/03) days prior to the first magnetix recall. The earliest MAG A total of 1,459 magnetic toy products were identified, review preceded that recall by a more modest 87 (1/3/06) within that group 270 SKUs were reviewed a total of days. Similar patterns were found for other safety-related 1,722 times. Of those reviews, 94 were flagged as safety reviews. In sum, while several safety-related reviews do, related and 31 were specific to ingestion of magnets. It in fact, precede major recalls, the majority seem to coin- was found that reviews that flagged safety issues (FLG) cide with them reflecting what’s in the news. were longer than other types and magnet (MAG) hazards were longest when compared to reviews with no hazard Conclusions and Discussion identified (NOH): 1110 vs. 905 vs. 464 characters, respectively. FLG~NOH and MAG~NOH review length In conclusion, we were able to confirm that consumer differences were both significant: p < 0.000. Customers reviews such as the ones provided on the online retailer reported to a statistically significant degree that they website do contain safety data that consumers can and do found safety-related reviews more helpful, p = 0.045, than access. The information appeared to be read at least as other types. Furthermore, controlling for duration of ex- frequently as non safety-related information and is gener- posure, customers reported that they both read and found ally perceived to be helpful. In general, safety concerns useful MAG reviews more frequently that other types at did negatively impact reviews in terms of star ratings. the p < 0.000 level. Finally, while some safety related reviews occurred in advance of the national recall campaigns and to varying On the website of interest, each review was accompanied degrees, the majority occurred about the same time as by a star rating on a scale of 1 to 5 stars. Overall, positive those recalls suggesting that they may have served to reviews (4 and 5 stars) were more common than negative amplify rather than predict the emergence of notable reviews (1 and 2 stars). However, when safety problems hazards. A CPSC website with a dedicated safety-related were a focus of a review, the proportion of 1-star ratings purpose for product review could potentially produce a increased significantly. This shift was most pronounced greater quantity of early-warnings. This analysis does not for ratings associated with MAG hazards where the num- speak to the possibility of false alarms and the represen- ber of 1-star ratings exceeded all other types. These tativeness (or lack thereof) of negative and positive re- differences were confirmed by χ2 tests comparing the 1- view data which would pose significant challenges for and 2-stars to the 4- and 5-star counts which found any researcher hoping to use such a database in a predictive manner.

For further information contact: Gavin Huntley-Fenner Exponent 320 Goddard Ave., Suite 200 Irvine, CA 92618 Phone: 949.242.6041 Fax: 949.242.6099 E-Mail: [email protected]

156 Marketing and Public Policy Conference Proceedings / 2009 A CUSTOMER SERVICE CONCEPTUALIZATION OF THE EHAVIORAL ADVERTISING PRIVACY DEBATE

JoAnna Williamson, The University of North Carolina at Chapel Hill

SUMMARY likelihood that customers will barter for some level of privacy. The literature includes research that concludes, Technological advances are increasingly enabling the consistent with the premise of this paper, that additional searching, browsing, shopping, and e-mail behaviors of multi-dimensional, customer-driven approaches are Internet users to be tracked, compiled, stored, and re- needed. trieved. This information can then be used to target advertisements to consumers based on their individual To help fill this literature gap, and in response to the FTC’s behavioral or “ehavioral” profiles. The proper balance call for feedback regarding its proposed self-regulatory between the benefits of this type of marketing principles, this paper addresses the following research customization versus the risks of data privacy intrusions question: what, if any, guidance can customer-defined continues to be questioned. This paper proposes that dimensions of online service quality provide to help customer-defined dimensions of online service quality be evaluate the appropriate balance between online advertis- used to frame proposed solutions in this ehavioral adver- ing customization and online consumer privacy? tising privacy debate. Methodology and Findings to Date Online Service Quality Transcripts from the FTC ehavioral advertising town hall Contemporary service research has shown that evalua- are being analyzed to address this research question. tions of online service quality are multi-faceted. In addi- Textual analysis methodology is being used to align input tion to data privacy and security, other service dimensions received from town hall participants with online service raised by online customers, as identified in the most quality dimensions. Findings to date provide the follow- widely-cited research on this topic, include: (1) effi- ing examples of how key issues addressed in the town hall ciency – the ease and speed of using a website, (2) can be informed by customer-defined dimensions of online fulfillment – the extent to which a site’s promises are service quality: fulfilled, (3) responsiveness – effective handling of problems encountered through a website, (4) compensation – 1. Efficiency – ehavioral marketers appeared to be the degree to which the site compensates customers for sensitive to the need for customers to quickly and problems, and (5) contact – the availability of customer easily find the information they need on a website. assistance if needed to handle problems. For example, website owners offered lengthy explanations of their efforts to find prominent placement Online Privacy and Ehavioral Advertising and user-friendly designs for their posted privacy policy notices. The Federal Trade Commission (FTC) hosted a town hall in November 2007 to bring together the collective input of 2. Fulfillment – consumer advocates argued that consumer advocates, industry members, academics, in- ehavioral marketers as a group have a questionable formation technologists, and interested publics. The after- track record of self-regulating their adherence to math of the town hall included the FTC’s December 2007 privacy commitments they have made to customers. issuance of proposed self-regulatory principles for orga- This would appear to run afoul of the service quality nizations engaging in online behavioral advertising, along dimension of fulfillment of promises made, and with an appeal for public comment and discussion that strengthen the arguments made for government regu- could lead to further development of the principles. lation of privacy promises.

Scholarly literature already suggests a variety of research 3. Responsiveness – the service quality dimension of constructs to assess the problems and solutions of ehavioral responsiveness in handling problems appears to be privacy management. Examples include using trust fac- better reflected by arguments in favor of allowing tors to determine consumers’ willingness to provide infor- customers to opt in, or affirmatively agree to, ehavioral mation, and developing economic models to assess the marketing practices that they find to be problematic.

Marketing and Public Policy Conference Proceedings / 2009 157 The current practice of assuming that customers means such as technologically removing “cookies,” agree to such practices unless they specifically opt or text files, embedded in their Internet browsers and out seems to lend credence to concerns expressed in used by ehavioral marketers. However, this argument the town hall regarding the perceived surreptitious, can be refuted based on the definition of privacy as an unresponsive nature of ehavioral marketing. online service quality dimension, which suggests that consumers gauge the extent to which the site itself 4. Compensation – some town hall discussants cau- protects their privacy, rather than presuming that tioned that the idea of compensating customers for consumers have the knowledge or expertise to pro- relinquishing some of their privacy rights could lessen tect themselves. the Internet’s appeal to advertisers who fund the free Internet content enjoyed by consumers. The study in progress takes a unique step toward integrating the work of data privacy and service quality research- 5. Contact – of the alternatives discussed in the town ers. The resulting conceptual framework will be based on hall, a centralized do-not-track registry would appear constructs that strike a service-driven balance between the to be the most efficient way of allowing consumers to risks and rewards of ehavioral advertising. This type of contact website owners regarding ehavioral advertis- customer-centric approach appears to be consistent with ing practices that the consumer finds problematic. the FTC’s stated desire to ensure the ongoing integrity of the Internet as a source of innovation in consumer services 6. Privacy – proponents of ehavioral advertising argued and products. References are available upon request. that consumers can control their own privacy, by

For further information contact: JoAnna Williamson The University of North Carolina at Chapel Hill 388 Carroll Hall Chapel Hill, NC 27599–3365 Phone: 614.313.0782 E-Mail: [email protected]

158 Marketing and Public Policy Conference Proceedings / 2009 CONSUMER COMPLAINT CHANNEL CHOICE IN ONLINE AND OFFLINE PURCHASE ENVIRONMENTS: IF I BUY ONLINE, WILL I COMPLAIN ONLINE?

Seul Lee, University of Georgia, Athens Brenda J. Cude, University of Georgia, Athens

SUMMARY Considering that consumers who are or have been online to make a purchase may have an additional set of com- With the emergence of the Internet consumers can pur- plaint options that are unavailable or less available to chase a product or a service via virtual routes. E-com- consumers who make their purchase offline, Hypoth- merce sales continue to increase; total e-commerce sales esis 1 is proposed as follows: Consumers will choose a for 2008 were estimated at $133.6 billion, an increase of 4.6 percent from 2007 (U.S. Census Bureau). The change different complaint channel in an online shopping envi- in purchase environments from a physical exchange to a ronment than in an offline shopping environment. virtual one also potentially affects consumers’ complaining behavior (CCB). Along with the pivotal role of an Consumers who experience greater dissatisfaction are alternative purchase route, the Internet also provides likely more motivated to complain and may be more likely consumers with a new way to communicate their experi- to seek a complaint channel that makes complaining ences to companies and to other consumers as well. The easier. Given that the Internet may reduce consumers’ Internet plays an important role in fostering consumer articulation of their concerns and complaints against com- psychological costs of complaining and, consequently, panies. That is, consumers can complain online at ex- increase their intention to voice their dissatisfaction (Hong tremely low cost and independent of distance and time. and Lee 2005), Hypothesis 2 is proposed as follows: The growing popularity of individual hate sites (i.e., There will be an interaction between the purchase envi- paypalsucks.com) and general complaint sites (i.e., ronment and the consumers’ degree of dissatisfaction. complaints.com) are supportive of the importance. Since Specifically, those experiencing stronger dissatisfaction the Internet allows consumer-to-consumer interactions to with an online purchase will be more likely to choose an be no longer purely private but to be more public (Harrison- online complaint channel than those experiencing weak Walker 2001), online consumer complaining behavior (OCCB) should be understood from a different perspec- dissatisfaction with an online purchase. tive than traditional consumer complaining behavior (CCB). Previous research (Bearden, Crocket, and Gra- A 2 (purchase environment: offline versus online) × 2 (the ham 1970; Cho, Im, Hiltz, and Fjermestad 2002; Landon degree of dissatisfaction: weak versus strong) online 1977) has suggested that consumers are more likely to experimental design was used to investigate the hypoth- complain if they are more dissatisfied, the purchase is eses. Four different scenarios were offered to respon- relatively important to them, the perceived benefits of dents. A furniture purchase was chosen as the research complaining exceed the costs, the consumer sees him/ context, as the product is delivered in both the offline and herself as more personally competent, and when there are problems with company response times. Cho et al. (2002) the online environments, making both purchase situations developed the studies and demonstrated that there were comparable. To investigate the clearer relationship be- differences between the antecedents of CCB in the online tween the degree of dissatisfaction and the complaining and the offline purchase environments. However, no channel choice, the degree of dissatisfaction was manipu- other studies have focused on customers’ choice of an lated via the scenario descriptions, instead of measured. online vs. offline complaint channel. Therefore, this re- Based on the likelihood respondents would choose, the search adds to the previous work by looking into two main consumer’s complaint channel preference was computed factors that influence the online vs. offline complaining as a relative ration of his/her average score for use of channel choice: the environment (online vs. offline) in which the consumer makes the purchase and the degree of online channel choices to his/her average score for use of consumer satisfaction/ dissatisfaction with the purchase. offline channel choices. Data were collected via a self-

Marketing and Public Policy Conference Proceedings / 2009 159 administered online experiment. A total of 511 partici- or an offline environment. Specifically, in an online pants participated, and after filtering, 480 undergraduate shopping environment, a preference for an online com- college students’ data was analyzed. plaint channel preference was amplified by strong dissatisfaction with the purchase. These findings further imply With a proper manipulation on degree of dissatisfaction, that how a different managerial effort is needed in each the research reported here confirmed that a consumer’s purchase environment and for each level of dissatisfac- choice of a complaint channel is influenced by the online tion. vs. offline purchase environment as well as the consumers’ level of dissatisfaction. Hypothesis 1 was supported, The limitations of this research include a student sample, indicating that the complaint channel choice differed in a single purchase situation, and limited variables. None- the online and offline shopping environments. Specifi- theless, the results suggest that future research into con- cally, online customers were more likely to use online sumer complaining behavior should consider not online channels to complain than to use offline channels. Fur- the online vs. offline purchase environment but also ther, Hypothesis 2 was also supported, suggesting that the whether the complaint channel is online or offline. Refer- effect of dissatisfaction on complaint channel choice ences are available upon request. depends on whether the purchase was made in an online

For further information contact: Brenda J. Cude University of Georgia Dawson Hall Athens, GA 30602–2622 Phone: 706.542.4857 Fax: 706.583.0313 E-Mail: [email protected]

160 Marketing and Public Policy Conference Proceedings / 2009 PRETEEN CHILDREN’S ONLINE PRIVACY: HOW ONLINE USAGE MOTIVES AFFECT PRIVACY CONCERNS AND INFORMATION DISCLOSURE

May O. Lwin, Nanyang Technological University, Singapore Anthony D. Miyazaki, Florida International University, Miami Andrea J.S. Stanaland, Radford University, Radford Evonne Lee, Nanyang Technological University, Singapore Mindawati Wijaya, Nanyang Technological University, Singapore

SUMMARY upon individual needs (Rubin 1996) that are fulfilled via usage motivations. The number of children online is rapidly increasing worldwide, resulting in concerns regarding the vulnerability of Although the potential loss of privacy may deter Internet children with respect to their privacy and safety, particu- usage, prior work has found that online users may put larly with respect to online marketers who seek to gather aside their privacy concerns and disclose intimate details information about this increasingly important segment. when they are interacting with a website that they find Numerous studies have shown that children are willing to interesting (Berendt et al. 2005). In the cases of children, disclose personally identifying information such as name, who are more trusting and may not comprehend underly- school, address and phone number through online chats, ing dangers, personal information disclosure may be more games, and contests, which can jeopardize personal pri- likely to occur for websites that fulfill important usage vacy (Lwin, Stanaland, and Miyazaki 2008). The gather- motives (Carlson 2000). Thus, we propose that as particu- ing of children’s personal information by marketers has lar usage motivations such as information seeking, enter- been criticized by privacy advocates as an intrusion to tainment, and socialization increase, children will be less children’s privacy. concerned about privacy and more willing to disclose sensitive information. We further propose that the presen- Few researchers have explored the social contexts and tation of free gifts or memberships will enhance such consequences of the Internet with respect to children’s disclosure behavior while various parental mediation information disclosure behavior. Yet, children have been strategies may limit it (Lwin et al. 2008). deemed to be the most vulnerable population by academic researchers and government entities. Although some leg- To examine the hypothesized relationships, we developed islation, such as the U.S. Children’s Online Privacy Pro- two survey questionnaires that were administered to pre- tection Act (FTC 1998), has been enacted to address teen children (ages 10–12) and to their corresponding online privacy concerns of children, its effectiveness has parents. The children’s survey examined each child’s been repeatedly questioned (FTC 2007). Any new at- motivations to use the Internet, their online behaviors, tempts to legislate children’s online privacy require a their privacy concerns, and how much information they greater understanding of how children themselves view would disclose before and after knowledge of an incen- privacy, what traits or motivations lead to these views, and tive. The parents’ survey further assessed parental inter- how these views affect the propensity to disclose personal ventions and children’s online behaviors from the parent information online. perspective. Preliminary data were collected from eighty 10–12 year-old children and their corresponding parents The current research examines motivations for Internet after receiving permission at the university, elementary use and how these motivations relate to children’s privacy school, parent, and child levels. Initial analyses showed concerns and, subsequently, willingness to disclose per- that stronger motivations to use the Internet were associ- sonally identifiable information. Using the Uses and Grati- ated with lower concerns regarding online privacy, par- fication Theory from communications research (Blumler ticularly for social networking motives. However, for and Katz 1974), we examine three motivations – informa- information seeking, stronger motives were related to tion seeking, entertainment, and socialization – for higher, rather than lower, privacy concerns. Additional children’s use of the Internet. The UGT posits that media analyses showed that, as expected, lower privacy con- choice, usage, and ultimate gratification are contingent cerns led to greater willingness to disclose sensitive infor-

Marketing and Public Policy Conference Proceedings / 2009 161 mation. Finally, direct incentives to disclose information dren were found to disclose more information when given were shown to increase information disclosure as hypoth- an incentive as compared to when there was no incentive. esized. Results have implications for the protection strategies that Findings suggest that the Uses and Gratifications Theory can be used by parents, legislators, and educators to deter is relevant to understanding children’s online privacy children from information disclosure. As established by behaviors in terms of how differing motivations to use this study, lower privacy concerns lead to higher informa- online media affect privacy-related attitudes and behav- tion disclosure. Hence, children with social networking iors. In terms of information seeking needs, the Internet needs appear to be most vulnerable online and more prone allows children to acquire and process detailed informa- to information disclosure as they have lower levels of tion about current issues and events. Information seeking privacy concerns compared to those who are motivated by is a cognitive need associated with increased knowledge information seeking and entertainment needs. Since our through resources. An alternate explanation may be that findings support the notion that children’s motivations to those with stronger privacy concerns may limit their use the Internet can influence their privacy concerns, Internet use to information seeking. Consistent with past parents should try to understand children’s motivations research on privacy on networking sites such as Facebook, by understanding the types of online activities their chil- we found that children with stronger social networking dren seek. needs were less concerned about privacy. Finally, chil-

REFERENCES 2008 [http://www.ftc.gov/ogc/coppa1.htm]. ______(2007), “Implementing the Children’s Berendt, B., O. Günther, and S. Spiekermann (2005), Online Privacy Protection Act: A Report to Con- “Privacy in E-Commerce: Stated Preferences vs. gress,” (accessed July 27, 2008), [http://www.ftc.gov/ Actual Behavior,” Communication of the ACM, 48 opa/2007/02/copparpt.shtm]. (4), 101–6. Lwin, M.O., A.J.S. Stanaland, and A.D. Miyazaki (2008), Blumler, J. and E. Katz (1974), The Uses of Mass Commu- “Protecting Children’s Privacy Online: How Paren- nication. Thousand Oaks, CA: Sage. tal Mediation Strategies Affect Website Safeguard Carlson, S. (2000), “Teens Will Give Online Marketers Effectiveness,” Journal of Retailing, 84 (2), 205–17. Personal Information for Money: A Study Finds,” Rubin, A.M. (1996), “Uses, Gratification and Media Ef- Chronicle of Higher Education, 46 (39), 50. fects Research,” in Perspective on Media Effects, Federal Trade Commission (1998), “Children’s Online J. Bryant and D. Zillmann, eds. Hillsdale, NJ: Privacy Protection Act of 1998,” accessed July 12, Lawrence Erlbaum, 281–301.

For further information contact: Anthony D. Miyazaki Department of Marketing College of Business Administration Florida International University 11200 SW 8th Street, RB 307B Miami, FL 33199 Phone: 305.348.3330 Fax: 305.348.3792 E-Mail: [email protected]

162 Marketing and Public Policy Conference Proceedings / 2009 IT’S COOL: REVIEW AND ANTICIPATED IMPACTS OF THE NEW COUNTRY OF ORIGIN FOOD LABELING LAW

Christopher Newman, University of Arkansas, Fayetteville Steven W. Kopp, University of Arkansas, Fayetteville

SUMMARY the individual pineapples and bananas would be exempt from those labeling requirements. For about 75 years, federal law had required most items imported into the U.S. – including food items – to have a Most recently, the 2002 and 2008 Farm Bills amended the country of origin label presentable to the ultimate pur- Agricultural Marketing Act of 1946, drastically changing chaser, or the last person in the U.S. to receive the product U.S. food labeling laws. To fully understand the current in its imported form (19 U.S.C. § 1304 (a)). The law transition that has occurred with respect to U.S. food exempted products from labeling requirements that would labeling requirements, it is important to examine and be “economically prohibitive” to label (19 U.S.C. § 1304 review all the previous legislation that has led to the most (a)(3) (C)). Some of those exempt products included fish, recently enacted 2008 Farm Bill. These new amendments livestock, vegetables, fruit, live or dead animals, and nuts affect every level of the agricultural marketing and distri- (19 U.S.C. § 1304 (a)(3) (J)). The “immediate containers” bution process from the grower to the consumer. This of these products, however, were required to bear country paper provides a history of the progression of this legisla- of origin labels (19 U.S.C. § 1304 (b)). For example, tion, as well as the factors that helped shape the program. cartons containing pineapples from Costa Rica or bananas In addition, the impact of the new law on consumers, from Panama would be required by law to be labeled, but retailers, and producers will also be examined. Directions for future research to assess the law’s impact will also be provided. References are available upon request.

For further information contact: Christopher Newman University of Arkansas 1 University of Arkansas Fayetteville, AR 72701–1201 Phone: 479.575.4005 Fax: 479.575.8407 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 163 20th ANNIVERSARY SESSION

JPP&M’S HISTORY AND CONTRIBUTIONS – REFLECTIONS OF ITS EDITORS

Session Chair: Alan Andreasen, Georgetown University, Washington, D.C.

Panelists: Thomas C. Kinnear, University of Michigan, Ann Arbor Patrick E. Murphy, University of Notre Dame Michael B. Mazis, American University, Washington, D.C. Debra L. Scammon, University of Utah, Salt Lake City J. Craig Andrews, Marquette University, Milwaukee Joel B. Cohen, University of Florida, Gainesville Ronald Paul Hill, Villanova University

SUMMARY ing Association, and in 1991 the journal increased its coverage to two issues per year. As each journal editor The Journal of Public Policy and Marketing was founded served their terms (Mazis 1992–1995; Scammon 1996– in 1982 at the University of Michigan by Professor Tho- 1998; Andrews 1999–2001; Cohen 2001–2007; Hill 2007– mas Kinnear. According to Kinnear “it was a time of great present), they broadened and enriched the scope of the excitement for those of us interested in public policy journal, bringing their unique perspectives and insights to issues in marketing. . . . Unfortunately, there was a limited this developing area of scholarship. The journal now amount of space available in the standard marketing plays a central role in the infrastructure of the marketing journals as public policy research competed with many and society field. Over time, it has developed into the other substantive topics” (as cited in Miyazaki and Sprott premier journal for marketing scholars and public policy 2002). Kinnear served as the first editor of the journal, makers interested in the intersection between business with an inaugural issue covering topics including FTC’s and government. affirmative disclosure program, the economics of advertising, reduction of air pollution, several perspectives on In the final 20th Anniversary session, the history of this energy conservation, product liability, consumerism, and important journal will be reviewed in detail, and assessed political advertising. In the early years, 1982–1990, the by the panel of all the Editors in its history. Included in the journal had one issue per year, with a steadily expanding discussion will be lessons that each Editor would pass on reputation. In 1986–1987 Kinnear turned the journal to people wishing to successfully publish in this area. In editorship over to Patrick Murphy, who served until 1992. addition, the Editors will each nominate certain papers During this period, in 1990, the journal publisher shifted published during their tenures that they believe everyone from the University of Michigan to the American Market- interested in this field would be well advised to read.

164 Marketing and Public Policy Conference Proceedings / 2009 SPECIAL SESSION

UNSUSTAINABLE CONSUMPTION HABITS: INSIGHTS AND IMPLICATIONS FOR MARKETING AND PUBLIC POLICY

Session Co-Chairs: Leona Tam, Old Dominion University, Norfolk Brennan Davis, Baylor University, Waco

Discussant: Alan Levy, Food and Drug Administration, College Park

SESSION OVERVIEW for youth at school. Past research has shown that highly sociable students whose schools are near fast food, versus Much attention has been given to the rising incidence of those whose schools are not near fast food, have higher children obesity. Adding to widely accepted criticisms of rates of obesity, body mass index levels, and unhealthy targeted advertising (Desrochers and Holt 2007; Moore food consumption (Davis and Carpenter 2009). The paper and Rideout 2007), researchers have identified a number by Davis and Pechmann examines the mental processes of key contributors to the unhealthy eating behaviors of that mediate and moderate such relationships. children and adults (Colapinto et al. 2007; Nestle 2006; Roemmich et al. 2002). We present two sets of work: each The second paper offers solutions to the problems of set with a paper introducing new knowledge about a topic healthy eating by youth at school. Past research suggests on unhealthy eating followed by a paper focusing on that banning less healthy foods may lead students to sustainable, healthy eating on that same topic. instead eat more fast food off-campus and to eat more snacks at home (Vecchiarelli, Takayanagi, and Neumann In each set, we address research questions that are closely 2006). The paper by Wansink and Payne proposes ways related to the theme of the 2009 Marketing and Public to offer students the freedom of choice in selecting food Policy conference, “Contributions, Controversies, and while subtly influencing them toward healthier options. Continuing Challenges in the Worlds of Consumer Pro- tection and Competition.” Because these questions have The third paper introduces a second set of issues by not been adequately addressed in the literature, we discuss showing that product factors like package size influence the theoretical issues emerging from our research and food consumption (Scott, Nowlis, Mandel, and Morales highlight promising research opportunities and policy 2008). Ultra-concentrated food and household products implications. Each set asks the following research ques- have become more popular in the marketplace. The paper tions. by Scott, Nowlis, and Mandel investigates the factors causing over-consumption of these ultra-concentrated Set 1: (1) How do choices at school influence un- food and products. healthy consumption by youth? (2) How can we design food choices in the The last paper complements the third by offering insights school environment to maximize sustain- into potential long-term solutions to over-consumption. ablyhealthy decisions? When consumers repeatedly eat bigger portions or eat unhealthy food, eating habits develop as such consump- Set 2: (3) How does the package influence over-con tions repeat frequently under stable circumstance over sumption? time (Ouellette and Wood 1998). Such habits are persis- (4) Given the problems of over-consumption, tent and hard to change (Wood, Tam, and Guerrero Witt how can we sustainably change unhealthy 2005). The paper by Tam, Spanjol, and Rosa studies how consumption habits? consumers can find the “right” strategies to help change their unhealthy snacking habits by making implementa- The first paper highlights that eating is often a social event tion plans based on their regulatory motivation.

Marketing and Public Policy Conference Proceedings / 2009 165 Special Session: Unsustainable Consumption Habits: Insights and Implications for Marketing and Public Policy

CONSTRUAL LEVEL THEORY AND UNHEALTHY CONSUMPTION

Brennan Davis, Azusa Pacific University, Azusa Cornelia Pechmann, University of California at Irvine

SUMMARY Specifically, we test the moderating effects of sociability and event type on the mental construal of an event that is We investigate the relationship between physical dis- spatially proximal versus spatially distant. We also test tance, social distance, fast food retail sites and unhealthy self control as a mediator of construal and unhealthy consumption of fast food. Past research has shown that consumption. We measure sociability via the Social Iden- highly sociable students whose schools are near fast food, tity Scale (Karasawa 1991) and the Sociability Scale versus those whose schools are not near fast food, have (Cheek and Buss 1981). We measure construal level via higher rates of obesity, body mass index levels, and Descriptor Abstraction (Fujita, Henderson, Eng, Trope, unhealthy food consumption (Davis and Carpenter 2009). and Liberman 2006); and via Action Identification from We explore the mental processes that mediate and moder- the Behavioral Identification Form (Vallacher and Wegner ate this relationship. 1989; Fujita, Henderson, Eng, Trope, and Liberman 2006). And we measure self-control as Delayed Gratification and Liberman and Bar-Anan of Tel Aviv University and Temptation Evaluation (Fujita, Trope, Liberman, and Trope of New York University introduced Construal Leven-Sagi 2006). Level Theory (2006). Their theory states that spatial and social distance is a subset of psychological distance. We show that spatial distance significantly influences Events that are spatially distant are construed more ab- mental construal of social events and intentions to fre- stractly, for example. Psychologically proximal events quent food restaurant; social distance significantly mod- lead to low-level construal (Trope, Liberman, and Wakslak erates the role of spatial distance on construal and inten- 2007) and lower self-control over temptation to consume tion to frequent fast food. We show that self-control (Fujita, Trope, Liberman, and Levin-Sagi 2006). We test mediates the relationship between construal level and the link between construal level, self-control and un- consumption. We also show that visiting fast food is healthy consumption of fast food, shedding light on perceived primarily as a social event rather than as a place unhealthy consumption by students more proximal to fast to eat when busy with life’s activities. food.

166 Marketing and Public Policy Conference Proceedings / 2009 Special Session: Unsustainable Consumption Habits: Insights and Implications for Marketing and Public Policy NUDGING BETTER CHOICES IN SCHOOL CAFETERIAS

Brian Wansink, Cornell University, Ithaca Collin Payne, New Mexico State University, Las Cruces

SUMMARY more fast food off-campus and to eat more snacks at home (Vecchiarelli, Takayanagi, and Neumann 2006). Childhood obesity has doubled since the early 1980s (Oliveira and Variyam 2003), and some critics have The challenge is how to encourage healthier choices while attributed part of this problem to the content of school still having these choices not appear to be unduly influ- lunches (Nestle 2002). School lunches are eaten by 75 enced. One solution may be to implement payment sys- percent of children between 3 and 18 years old and tems that make it less convenient to purchase less healthy account for approximately 1/3 of the weekday calories – foods. Such systems might be useful in quietly – almost 40–120 calories more than bagged lunches (Schanzenbach invisibly – guiding students to seriously consider healthier 2005). Over time, the accumulation of unhealthy choices trade-offs they might have otherwise ignored. By still in the lunchroom could easily contribute to weight gain providing students the option of purchasing whatever and obesity (Moore 2007). they wish, however, this payment system could leave the students with a choice. School lunch programs have substantial freedom over many aspects of their operations, including content, por- The policy-relevant research question is this: Are cashless tion size, price, eating environment, payment, and the payment systems encouraging less healthy food selec- extent of parental control. In particular, high schools (and tions in school? More importantly, are different socioeco- many middle schools) allow a wide array of food choices. nomic, racial, and ethnic individuals differentially af- These food choices are often facilitated by the use of fected by these payment systems? If they are, important prepayment accounts which are in the form of debit cards policy, health, and economic policies may need to be or pre-charged ID cards. changed before the food choice behavior of millions of children are adversely affected. Much scrutiny has focused on the content of school lunches and how students could be encouraged to eat Understanding how behavioral interventions influence better, such as eating more fruits and vegetables (French food choice can inform school boards as they design and Stables 2003) or less fat (French et al. 2004). One wellness policies to combat obesity and promote healthier proposed solution is to ban unhealthy food, and a second eating. For example, knowing how changes in payment is to raise its price or reduce the price of healthier food mechanisms affect food choice may help them design through subsidies (Brownell and Battle-Horgen 2004). programs that encourage better eating without alienating Neither has proven consistently effective. One survey students, potentially reducing profitability of these par- reported that banning less healthy foods decreased their tially subsidized programs. consumption but may have led students to instead eat

Marketing and Public Policy Conference Proceedings / 2009 167 Special Session: Unsustainable Consumption Habits: Insights and Implications for Marketing and Public Policy OVER-CONSUMPTION OF ULTRA-CONCENTRATED PRODUCTS

Maura L. Scott, University of Kentucky, Lexington Stephen M. Nowlis, Arizona State University, Tempe Naomi Mandel, Arizona State University, Tempe

SUMMARY and fail to adjust appropriately, leading to over-consumption (Tversky and Kahneman 1974). Consumers use more Product and package innovations frequently result in of a product when dispensing from larger packages than smaller packages with greater product potency. This wide- smaller ones (Wansink 1996). We propose that product spread phenomenon of product “ultra-concentration” is potency changes exert a unique force that causes consum- present in numerous product categories including laundry ers to overuse ultra-concentrated products. detergent, medication, and energy drinks; however, ultra- concentration and its impact on consumption behavior Four experiments demonstrate that though consumers has not been examined. This research explores the cir- understand the required amount to use, they tend to cumstances under which consumers overuse ultra-con- significantly over-consume ultra-concentrated products. centrated products, and which cues help them consume Conflicting hypotheses examine the role of perceptual the prescribed amount. distortions and anchoring and adjusting to explain these consumption patterns. Understanding what factors drive Factors influencing consumption of ultra-concentrated wasteful consumption rates and what cues help consum- products include prior usage behavior, package size, and ers more accurately use products has marketing theory, product potency. Consumers anchor on their prior usage practice, and public policy implications.

168 Marketing and Public Policy Conference Proceedings / 2009 Special Session: Unsustainable Consumption Habits: Insights and Implications for Marketing and Public Policy SUSTAINABLE SNACKING: ALIGNING REGULATORY ORIENTATION AND SNACK DECISION STRATEGIES FOR BETTER RESULTS

Leona Tam, Old Dominion University, Norfolk Jelena Spanjol, University of Illinois at Chicago José Antonio Rosa, University of Wyoming, Laramie

SUMMARY consumption habits. Consumers with entrenched vs. weak unhealthy snacking habits obtained differential effective- Consider that about 90 percent of Americans snack on any ness in changing such habits using different types of given day with an average of two snacks per day (Driskell, implementations plans. In particular, consumers with Kim, and Goebel 2005; Zizza, Tayie, and Lino 2007), and weak unhealthy snacking habits snacked more healthily habits account for approximately one-half of total food when they developed any kind of implementation plans, consumption (Naik and Moore 1996). However, snacks whereas consumers with entrenched unhealthy snacking only account for 3–8 percent of our daily vegetable and habits broke their habits and snacked more healthily only fruit consumption (Bowman 1997), potentially posing a when they developed implementation plans that matched significant threat to healthful living. Constructively shap- their regulatory orientations. ing the emerging eating habits of children and young adults to promote sustainable (long-term) health and well- Our findings point to more effective methods for identify- being is clearly a primary concern (Cartwright et al. 2003; ing the “right” strategies to promote healthy eating. Con- Epstein et al. 2007). The idea of inducing sustainable sumers with goals to improve their eating behaviors may snacking behaviors in the young adult population raises get frustrated if they find themselves failing to overcome the important question of how can established unhealthy unhealthy food habits. Our study has implications for eating habits be modified effectively to promote sustain- programs aimed at promoting sustainable snacking be- able snacking behaviors. haviors in teenagers and young adults, a problem with national implications and of importance to government, Through the lens of disposition and self-imposed goals, the healthcare industry, and responsible food producers. we investigate the effects of self-regulatory orientations Admonition is not enough, and hence our emphasis is on (Higgins 1997) and different implementation plans isolating mechanisms that underlie the over-consumption (Gollwitzer and Brandstatter 1997) on changing everyday decision and developing interventions that will help con- snacking habits. Across three longitudinal studies, we sumers embrace sustainable snacking behaviors success- examined how young people’s (i.e., college students’) fully. Our research also carries implications for policies unhealthy snacking habits can be changed to ones that pertaining to food marketing in school environment, as promote extended well-being – what we call sustainable well as healthy lifestyle campaigns.

REFERENCES ent Income Groups,” Family Economics and Nutri- tion Review, 10 (4), 45–49. Bar-Anan, Yoav, Nira Liberman, and Yaacov Trope Brownell, Kelly D. and Katherine Battle-Horgen (2004), (2006), “The Association between Psychological Dis- Food Fight. New York: McGraw-Hill. tance and Construal Level: Evidence from an Implicit Cartwright, Martin, Jane Wardle, Naomi Steggles, Alice Association Test,” Journal of Experimental Psychol- E. Simon, Helen Croker, and Martin J. Jarvis (2003), ogy, 135 (4), 609–22. “Stress and Dietary Practices in Adolescents,” Health Bowman, Shanthy A. (1997), “Snacking Habits of Differ- Psychology, 22 (4), 362–69.

Marketing and Public Policy Conference Proceedings / 2009 169 Cheek, Jonathan M. and Arnold H. Buss (1981), “Shyness Online Marketing of Food to Children: Is it Just Fun and Sociability,” Journal of Personality and Social and Games?” Journal of Public Policy & Marketing, Psychology, 41 (2), 330–39. 26 (2), 202–20. Colapinto, Cynthia K., Angela Fitzerald, L. Janette Taper, ______(2007), “Perspectives on Food Marketing and Paul J. Veugelers (2007), “Children’s Preference and Childhood Obesity: Introduction to the Special for Large Portions: Prevalence, Determinants, and Section,” Journal of Public Policy and Marketing, 26 Consequences,” Journal of the American Dietetic (2), 157–61. Association, 107 (7), 1183–90. Naik, Narayan Y. and Michael J. Moore (1996), “Habit Davis, Brennan and Christopher Carpenter (forthcoming, Formation and Intertemporal Substitution in Indi- 2009), “Fast-Food Placement Around Schools and vidual Food Consumption,” The Review of Econom- Youth Obesity,” American Journal of Public Health, ics and Statistics, 78 (2) 321–28. 99 (3). Nestle, Marion (2002), Food Politics: How the Food Desrochers, Debra M. and Debra J. Holt (2007), Industry Influences Nutrition and Health. Berkeley “Children’s Exposure to Television Advertising: Im- and Los Angeles, CA: University of California Press. plications for Childhood Obesity,” Journal of Public Oliveira, Victor and Jayachandran N. Variyam (2003), Policy & Marketing, 26 (2), 182–201. “Childhood Obesity and the Role of USDA,” Food Driskell, Judy A., Yong-Nam Kim, and Kelly J. Goebel Assistance and Nutrition Research Report, (34–11), (2005), “Few Differences Found in the Typical Eat- (July). ing and Physical Activity Habits of Lower-Level and Ouellette, Judith A. and Wendy Wood (1998), “Habit and Upper-Level University Students,” Journal of Ameri- Intention in Everyday Life: The Multiple Processes can Dietetic Association, 105 (5), 798–801. by Which Past Behavior Predicts Future Behavior,” Epstein, Leonard H., Rocco A. Paluch, James N. Psychological Bulletin, 124 (1), 54–74. Roemmich, and Meghan D. Beecher (2007), “Fam- Roemmich, James N., Suzanne M. Wright, and Leonard ily-Based Obesity Treatment, Then and Now: Twenty- H. Epstein (2002), “Dietary Restraint and Stress- Five Years of Pediatric Obesity Treatment,” Health Induced Snacking in Youth,” Obesity Research, 10 Psychology, 26 (4), 381–91. (11), 1120–26. French, Simone A. and Gloria Stables (2003), “Environ- Schanzenbach, Diane Whitmore (2005), “Do School mental Interventions to Promote Vegetable and Fruit Lunches Contribute to Childhood Obesity?” Harris Consumption among Youth in School Settings,” Pre- School Working Paper Series 05.13, University of ventative Medicine, 37, 593–610. Chicago. ______, Mary Story, Jayne A. Fulkerson, and Scott, Maura L., Stephen M. Nowlis, Naomi Mandel, and Peter Hannan (2004), “An Environmental Interven- Andrea C. Morales (2008), “The Effects of Reduced tion to Promote Lower Fat Food Choices in Second- Food Size and Package Size on the Consumption ary Schools: Outcomes of the TACOS Study,” Ameri- Behavior of Restrained and Unrestrained Eaters,” can Journal of Public Health, 94 (9), 1507–12. Journal of Consumer Research, 35, 391–405. Fujita, Kentaro, Marlone Henderson, Juliana Eng, Yaacov Trope, Yaacov, Nira Liberman, and Cheryl Wakslak Trope, and Nira Liberman (2006), “Spatial Distance (2007), “Construal Levels and Psychological Dis- and Mental Construal of Social Events,” Psychologi- tance: Effects of Representation, Predictio, Evalua- cal Science, 17 (4), 278–82. tion, and Behavior,” Journal of Consumer Psychol- ______, Yaacov Trope, Nira Liberman, and Maya ogy, 17 (2), 83–95. Levin-Sagi (2006), “Construal Levels and Self-Con- Tversky, Amos and Daniel Kahneman (1974), “Judgment trol,” Journal of Personality and Social Psychology, under Uncertainty: Heuristics and Biases,” Science, 90 (3), 351–67. 185 (4157), 1124–31. Gollwitzer, Peter M. and Veronika Brandstatter (1997), Vallacher, Robin R. and Daniel M. Wegner (1989), “Lev- “Implementation Intentions and Effective Goal Pur- els of Personal Agency: Individual Variation in Ac- suit,” Journal of Personality and Social Psychology, tion Identification,” Journal of Personality and So- 73 (January), 186–99. cial Psychology, 57 (4), 660–71. Higgins, E. Tory (1997), “Beyond Pleasure and Pain,” Vecchiarelli, Stephanie, Sumiko Takayanagi, and Char- American Psychologist, 52 (December), 1280–1300. lotte Neumann (2006), “Students’ Perceptions of the Karasawa, Minoru (1991), “Towards an Assessment of Impact of Nutrition Policies on Dietary Behaviors,” Social Identity: The Structure of Group Identification Journal of School Health, 76, 525–31. and its Effects on In-Group Evaluation,” British Wansink, Brian (1996), “Can Package Size Accelerate Journal of Social Psychology, 30 (4), 293–307. Usage Volume?” Journal of Marketing, 60 (July), 1– Moore, Elizabeth S. and Victoria J. Rideout (2007), “The 14.

170 Marketing and Public Policy Conference Proceedings / 2009 Wood, Wendy, Leona Tam, and Melissa Guerrero Witt Zizza, Claire A., Francis A. Tayie, and Mark Lino (2007), (2005), “Changing Circumstances, Disrupting Habit,” “Benefits of Snacking in Older Americans,” Journal Journal of Personality and Social Psychology, 88 of American Dietetic Association, 107 (5), 800–6. (6), 918–33.

For further information contact: Brian Wansink Department of Marketing Cornell University 110 Warren Hall Ithaca, NY 14853 Phone: 607.254.6302 Fax: 607.255.4776

Collin Payne Department of Marketing College of Business Administration and Economics 310 Guthrie Hall New Mexico State University P.O. Box 30001/MSC 5280 Las Cruces, NM 88003–001 Phone: 575.646.6693 Fax: 575.646.1498 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 171 SPECIAL SESSION

SMOKERS ARE NOT ALL THE SAME: THE IMPORTANCE OF UNDERSTANDING MOTIVATIONAL DIFFERENCES AMONG VARIOUS SMOKER SEGMENTS WHEN DEVELOPING POLICY STRATEGIES

Session Chair: Jeremy Kees, Villanova University

Discussant: Jeremy Kees, Villanova University

SESSION OVERVIEW tobacco influences. The final presentation interviews smokers who started out as social smokers and have quit The objective of this special session is to investigate the smoking. The author identifies various types of strategies influence of various types of marketing communication and thematic messages that seem to be more effective in strategies on various segments that are attractive to to- encouraging this type of cessation behavior. bacco companies. The three papers use three different methodologies to address the objective of this special This special session addresses the key issue of how public session. The first presentation investigates, using survey policy needs to recognize the critical differences that research, different segments of “at-risk” teens to deter- exists among various smoking publics including the “at- mine what thematic approaches can be most effective at risk of smoking public,” both teens and college students. either stopping experimentation or preventing teens from The research presented is of interest to policy researchers starting to smoke. The authors segment this market into and policy makers with a focus on risk communication. two very different groups and compare the similarities and The importance of this session rests on the belief that risk differences in how they respond to various PSAs. The communication must be targeted to affect the behavior of second presentation using an experiment focuses on an- the segment of interest. Thus, understanding the underly- other at-risk population, college students who are experi- ing motivation for considering the risky behavior or menting with smoking without fully committing to the actually undertaking the risky behavior is critical to effec- behavior. These “social smokers” pose an especially tive communication. This session makes a contribution to promising market for tobacco companies because these the risk communication literature by demonstrating that individuals do not perceive themselves to be smokers. The new strategies and objectives are needed to present risk importance of the type of thematic approach that can information to the public. effectively reach this group is also critical to countering

172 Marketing and Public Policy Conference Proceedings / 2009 Special Session: Smokers Are Not All the Same: The Importance of Understanding Motivational Differences among Various Smoker Segments When Developing Policy Strategies IDENTIFYING AT-RISK YOUTH SEGMENTS AND REACTIONS TO ANTI-TOBACCO CAMPAIGNS

Marlys J. Mason, Oklahoma State University, Stillwater Joshua L. Wiener, Oklahoma State University, Stillwater

SUMMARY filing teens, tobacco documents, and several teen interviews. A 3x2 between subject experiment showed anti- A segmentation and targeting approach is advocated by tobacco ads, which varied in theme and tone, and assessed the CDC in their Designing and Implementing an Effec- reactions. Three schools were selected based on demo- tive Tobacco Counter-Marketing Campaign (2002) graphic factors to serve as research sites. manual. A segmentation approach is critical because the efficacy of anti-tobacco efforts may vary based on target- Distinct teen segments emerged in the cluster analysis ing effectiveness. A common image of the at-risk teen to with respect to psychographic, behavioral, and environ- target is defined as a rebellious, sensation-seeking youth mental variables. Then as predicted, within these seg- with a desire to express his/her independence and chal- ments significant differences in smoking initiation, envi- lenge authority (Legacy First Look Report 2000). How- ronmental influences, and attitudes toward smoking were ever, models of youth risk behavior (Jessor 1992) suggest found (e.g., 67% vs. 17% trial in segment). Specifically, a more complex array of factors function as risk/protec- an Uninvolved/ Disconnected segment is at greatest im- tive factors that encourage/deter a teen to engage in mediate risk for uptake given many risk factors (e.g., unhealthy behaviors. Based on such frameworks, a broader smoking influences, low optimism about future) and lack conceptualization and segmentation of at-risk youth is of protective factors (e.g., little involvement in school undertaken. In addition, reactance to anti-tobacco mes- activities). An Involved/Peer Connected segment also sages may occur in teens, and particularly in those at-risk appears at risk given risk-seeking attitudes and desire for segments which feel marginalized, feel little power, or peer acceptance, while limited in protective factors such feel their freedom is threatened. Some anti-tobacco cam- as parental involvement. ANOVAs further revealed dif- paigns may contribute to this reactance due to the framing ferences in reactions to the anti-tobacco campaigns. Over- of the message. Research is needed into the impact of all, teens who had not engaged in tobacco experimenta- different anti-tobacco messages on the at-risk segments. tion reported similar attitudes regardless of exposure to different ads. However, for teens who had engaged in trial The purpose of this research was (1) to develop a portrait (and at-risk segments) the hard-hitting health messages of teen segments in one Midwestern state that are at risk and humorous social messages led to greater undesirable for tobacco uptake and (2) to assess their reactions toward responses in smoking intentions, smoking acceptability, anti-tobacco appeals. Six hundred 14 and 15-year-old and reactance. Our research highlights the need to better teens were surveyed to gather segmentation data. Specific understand the nature of at-risk youth and their reactions items were adapted from prior studies, publications pro- to media efforts.

For Further Information contact: Marlys Mason Oklahoma State University 418A Business Bldg. Stillwater, OK 74078 Phone: 405.744.5109 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 173 Special Session: Smokers Are Not All the Same: The Importance of Understanding Motivational Differences among Various Smoker Segments When Developing Policy Strategies

THE DILEMMA OF SOCIAL SMOKERS BELIEVING THAT THEY ARE NOT SMOKERS

Ingrid M. Martin, California State University at Long Beach Michael Kamins, Stony Brook University Aditi Grover, Plymouth State University

SUMMARY esteem on whether college-age smokers will comply with anti-smoking messages. Two hundred and 98 participants Marketers and public policy makers continue to be plagued saw either two social death or two social exclusion or two with the problem of creating effective communications health effects PSAs used in various anti-smoking cam- that can increase the probability of complying with risk paigns in the northeast U.S. Initial descriptive statistics avoidance behavior related to smoking. Using Terror find that participants smoked an average of 12 cigarettes Management Theory and Self-Affirmation as the theo- per month and 73 percent had friends and family members retical bases, we provide a rationale as to why traditional that were smokers. In this group of social smokers, 48 anti-smoking appeals focusing on negative health conse- percent of the participants admitted “bumming” or bor- quences are not impactful for social smokers. We focus on rowing cigarettes from friends. The low number of ciga- this group of “smokers” who do not consider themselves rettes smoked per month is an indication that this segment to be smokers because of their recent and opportunistic of the smoking “market” is still not what we would call behavior (“I only smoke when I am out with friend.” or “I “avowed” smokers providing some potentially interest- only smoke when I am under pressure with final exams,” ing policy implications. etc.). We focus on three types of thematic approaches to test the effectiveness of this type of communication strategy. We use the implications of these two theories to Overall, our initial results show that mortality salience predict and show that PSAs that focus on social death and interacts with self esteem in terms of influencing the social exclusion should be more motivating than PSAs probability of smoking in the short run in a more impactful that focus on physical and health effects of smoking to for social death and social exclusion than emphasizing the encourage smokers to quit in the short and long run. salience of health related consequences of smoking. Data analysis is currently being done with the final write-up of Specifically, we conducted an experiment designed to the results and policy implications. investigate the impact of mortality salience and self-

For further information contact: Ingrid M. Martin Department of Marketing College of Business Administration California State University at Long Beach 1250 Bellfower Boulevard Long Beach, CA 92677 Phone: 949.632.5070 Fax: 949.481.8204 E-Mail: [email protected]

174 Marketing and Public Policy Conference Proceedings / 2009 Special Session: Smokers Are Not All the Same: The Importance of Understanding Motivational Differences among Various Smoker Segments When Developing Policy Strategies

COUNTERING REACTANCE, ENTITLEMENT, AND MISGUIDED OPTIMISM IN SMOKING CESSATION STRATEGIES

Joyce M. Wolburg, Marquette University, Milwaukee

SUMMARY quitting and required multiple attempts before succeeding, if they succeeded at all (Wolburg, in press). College student non-smokers are generally supportive of anti-smoking messages in the media because they give Several themes emerged to account for the behavior voice to their frustrations over exposure to cigarette related to optimism bias (Weinstein 1980, 2001). For smoke. In fact, the more insulting the messages, the better. example, when students started smoking they were con- On the other hand, many college student smokers dislike vinced that they could do so safely because they would anti-smoking messages to such an extent that they exhibit smoke for only a short duration. They consistently be- reactance effects, particularly for messages they find lieved they would quit before becoming addicted, and judgmental (Wolburg 2006). Smokers also dislike many nearly all had a quit date in mind, such graduation from ads from the industry and regard them as thinly disguised college, first job, becoming parents, etc. Student smokers public relations efforts. Research has pointed out the need were also overly optimistic about their ability to quit, to communicate with smokers in ways that avoid evoking which made their initial attempts poorly planned and reactance so that they don’t become more entrenched in poorly executed. Their comments illustrate the need to the behavior. target college students with campaigns that emphasize how to quit more so than why to quit. Creative messages Because anti-smoking messages haven’t appeared to play are also needed to address the addictiveness of the product a major role in motivating college students to quit, a more in order to discourage non-smokers from starting without recent study attempted to learn what motivated the suc- discouraging smokers from quitting. Several strategies cessful quitters and how they accomplished the task. emerged that have potential application for cessation Health scares were the most frequent reason for quitting, campaign messages that may aid marketers and public and various quit-strategies were utilized; however, results policy makers. showed that students were unprepared for the difficulty in

REFERENCES Wolburg, Joyce M. (2006), “College Students’ Responses to Anti-Smoking Messages: Denial, Defiance, and Weinstein, Neil D. (1980), “Unrealistic Optimism about Other Boomerang Effects,” Journal of Consumer Future Life Events,” Journal of Personality and Affairs, 40 (2), 293–323. Social Psychology, 39 (5), 806–20. ______(in press), “Misguided Optimism among ______(2001), “Smokers’ Recognition of Their College Student Smokers: Leveraging Their Quit Vulnerability to Harm,” in Smoking: Risk, Percep- Smoking Strategies for Smoking Cessation Cam- tion, and Policy, Paul Slovic, ed. Thousand Oaks, paign,” Journal of Consumer Affairs. CA: Sage Publications, 81–96.

For further information contact: Joyce Wolburg Marquette University 1131 W. Wisconsin Ave. Milwaukee, WI 53233 Phone: 414.288.7309 Fax: 414.288.5227 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 175 MARKETING CLAIMS REGARDING CARBON OFFSET PROGRAMS: A COMPARISON OF U.S. AND AUSTRALIAN CONSUMER KNOWLEDGE AND BEHAVIOR

Stacy Landreth Grau, Texas Christian University, Fort Worth Michael Jay Polonsky, Deakin University, Australia Romana Garma, Victoria University, Australia

SUMMARY mental quality of some retail providers. Without regulation, it is likely that the multiple codes of standards, which This paper explores consumers understanding of environ- have emerged in recent years, will create additional opmental issues and how this effects behavior, both for portunities for confusion in the market. general environmental information and information related to carbon offsets, a popular solution to decreasing Second, one issue of debate is whether the project would carbon emissions. This research examines several re- have happened regardless of its value as a “carbon offset.” search questions. First, do consumers (both from Austra- Projects should be what is called “additionally,” which lia and the U.S.) understand carbon offset programs and represents a high quality offset because it happens above are there differences in their level of understanding? and beyond what was already planned (Clean Air-Cool Second, is consumers’ level of knowledge about carbon Planet 2006). This is not always the case, but the informa- offsets related to their general environmental knowledge tion is presented as such. Third, there is a question and are there differences between the two countries? surrounding where the money actually goes after a com- Third, does ones level of knowledge affect their behavior pany purchases a carbon offset. Much of the money is and are there differences between countries? And last, taken by retail providers in exchange for searching for what are the public policy implications for regulating carbon projects on behalf of clients. There is a lack of something so complicated as carbon offsets. transparency.

These relationships are examined using two online con- Fourth, critics point out that these carbon offset programs sumer panel samples of U.S. (N = 352) and Australian are really convenient for many companies in that they do (N = 356) consumers, as these two countries have adopted not have to change their behavior regarding the environ- similar approaches to the regulation of environmental ment in any way. Rather they can pay their way out of it marketing claims. We found that in both countries con- and still get credit for being “environmentally friendly.” sumers generally have a low understanding of carbon This could lead to a false halo effect over many companies offsets and that this knowledge is unrelated to their who are among the worst environmental offenders, but general environmental knowledge. We also found that in become “carbon neutral” by investing in external carbon both countries, higher levels of knowledge (general envi- offset programs. Consumers will also be less likely to ronmental or related to carbon offsets) do not translate curtail their behavior because they feel that since they are into differences in consumer behavior. Thus, there is a purchasing from an “environmentally friendly” company, need for regulators in both countries to expand their green their work is done. Environmentalists are advocating for marketing guidelines to include information related to real behavior change. Fifth, it is difficult to assess the carbon offsets. actual environmental benefit produced by the carbon offset companies (ACCC 2008). Most offer little informa- Issues Important to Consider for Regulation tion about the selection criteria for carbon offset projects, provide no information about RECs and claim that their While there are several issues of concern, regarding the reductions are certified when no such certification exists. use of carbon offsets in marketing, one of the primary In addition, the effectiveness is unclear. issues surrounding carbon offset programs is the ability of retail carbon offset providers to deliver on their promises. These problems highlight the fact that there needs to be Since demand has grown to reduce carbon footprints, some consistent regulation or accreditation guidelines to many unregulated companies have started to offer carbon ensure that carbon offsets achieve their desired goal. This offset projects (Gogoi 2008). Critics question the environ- will provide consumers and firms seeking to behave

176 Marketing and Public Policy Conference Proceedings / 2009 responsibly with some certainty regarding the ‘reliability’ thereby causing confusion in the market and possibly of the schemes. A market focused approach has the more importantly, reducing levels of environmental im- potential to allow a proliferation on multiple schemes, provements, which is ultimately the aim of carbon offset which have varying degrees of environmental impact, in the first place. References are available upon request.

For further information contact: Stacy Landreth Grau Neeley School of Business Texas Christian University TCU Box 298530 Fort Worth, TX 76129 Phone: 817.257.6540 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 177 ENVIRONMENTALLY FRIENDLY CONSUMPTION PRACTICES: THE ECOLOGY OF CULTURAL CONTEXT

Marie Hafey DeVincenzo, University of Utah, Salt Lake City

SUMMARY the program. Finally, the purchase is very low cost. At $1.95 per block per month individuals may spend very There is growing concern that the energy consumption of little in order to be participants in the program. This fact the developed world is leading to a warming of the earth’s makes the purchase of wind energy, a relatively new atmosphere. Global warming is caused by carbon dioxide option, similar to other low-cost socially conscious con- emissions from the burning of fossil fuels. The United sumption practices, such as curbside recycling, that are States is the biggest contributor of carbon dioxide emis- more firmly established. These types of small behaviors sions in the world. Non-renewable fossil fuels, such as are the kind that, when practiced by many individuals, add coal (49%) and natural gas (20%), are the most common up to large impacts and the resolution of social problems. fuels used to generate the electricity for U.S. household use (Energy Information Administration 2007). A pos- These “little behaviors with big impacts” have been re- sible explanation for the reliance on fossil fuels rather than ferred to as “large-scale social dilemmas” in previous renewable sources is that the cost of large-scale changes research. As defined by Weiner and Doescher (1994), a to our energy systems seems to be prohibitive. While large-scale social dilemma has a “community goal, such expense is certainly an issue, there may be other issues that as preserving the earth’s resources, which can only be thwart change. In fact, enculturation is one of the main achieved if almost all community members sacrifice” factors that may be obstructing widespread adoption of (p. 259). This study builds on previous social dilemma renewable energy. The use of fossil fuels has been in- research and adds new insights through an examination of grained in our culture, and because they have been used cultural perceptions of individuals who participate in low for many generations, it may be difficult for people to cost, inconspicuous behaviors that solve social dilemmas. acknowledge their dangers. A written projective task was administered to gain a In order to address the issues of pollution and non- cultural level understanding of consumers’ perceptions of renewable resource consumption, an innovative alterna- wind power purchasers. A secondary goal of this research tive energy program called Blue Sky is currently being was to compare perceptions of wind power purchase with offered to consumers. Through the Blue Sky program, that of other types of environmentally friendly consump- Rocky Mountain Power offers its customers the option of tion behaviors at the broad cultural level. Of particular purchasing blocks of wind power at an increased cost interest was how the purchase of wind power is perceived ($1.95 per hwh block). to compare with other private, low-cost behaviors, such as turning off the lights when leaving a room. The Blue Sky program has several characteristics that make it particularly interesting and relevant to the study of Comparison results are given for three types of low cost, socially conscious consumption. First of all, power that is private behaviors: turning off the light when leaving a purchased through the Blue Sky program is completely room at home, turning off the water while brushing teeth, undifferentiated to the user, compared to coal generated and the purchase of blocks of wind power for household power. The wind energy is put into the total energy pool electricity use. The descriptions of individuals who par- that is supplied to the grid. Users cannot determine the ticipate in each of these activities are compared to each source of their particular power. Because the power is other and to the descriptions of individuals who partici- undifferentiated, the biggest benefit of the purchase is at pate in a high cost, public behavior: the purchase of a the level of the society. Second, the purchase decision is hybrid vehicle. a private one. There are no outward signs that an individual has participated in the program. It is only reflected The results of the comparisons illustrate how cultural in the participant’s power bill. This fact reduces the perceptions of low cost, private behaviors may be orga- likelihood that the motivation for purchase is related to nized and understood using an ecological model. A con- social pressure from others, either negatively or posi- tribution has been made to the existing literature through tively, because others will not know who participates in an approach that considers the broader theoretical context

178 Marketing and Public Policy Conference Proceedings / 2009 of the behavior, such as familial, community, and cultural of wind power is a very private decision with no outward influences on individual behavior. indicators, the perceptions of the informants suggested a need for further research into the decision to purchase The results also reveal cultural perceptions of low cost, wind power. private behaviors and how they compare with perceptions of other socially conscious consumption behaviors. Indi- These results have implications for managers and policy viduals who purchased wind power were perceived to be makers, whose decisions can be informed through an very different from individuals who participated in other understanding of cultural perceptions of new offerings, types of low cost environmentally friendly behaviors. such as wind power. The present research begins this Instead, wind power purchasers were perceived to be process, and future research can add to this understanding similar to purchasers of hybrid vehicles. This is an inter- and provide policy makers and managers with informa- esting finding given the fact that the purchase of a hybrid tion they can use to structure their policies, offerings, and vehicle is a much higher cost, public consumption behav- promotions in ways that will increase participation in ior. Wind power purchasers and hybrid vehicle purchas- renewable energy programs. References are available ers were perceived to desire public awareness of their upon request. environmentally friendly activities. Because the purchase

For further information contact: Marie Hafey DeVincenzo University of Utah 1645 E Campus Center Drive Salt Lake City, UT 84112–9305 Phone: 801.486.8551 Fax: 801.581.7214 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 179 SUSTAINABLE CONSUMER BEHAVIORS: A CONCEPTUAL FRAMEWORK

Kristin Scott, Oklahoma State University, Stillwater

SUMMARY The environment and cultural values will influence more specific motivation factors such as anti-materialism, envi- Sustainability has become an increasingly important issue ronmentalism, and fiscal responsibility that will influence as environmental concerns such as global warming, ozone slow consumption during the three stages. It is important depletion, and loss of species rise in prominence on public to consider different motivational factors (e.g., frugality) and political agendas. One main cause of these environ- for engaging in sustainable consumer behaviors because mental problems is the overconsumption of natural re- they may have the same results as being environmentally sources (Tanner and Kast 2003). Consumers can play a concerned. As such, there may be other ways to motivate large role in reversing this trend and in turn, become a sustainable behaviors that do not involve attitude changes more sustainable society. For maximum environmental toward the importance of the environment. benefit, consumers should participate not only in activities such as energy reduction but also engage in more The framework describes consumption behaviors in each sustainable actions throughout the consumption process. of the three stages that would be considered sustainable. Ideally, sustainable consumer behaviors should be ac- In the acquisition stage, sustainable consumer behaviors tions taken during the acquisition, usage, and disposition would include buying fewer products (minimal consump- stages of consuming goods. These three areas all provide tion), buying durable products, and buying products that potential avenues for creating a sustainable society as well were pre-owned or have recycled components. During the as slowed consumption across integrated areas. However, usage stage, sustainable behaviors would involve increas- research suggests that very few people adopt a consistent ing the length of time that a consumer keeps and uses a approach toward product life spans and thus fail to exhibit product before acquiring another one. Four different sustainable behaviors in all three of the stages (Cooper behaviors are discussed: care and maintenance, repair, 2005). As such, investigating the differences in the three non-use, and storage. These behaviors are important be- stages and their influence on each other may be beneficial cause, if practiced, they will help extend the life of a in terms of behavioral change. This research develops a product. Non-use, however, should be decreased because new framework to consider sustainable consumer behav- if products go unused, the natural resources are wasted. In iors using the three stages of consumption: acquisition, the disposition stage, a distinction between triggers of usage, and disposition and the motivations that propel disposition and method of disposition is made. Triggers of them. This new framework provides guidance for future disposition include product obsolescence and product research in sustainability and provides a broader perspec- repair which may influence disposition method. To main- tive of sustainable consumer behaviors. tain a more sustainable society through disposition, researchers should examine the triggers of disposition and This framework examines behaviors in the acquisition, attempt to lessen the effect they may have. Also, research usage, and disposition stages that are considered sustain- could involve attempting to extend the useful life of the able, the interaction between the stages, and the multiple product once the original owner decides to dispose of it. motivational factors that propel these behaviors. Two This would influence the “how” of disposal or how a previous frameworks are drawn on: Boyd and person disposes of the product once they are finished with McConocha’s (1996) framework of the movement of it. Research might concentrate on donating and recycling consumption from acquisition to disposition and Cooper’s behaviors and thus extending the life of a product even (2005) perspective of “slow consumption.” The term after disposition has occurred. The relationship between “slow consumption” is defined as a need to slow the rate these three stages and the sustainable consumer behaviors at which raw materials are transformed into products and engaged during each are also considered. For example, eventually discarded. To achieve this slower consump- those that engage in minimal consumption may be more tion, consumers may need to develop greater attachment inclined to engage in care and maintenance of their to their possessions and no longer aspire to update them as possessions. Investigating the relationship between stages soon as new models come out (Cooper 2005). The goal is will provide a better understanding of the sustainable to stretch out the consumption process as goods move into consumption process and influences that can be utilized to and out of a person’s possession. increase sustainable behaviors in all three stages.

180 Marketing and Public Policy Conference Proceedings / 2009 This framework provides implications for social market- but sustainable behaviors during the usage stage are less ers, profit marketers, and policymakers. Social marketing ubiquitous. More emphasis by companies should be given may be conducted to make consumers aware of the con- to the repair, maintenance, and storage aspects of usage cept of slow consumption and the benefits behind it and the company’s ability to provide each. Companies (whether environmental or financial). Making salient the should strive to make their products repairable, provide idea that if consumers buy products that are made to last information on maintenance and storage, and also provide longer it will not only save them money but also time in the necessary storage material. For policymakers, impor- replacement purchases will benefit not only the environ- tant topics include regulations and guidelines to provide ment but also the consumer. Consumers may be aware of consumers the necessary information to make informed behaviors that are sustainable during the acquisition stage decisions. References are available upon request.

For further information contact: Kristin Scott Oklahoma State University 405D Spears School of Business Stillwater, OK 74075 Phone: 405.744.5418 Fax: 405.744.5180 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 181 GENERATIONAL COHORT EFFECTS ON GREEN CONSUMER BEHAVIOR

Gail M. Zank, Texas State University, San Marcos Karen H. Smith, Texas State University, San Marcos

ABSTRACT “enviropreneurial marketing strategy,” as a way to take environmental issues into consideration when developing The research reported in this paper examines the effect of a marketing strategy. The importance of this approach is generational cohorts on consumer attitudes and behavior supported by a recent survey of CEOs from 43 countries with respect to the environment. Younger generational that found 87 percent believed that environmental cohorts are significantly less “green” than their older sustainability is important to a company’s profits counterparts, a finding which has substantial implications (Krebsbach 2005). for future marketing strategies and public policy initiatives. Although CEOs note the importance of environmental sustainability, Peattie and Crane (2005) suggest that past INTRODUCTION efforts of green marketing have not been as successful as hoped. Bonini and Oppenheim (2008) reported that only The quality of the environment is an issue of importance 33 percent of consumers say they are ready to buy green worldwide. Many consumers (87%) are concerned about products or have already done so. Pickett-Baker and the environmental and social impacts of the products they Ozaki (2008) findings suggest that most consumers can- buy (Bonini and Oppenheim 2008). A study by American not easily identify greener products although they would Demographics found that 80 percent of Americans say favor products manufactured by greener companies. Given that whether or not a product is safe for the environment the importance of the environment and the need for does influence their decision to buy that product (Gardyn improvement in green marketing, it is important for mar- 2003). But a May 2008 poll by Harris Interactive found keters and policy makers to better understand the behavior that most people have done little to change their lifestyles of consumers and what influences them to behave in a with only two percent of respondents indicating that they “green” manner. were avoiding over-packaged goods. However, the poll also showed that Americans think of themselves as doing Generational cohorts are a useful basis of market segmen- a great deal, with 72 percent identifying their personal tation (Schewe and Meredith 2004). Generational cohorts actions as “significant on the environment” (Solman 2008). are based on the idea that a shared historical or socially Therefore, there is a gap between consumers’ stated structured life experience influences value, attitudes, and concern for environment and their behavior. behaviors. Several studies have examined the effects of generational cohorts, but none have done so in studying Consumer concern for the environment has impacted differences in the level of environmental concern and marketers in a number of industries. For example, energy green behavior. To gain better understanding of consumer production is directly affected by this concern. There are green behavior, this paper examines the differences be- a number of states where consumers have the option of tween generational cohorts in terms of their environmen- buying “green power,” that is energy produced by envi- tal concern, environmental knowledge, attitudes toward ronmentally non-damaging technologies. And consumers the environment, perceived consumer effectiveness, and have indicated some willingness to pay a higher price for their green behavior. This paper first reviews the relevant such options (Raymond 2001). The housing market is also literature and the hypotheses. Next it discusses the meth- experiencing an increase in interest and availability of odology and the results. In conclusion, implications for environmentally friendly dwellings and home improve- policy makers are examined. ment products (Lipke 2001). There is also a large demand for energy efficient automobiles (Marketing Week 2005). LITERATURE REVIEW AND HYPOTHESES Additionally, a number of researchers suggest that marketing with environmental consciousness is a strategic Perceived Consumer Effectiveness (PCE), the belief by decision (Hart 1997; Walley and Whitehead 1994). Menon the consumer that his/her actions can have a positive and Menon (1997) propose what they term an effect on the environment, has been found to relate to

182 Marketing and Public Policy Conference Proceedings / 2009 consumer environmentalism (Allen 1982; Allen, Dahlstrand, and Biel (2004) determined that Environ- Callantone, and Schewe 1982; Antil 1984; Kinnear, Tay- mental Concern was positively related to the effect of lor, and Ahmed 1974). Specifically, Kinnear, Taylor, and information about the environment but there was a weak Ahmed (1974) examined the effect of consumers’ socio- correlation between environmental attitudes and pur- economic and personality variables and PCE on the amount chase behavior. Whereas, Schlegelmilch, Bohlen, and of ecological concern they indicate. They found that Diamantopoulos (1996) reported that both environmental personality variables were better predictors than demo- attitude and environmental knowledge (EK) are linked to graphic variables and that PCE was positively related to green purchasing behavior. environmental concern (EC). Others, in examining PCE, have found that it is related to the purchase of low- An important environmental issue for marketers to under- phosphate detergents (Webster 1975), the subject’s locus stand is the effect of the inclusion of recycled materials as of control (Henion and Wilson 1976), and consumer’s components of a product on consumer product evaluation. energy consumption patterns (Seligman et al. 1979). Mobley et al. (1995) examined this issue using a lab Balderjahn (1988), in his study of consumers in the experiment. They reported that recycling enhanced the Federal Republic of Germany, found that consumers’ image that subjects had of brands but had no real effect on belief that they could positively affect the environment perceived quality. They suggested this implied that recy- (i.e., PCE) had a positive impact on the purchase of home cling impacts the affective rather than the cognitive ele- insulation. ments of consumer attitudes. However, they point out that this was true only for recognized brands. Attitudes about A study by Ellen, Wiener, and Cobb-Walgren (1991) set unknown brands were not significantly affected by the the stage for further examination of this topic. Their study fact that those products were made of recycled material. examined the concept of PCE along with some other Whereas, Biswas et al. (2000) determined that the attitude issues related to consumer environmentalism. They re- toward recycling had a significant effect on recycling ported that environmental concern and PCE were separate waste and the purchase of recycled/recyclable products. constructs. PCE was also found to relate to consumers Cornwell and Schwepker (1995) found that a favorable engaging in such activities as recycling, purchasing envi- attitude toward environmentally-conscious living was ronmentally friendly products, and contributing to envi- positively related to consumer propensity to purchase ronmental causes. Straughan and Roberts (1999) found products with environmentally-friendly packaging. that environmental concern was less important than PCE although environmental concern was significant in the full model. Generational Cohorts

In an examination of antecedents to green behavior, While no prior research was found that examined differ- Kinnear and Taylor (1973) found that consumers’ level of ences in green behavior across generational cohorts, there Environmental Concern affect their purchase behavior for have been studies of the impact of age on green attitudes laundry detergent. However, they concluded that Envi- and behaviors, although their findings have been mixed. ronmental Concern by itself did not have sufficient impact Van Liere and Dunlap (1981) and Zimmer, Stafford, and to get polluting products off the shelves. Several studies Stafford (1994) found age to be negatively correlated with have sought to examine the underlying factors that are environmental attitudes and behavior. The explanation is critical in environmentally-related consumer behavior. that younger people will be more aware of environmental Zimmer, Stafford, and Stafford (1994) in examining con- issues. Conversely, Samdahl and Robertson (1989) found sumer attitudes, suggest that one of the difficulties facing a positive correlation between age and environmental marketers is that consumer attitudes are multi-dimen- attitudes. The belief is that as people get older, they will sional and complex. Smith and Haugtvedt (1995) suggest have a greater concern for charitable activities. that attitudes toward the environment (EA) are enhanced Diamantopoulos et al. (2003) found that older people by the strength of the attitude, whether the attitude was were found to undertake higher levels of recycling activi- held for social or utilitarian reasons, whether the attitude ties. They also offer a possible explanation for why studies is a personally held one or is one prescribed by society, have found contradictory findings regarding the influence and whether the attitudes had an affective or cognitive of age and environmental behavior. They suggest that basis. Johnson and Johnson (1995) examined attitudes studies using intentional commitment measure of behav- and behavior of subjects in Germany. They found that ioral domain often found that age is negatively related to attitudes toward the environment have four distinct di- intended behavior while those using indicators of current mensions, which are recycling, water pollution, economic behavior have found that older people display higher development, and government involvement. Grankvist, levels of green behavior.

Marketing and Public Policy Conference Proceedings / 2009 183 This study examines age from the perspective of genera- green purchasing behavior by Hong Kong adolescents, tional cohorts: Generation Y, Generation X, Baby these respondents would be classified as Generation Y Boomers, and Depression/War/Post-war. Generational (Lee 2008). A Harris Interactive Poll (2008) found that 53 cohorts are defined as a group of people who share a given percent of respondents (including the majority of historical or socially structured life experience and the Generation Y) believed that Gen Y was the most self- generational cohort experience has a relatively stable indulgent generation. O’Donovan (1997) suggests that influence over the decision-making of individuals, distin- Generation X is not motivated by civic duty and therefore guishing one generation from another (Conger 1997). The might be less concerned about the environment. Further, idea is that events that are happening when one comes of Littrell, Ma, and Halepete (2005) found that baby boomers age, imprints core values and these “defining moments” differed from Generation X when evaluating fair trade can include things such as wars, assassinations, or eco- consumers. Generation X was showed to have less concern nomic difficulties (Meredith and Schewe 2002). Each on qualities of comfort, value and quality, and less cohort has experienced different events that have helped preference for authentic product and less local activism to shape their values. behavior.

The Depression/War/Post-war cohort experienced the American Demographics (Gardyn 2003) found that older depression, World War II and the end of World War II. Americans seem to be the best consumers of environmen- These people were born between 1912 and 1945. Those in tally friendly products. In addition, those over the age of the Baby Boomer cohort were born between 1946 and 65 are more likely to believe that the measures an organi- 1964. Defining moments for this group include the Viet- zation takes to practice environmental safety make a nam War, the assassination of JFK, RFK, and Martin difference. Therefore, it is believed that: Luther King, Jr. and the resignation of President Nixon. Generation X individuals were born between 1965 and Compared to older generational cohorts, younger genera- 1980. Events shaping their lives include the Gulf War, tional cohorts: AIDS crisis, the fall of the Berlin Wall and the Stock Market crash of 1987. Gen Y includes those born after H1: Are less Environmentally Concerned, 1980. Defining moments and experiences for this group include the Internet, good economic times, and the Sep- H2: Have less Environmental Knowledge, tember 11th attacks. H3: Hold less favorable Environmental Attitudes, Schewe and Meredith (2004) recommend that marketers segment by generational cohort, and several studies have H4: Have lower Perceived Consumer Effectiveness, examined the differences between generational cohorts. Jurkiewicz (2000) examined the difference between Gen- H5: Are less likely to engage in Green Shopping, eration X and Baby Boomers employees in terms of motivation to work. Cutler (2005) evaluated generational H6: Are less likely to engage in Green Activism, and differences in terms of saving money. And Jurkiewicz and Bradley (2002) examined three generations of healthcare H7: Are less likely to engage in Recycling. executives to determine whether generational cohort was the key determinant of ethical values. Rodriquez, Green, METHODOLOGY and Ree (2003) examined the preference of leadership behavior association with generational themes. In addi- A survey was used to gather data for hypothesis testing. tion to research that has looked at differences between The survey was pre-tested using several graduate students groups, there has been characterization of each genera- and colleagues in order to obtain feedback about the tional cohort. survey in terms of completeness, understandability, and ease to complete. Generation Y is viewed as being very brand conscious so they may be less likely to worry about environmental Sample and Procedure concerns when purchasing products. Also, a study by Cui et al. (2003) determined that Generation Y responded The finalized survey was administered by mail (response more favorably to cause related marketing offer when it rate 11.3%) to a nation-wide sample of adults, ranging in helped a disaster rather than an ongoing cause. The issue age from 18 to over 75, supplemented by a local sample of the environment would be an ongoing cause. Social administered in person with young adults, resulting in 299 influence was found to be the most important predictor of usable surveys. Other published studies have combined

184 Marketing and Public Policy Conference Proceedings / 2009 two different samples in order to achieve the required prove the environmental friendliness of its products.” sample characteristics (e.g., Vitell and Muncy 2005). The Gender and ethnicity are also measured. mail survey sample was obtained by purchasing a CD with addresses of U.S. residents and selectively sampling by ANALYSIS AND RESULTS gender, age, and geographic distribution. However, the number of Generation Y adults on the list was smaller than Hypothesis Tests the other generational cohorts. A follow-up letter was used to remind potential participants about the survey, One-way analyses of variance, with each of the seven although no incentive was provided for participation. constructs as dependent variable and generational cohort Comparison of respondents and non-respondents yielded as the fixed factor, as well as multiple comparisons no significant differences in the variables of interest. In between cohorts, were used to test H1–H7. The ANOVA addition, because no geographical differences were found results were significant for six of the seven variables. As for the environmental variables, it is believed that the use shown in Table 1, the one-way ANOVAs indicate signifi- of the local sample does not impact the results. cant cohort effects at the .01 level for EC, EK, PCE, Green Shopping, and Recycling, and at the .05 level for Green The sample comprises 53 percent males and 47 percent Activism. The ANOVA for ATT approached significance females; and 85 percent white, 4 percent African-Ameri- at the .08 level. can, 5 percent Hispanic, and 6 percent other. The sample consisted of approximately 25 percent in each genera- LSD Post Hoc tests were used to identify significant tional cohort, defined as Depression/War/Post-war (1945 differences among individual cohorts for each dependent and earlier, N = 76), Baby Boomer (1946–1964, N = 73), variable. In Table 1, lower case letter superscripts specify Generation X (1965–1980, N = 75), and Generation Y significant differences at the .05 or better level. Means for (1981 and later, N = 75). Generation Y were significantly lower than Depression/ War/Post-war and/or Baby Boomer for all dependent Measures variables, except Green Activism. Generation Y was significantly less “green” than both older generational The three variables measuring Green Behavior are Green cohorts for four variables (EK, PCE, Green Shopping, and Shopping, Green Activism, and Recycling. Green Shop- Recycling) and was significantly less “green” than Baby ping is measured on a 7-point Likert-type scale with seven Boomer for EC and ATT. Means for Generation X were items (average of item responses, alpha = .87), such as “If below those of Depression/War/Post-war and/or Baby I think a company is not environmentally responsible, I Boomer for five of the seven variables. Means for Genera- will be much less likely to buy its products” (reverse- tion X were significantly below those of both older scaled). Green activism is measured on a 4-item, summed generational cohorts for four variables (EK, PCE, Green yes/no scale (alpha = .54), with items such as “Are you Shopping, and RE), and lower than those of Baby Boomers actively involved in environmental organizations.” Recy- for two variables (EC and ATT). cling is measured on an 8-item, summed yes/no scale (alpha = .76), with items such as “Do you recycle alumi- Of the generational cohorts, Baby Boomers had the most num cans.” favorable ATT, the second highest EK, the highest PCE, the greatest EC, the highest level of Green Shopping, and The four covariates, Environmental Concern, Environ- the highest level of Recycling. Depression/War/Post-war mental Knowledge, Attitude toward the Environment, had the highest EK, and was second highest in PCE, Green and Perceived Consumer Effectiveness, are measured via Shopping, and Recycling. In contrast, Green Activism 7-point multiple-item Likert-type scales. Environmental was significantly greater for Generation Y than for De- Concern (alpha = .72) is measured with items such as “The pression/War/Post-war, and was very similar for Baby concern about running out of natural resources is over- Boomer and Generation X, refuting H6. Therefore, H1 – blown” (reverse-scaled). Environmental Knowledge (al- H5 and H7 are supported, but H6 is not supported. pha = .73) is measured with items such as “I understand environmental issues better than most people.” Attitude Interaction and Covariate Effects (alpha = .70) is measured with items such as “I think that a person should urge friends not to use products that As a stronger test of the relationships between genera- pollute the environment.” Perceived Consumer Effective- tional cohort and green behavior, the data were analyzed ness (alpha = .71) is measured with items such as “Con- using MANCOVA with gender and cohort as fixed fac- sumers like myself generally can get companies to im- tors; EC, EK, ATT, and PCE as covariates; and Green

Marketing and Public Policy Conference Proceedings / 2009 185 3.03 4.69 3.42 3.31 4.84 4.31 1.20 3.50 5.23 4.32 4.65 5.15 4.62 1.00 4.26 5.78 5.21 5.42 5.53 4.79 1.01 4.58 5.53 5.01 4.97 4.97 4.59 0.70 (1.20) (1.13)(1.03) (0.95)(1.11) (0.88) (1.00) (0.92) (1.04)(1.86) (1.02) (0.93) (1.70) (0.98) (2.24) (2.55) (1.38) (1.05)(1.26) (1.15) (1.14) (0.94) (1.15)(0.69) (1.08) (1.09) (1.14) (1.33) N = 76 N = 73 N = 75 N = 75 Depression/ War/Post-war Baby Boomer Generation X Generation Y TABLE 1 Table 1 Green Consumer Behavior by Age Cohort ANOVA ResultsANOVA Means Age (Standard Cohort by Deviations) Green Behavior Age Consumer Cohort by 5.10 0.00 2.25 0.08 2.78 0.04 32.29 0.00 17.2847.55 0.00 0.00 13.84 0.00 Depression/War/Post-war is significantly Baby from different Boomer; Depression/War/Post-war is significantly Generation from different X; Depression/War/Post-war is significantly Generation from different Y; Baby Boomer is significantly Generation from different X; Baby Boomer is significantly Generation from different Y; Generation X is Generation significantly X Generation from different Y; at better. or .05 level a b c d e f cdef e bcdef ade bcdef c cdef ConstructEnvironmental Concern F p Attitude toward the Environment the toward Attitude ShoppingGreen ActivismGreen Results paired comparisons: of Environmental Knowledge ConsumerPerceived Effectiveness Recycling

186 Marketing and Public Policy Conference Proceedings / 2009 Shopping, Green Activism, and Recycling as dependent important implications for both marketing and public variables. policy. The perhaps counter-intuitive finding that younger generational cohorts are less “green” than older genera- Multivariate results indicate significant relationships ex- tional cohorts emphasizes the need for continued environ- ist between the dependent variables and EC (Wilks’ mental education. A Harris Interactive Poll found that lambda = .91, F = 9.09, p < .01), EK (Wilks’ lambda = .79, Americans aged 63 – 83 would prefer to be named the F = 25.86, p < .01), ATT (Wilks’ lambda = .89, F = 11.55, Responsible Generation (2008). This might be consistent p < .01), and PCE (Wilks’ lambda = .94, F = 5.96, p < .01). with their greater effort for the environment. Perhaps, the Significant main effects of gender (Wilks’ lambda = .97, proliferation of green marketing has increased the level of F = 3.40, p < .05) and generational cohort (Wilks’ lambda = skepticism toward products claiming to be environmen- .76, F = 9.33, p < .01), as well as a significant interaction tally friendly. In fact, in response to unsubstantiated green effect (Wilks’ lambda = .93, F = 2.28, p < .05), were found. claims, EnviroMedia, an environmental marketing consulting company, has established a website Tests of between-subjects effects indicate that the (www.greenwashingindex.com) where consumers can univariate model for each dependent variable is highly submit ads and rate them on the “Greenwashing Index,” significant and explains a substantial amount of variance measured on a scale from 1 to 5, with a “5” signifying an (F = 59.91, p < .01, adjusted R2 = .69 for Green Shopping; outright lie (Walsh 2008). Also, better education of what F = 3.94, p < .01, adjusted R2 = .10 for Green Activism; and to do in order to help the environment needs to take place. F = 5.80, p < .01, adjusted R2 = .15 for Recycling). There Solman (2008) reported that 34 percent of respondents to were significant main effects of generational cohort on all a Harris Interactive poll indicated that they had made no three dependent variables, significant main effects of lifestyle changes because they did not know what to do. gender on Recycling, and significant interaction effects on Green Shopping and Recycling. Results from the It is well known that younger generational cohorts tend to multivariate and univariate analyses show that cohort be more skeptical of advertising because they have grown effects are significant, even after the effects of the four up with the mass media, and it is likely that this skepticism covariates are removed. With respect to gender, paired transfers to green products. Greater oversight of green comparisons indicated that females are more engaged in product claims, as well as development and enforcement Green Shopping and Green Activism than males, al- of strict standards on such claims, could enhance consum- though these differences are not significant at the .05 ers’ abilities to determine which products really are envi- level. In contrast, males engage in significantly more ronmentally friendly. Generation Y expects convenience Recycling than females (F = 5.26, p < .05). and instant gratification, so it might be necessary to develop strategies that make it easier to identify green Tests of between-subjects effects also indicate that Green products (e.g., labeling) and engage in green behaviors Shopping is significantly related to all four covariates, (e.g., curbside recycling). Such efforts may help over- whereas Green Activism is significantly related only to come the cohort effects found in the current study, en- EK, and Recycling is not significantly related to any of the hancing young consumers’ concern about the environ- four predictor variables. ment and likelihood of engaging in green consumption behaviors. Effects of the generational cohort by gender interaction can be seen in Figures 1–3. Green shopping is higher for It is especially alarming that Generation Y is less “green” females than males for three cohorts, but is lower than because of the tendency for members of a generational males for Generation Y. Green Activism is higher for cohort to continue their beliefs, and associated behaviors, females than males for all cohorts except for Depression/ throughout their lives (Meredith and Schewe 2002). There- War/Post-war. In contrast, recycling by females is below fore, it cannot be assumed that Generation Y will improve that of males for three of the four cohorts; female recy- their green behavior as they age. However, given that cling is higher only for Baby Boomers. In fact, male behavior of generational cohorts is strongly influenced by recycling is very similar for Depression/War/Post-war, “defining moments,” perhaps the recent rise in gas prices Baby Boomer, and Generation X, but drops considerably will cause younger cohorts to re-examine their use of for Generation Y. scarce resources and to engage in greater green behavior. If the “defining moments” of generational cohorts can be IDISCUSSION better understood, and if environmental education can be couched in terms of these moments, it should be more Based on this study, generational cohort effects on green effective. For example, because Baby Boomers’ defining consumer behavior appear to be significant, a finding with moments may include the Love Canal disaster in the late

Marketing and Public Policy Conference Proceedings / 2009 187 FSig. Mean Mean Square df SS Type III Sig. F FSig. Mean Mean Square Error df df df Hyp. Hyp. SS Wilks' Lambda Type III Table 2 TABLE 2 M ultivariate Te sts FSig. Tests of Between-Subjects Effects R e sults of M ultivariate Analysis of Covariance Results of Multivariate Analysis Covariance Mean Mean Square df Green Shopping Green Activism Recycling Adjusted R2 = .69 Adjusted R2 = .10 Adjusted R2 = .15 InterceptEnvironmental Concern KnowledgeEnvironmental AttitudeEnvironmental Effectiveness Consumer Perceived Gender 0.94Cohort 0.79Cohort by Gender 0.91 3 3 0.89 3 285 3 285 0.94 5.96 25.86 285 ** ** 285 3 9.09 11.55 ** ** 0.93 285 6.29 9 0.97 ** 0.76 693.8 3 2.28 9 * 285 693.8 3.40 9.33 * ** SS Type III Source Corrected ModelCorrected Intercept ConcernEnvironmental KnowledgeEnvironmental AttitudeEnvironmental Effectiveness Consumer Perceived GenderCohort 7.53 311.41Cohort by 35.61 11.78Gender Error 11 1 1 1 15.86Total 28.31 TotalCorrected 35.61 7.53 11.78 1 59.91 75.36 15.93 24.92 ** .05 level. the at 0.16 significance ** ** ** * indicates 15.86 .01 level; the at significance 33.56 **Note: indicates 1 ** 46.85 0.16 4.48 5.50 2.50 2.73 11 0.33 3 1.36 2.56 1 1 1 4.26 24.61 447.04 298 1 3.94 1 1.49 5.50 2.50 3 2.73 ** 135.63 6470.33 3.16 5.10 2.31 2.53 287 2.56 1.36 299 * 8.20 * 8.51 2.37 2.88 274.05 17.36 0.47 ** 11 1 24.91 6.31 3.77 8.51 8.14 0.13 5.80 21.67 7.88 ** 1 3 5.11 ** 1 1 0.71 3 1 6.31 1.26 1 8.14 0.13 356.84 7.22 309.98 1.47 1.16 36.38 298 1.89 0.03 287 5.11 6.69 0.71 ** 1.19 1 642.00 0.66 1.08 299 36.38 8.47 76.84 39.91 ** 3 3 22.58 25.61 13.30 1 5.96 1232.53 3.10 ** 287 22.58 * 1506.58 298 5.26 4.29 * 7793.00 299

188 Marketing and Public Policy Conference Proceedings / 2009 1970’s, they may be more in-tune to the potential prob- FIGURE 1 lems that can be caused by pollution.

Another implication of this research is the need to examine messages (both marketing and public policy messages) and whether they need to target certain generational cohorts to enhance effectiveness. Factors such as content, execution, and placement have strong effects on persuasion and may vary by cohort. Furthermore, the role of the consumer in determining appropriate disposition of waste products is changing and there will be a need to better educate consumers about their responsibilities in terms of improving the environment.

The model explains a very large proportion of the variance in Green Shopping, but accounts for a much smaller proportion of the variance in Green Activism and Recy- cling. Additional research is needed to investigate what variables have greater explanatory capability for Activ- FIGURE 2 ism and Recycling. In summary, although the model showed significant relationships among the variables, additional research on the effects of gender, generational cohort, and their interaction, on environmental attitudes and behavior are warranted.

REFERENCES

Allen, Chris (1982), “Self-Perception Based Strategies for Stimulating Energy Conservation, Journal of Consumer Research, 8 (March), 381–90. ______, Roger Calantone, and Charles Schewe (1982), “Consumers’ Attitudes About Energy Con- servation in Sweden, Canada, and the United States, With Implications for Policymakers,” Journal of FIGURE 3 Marketing and Public Policy, 1, 43–57. Antil, John (1984), “Socially Responsible Consumers: Profile and Implications for Public Policy,” Journal of Macromarketing, 4, 18–39. Balderjahn, Ingo (1988), “Personality Variables and En- vironmental Attitudes as Predictors of Ecologically Responsible Consumption Patterns,” Journal of Busi- ness Research, 17, 51–56. Biswas, Abhifit, Jane Licata, Karyl McKee, Chris Pullig, and Christopher Daughtridge (2000), “The Recy- cling Cycle: An Empirical Examination of Consumer Waste Recycling and Recycling Shopping Behav- iors,” Journal of Public Policy & Marketing, 19 (1), 93–105. Bonini, Sheila and Jeremy Oppenheim (2008), “Cultivat- ing the Green Consumer,” Standford Social Innova- tion Review, (Fall), 56–61. Conger, J. (1997), “How Generational Shifts Will Trans-

Marketing and Public Policy Conference Proceedings / 2009 189 form Organizational Life,” in The Organization of Effect of Ecological Concern on Brand Perceptions,” the Future, Frances Hesselbein, Marshall Goldsmith, Journal of Marketing Research, 10 (May), 191. Richard Beckhard, eds. San Francisco, CA: Jossey- ______, ______, and Sadrudin Ahmed Bass, 17–24. (1974), “Ecologically Concerned Consumers: Who Cornwell, T. Bettina, and Charles Schwepker, Jr. (1995), Are They?” Journal of Marketing, 38 (April), 20–24. “Ecologically Concerned Consumers and Their Prod- Krebsbach, Karen (2005), “The Green Revolution,” uct Purchases, in Environmental Marketing: Strate- USBanker, 115 (12), 28–32. gies, Practices, Theory, and Research, Michael Lee, Kaman (2008), “Opportunities for Green Marketing: Polonsky and Alma Mintu-Wimsatt, eds. 119–53. Young Consumers,” Marketing Intelligence & Plan- Cui,Yanli, Elizabeth S. Trent, Pauline M. Sullivan, and ning, 26 (6), 573–86. Grace N. Matiru (2003), “Cause-Related Marketing: Littrell, Mary, Ma Yoon Jin, and Jaya Halepete (2005), How Generation Y Responds,” International Jour- “Generation X, Baby Boomers, and Swing: Market- nal of Retail & Distribution Management, 31 (6), ing Fair Trade Apparel,” Journal of Fashion Mer- 310–20. chandising and Management, 9 (4), 407–19. Cutler, Neil (2005), “Saving Too Much for Retirement: Lipke, David J. (2001), “Green Homes,” American De- Are There Generational Differences?” Journal of mographics (January), 50–55. Financial Service Professionals, 50 (6), 28–31. Menon, Ajay and Anil Menon (1997), “Enviropreneurial Diamantopoulos, Adamantios, Bodo B. Schlegelmilch, Marketing Strategy: The Emergence of Corporate Rudolf R. Sinkovics, and Greg Bohlen (2003), “Can Environmentalism as Market Strategy,” Journal of Socio-Demographics Still Play a Role in Profiling Marketing, (January), 51–68. Green Consumers? A Review of the Evidence and an Meredith, Geoffrey and Charles Schewe with Janice Empirical Investigation,” Journal of Business Re- Karlovich (2002), Defining Markets, Defining Mo- search, 56, 465–80. ments: American’s 7 Generational Cohorts, Their Ellen, Pam Scholder, Josh Wiener, and Cathy Cobb- Shared Experiences, and Why Businesses Should Walgren (1991), “The Role of Perceived Consumer Care. New York: Hungry Minds, Inc. Effectiveness in Motivating Environmentally Con- Mobley, Amy, Todd Painter, Erich Untch, and H. Rao scious Behaviors,” Journal of Public Policy and Unnava (1995), “Consumer Evaluation of Recycled Marketing, 10 (Fall), 102–17. Products, Psychology & Marketing, 12 (May), 165– Gardyn, Rebecca (2003), “Eco-Friend or Foe,” American 76. Demographics, 25 (8), 12–13. O’Donovan, Cheryl (1997), “The X Styles,” Communica- Grankvist, Gunne, Ulf Dahlstrand, and Anders Biel (2004), tion World, (December/January), 17–20. “The Impact of Environmental Labeling on Con- Peattie, Ken and Andrew Crane (2005), “Green Market- sumer Preference: Negative and Positive Labels,” ing: Legend, Myth, Farce or Prophesy?” Qualitative Journal of Consumer Policy, 27, 213–30. Market Research: An International Journal, 8 (4), Hart, Stuart L. (1997), “Beyond Greening: Strategies for 357–70. a Sustainable World,” Harvard Business Review, Pickett-Baker, Josephine and Ritsuko Ozaki (2008), “Pro- (January/February ), 66–76. Environmental Products: Marketing Influence on Henion II, Karl, and William Wilson (1976), “The Eco- Consumer Purchase Decision,” Journal of Consumer logically Concerned Consumer and Locus of Con- Marketing, 15 (5), 281–93. trol, in Ecological Marketing, Karl Henion and Tho- Raymond, Joan (2001), “All Juiced Up,” American De- mas Kinnear, eds. American Marketing Association, mographics, (January), 42–44. Chicago, IL, 131–44. Rodriguez, Raul, Mark T. Green, and Malcolm James Johnson, Scott and Denise Johnson (1995), “Eco-Atti- Ree, “Leading Generation X: Do the Old Rules tudes and Eco-Behaviors in the New German States: Apply?” The Journal of Leadership and Organiza- A 1992 Perspective,” in Environmental Marketing: tion Studies, 9 (4), 67–75. Strategies, Practices, Theory and Research, Michael Samdahl, D. and R. Robertson (1989), “Social Determi- Polonsky and Alma Mintu-Wimsatt, eds. 119–53. nants of Environmental Concern: Specification and Jurkiewicz, Carole (2000), “Generation X and the Public Test of the Model,” Environment and Behavior, 21 Employee,” Public Personnel Management, 29 (1), (1), 57– 81. 55–74. Schewe, Charles and Geoffrey Meredith (2004), “Seg- ______and Dan Burr Bradley (2002), “Genera- menting Global Markets by Generational Cohorts: tional Ethics: Age Cohort and Healthcare Execu- Determining Motivations by Age,” Journal of Con- tives’ Values,” HEC Forum, 14 (2), 148–71. sumer Behaviour, 4 (1), 51–63. Kinnear, Thomas C. and James R. Taylor (1973), “The Schlegelmilch, Bodo, Greg Bohlen, and Adamantios

190 Marketing and Public Policy Conference Proceedings / 2009 Diamantopoulos (1996), “The Link Between Green Van Liere, K. and R. Dunlap (1981), “The Social Bases of Purchasing Decisions and Measures of Environmen- Environmental Concern: A Review of Hypotheses, tal Consciousness,” European Journal of Marketing, Explanations, and Empirical Evidence,” Public Opin- 30 (5), 35–55. ion Quarterly, 44 (2), 181–97. Seligman, C., M. Kriss, J.M. Darley, R.N. Fazio, L.J. Vitell, Scott J. and James Muncy (2005), “The Muncy- Becker, and J.B. Pryor (1979), “Predicting Summer Vitell Consumer Ethics Scale: A Modification and Energy Consumption from Homeowners’ Attitudes,” Application,” Journal of Business Ethics, 62 (3), Journal of Applied Social Psychology, 9, 70–90. 267–75. Smith, Stephen and Curtis Haugtvedt (1995), “Implica- Walley, Noah and Bradley Whitehead (1994), “It’s Not tions of Understanding Basic Attitude Change Pro- Easy Being Green,” Harvard Business Review, (May/ cesses and Attitude Structure for Enhancing Pro- June), 46–52. Environmental Behaviors, in Environmental Mar- Walsh, Bryan (2008), “Eco-Buyer Beware,” Time, 172 keting: Strategies, Practice, Theory and Research, (12), 71–72. Michael Polonsky and Alma Mintu-Wimsatt, eds. Webster, Jr., Frederick E. (1975), “Determining the Char- 155–76. acteristics of the Socially Conscious Consumer,” Solman, Gregory (2008), “We’re Not as Green as We Journal of Consumer Research, 2 (December), 188– Think,” Adweek, 49 (21), 7. 96. Straughan, Robert and James Roberts (1999), “Environ- “Widely Held Attitudes to Different Generations” (2008), mental Segmentation Alternatives: A Look at Green PR. (August 20). Consumer Behavior in the New Millennium, The Zimmer, Mary, Thomas Stafford, and Marla Royne Journal of Consumer Marketing, 16 (6), 558–75. Stafford (1994), “Green Issues: Dimensions of Envi- “The ‘Bigger is Better’ Dream of America is Getting Greener ronmental Concern,” Journal of Business Research, Too” (2005), Marketing Week, (October 6), 32. 30, 63–74.

For further information contact: Gail M. Zank Texas State University – San Marcos McCoy Hall 424 601 University Drive San Marcos, TX 78666 Phone: 512.245.3196 Fax: 512.245.7475 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 191 SPECIAL SESSION

CONSUMPTION CONSTRAINTS: HIGHLIGHTING PERSPECTIVES FROM THE SPRING 2009 SPECIAL ISSUE OF JPP&M WITH DIRECTIONS FOR FUTURE RESEARCH

Session Chair: Stacey Baker, University of Wyoming, Laramie

Session Participants: Gene Laczniak, Marquette University, Milwaukee John D. Mittelstaedt, Clemson University Julie L. Ozanne, Virginia Tech University, Blacksburg Nicholas Santos, Marquette University, Milwaukee

SUMMARY have visual impairments; 80 percent of the world’s population lives in poverty; one in five people in the U.S. is This session is designed to stimulate thinking on how a functionally illiterate. Thus, some global citizens, the perspective on constraint or restriction will illuminate or majority actually, live under conditions of restriction. influence marketing’s contributions to society. Whether the volume of articles across all marketing journals or the substantive focus of articles in marketing’s top journals is The presenters in this session tackle consumption con- assessed, the perception derived is that, in general, what straints from three different perspectives, while offering the marketing field cares about is creating or managing sound implications for marketers and policy makers. The abundant resources. Marketing and its sub-disciplines first presentation tackles business perspectives on how to tend to focus on (1) managerial problems such as creating relieve consumption constraints. The authors offer an efficiencies in scale, expanding the market, expanding integrative justice model on how businesses can authen- products in the market, or creating relationships with the tically engage consumers while co-creating value with the most profitable customers; (2) customer issues such as ultimate purpose of developing sustainable business ini- decision-making amongst a set of alternatives, informa- tiatives, empowering customers, and creating a fairer tion processing, or meaning making; or (3) darkside marketplace (Santos and Laczniak 2009). Though the consumer problems including materialism, compulsive presenters demonstrate application of this model within consumption, or greed. To be sure, these are exceedingly the context of poverty, the model applies equally well to important issues with great theoretical and practical sig- other situations of constraint. The second presentation nificance. However, for the most part, these concepts focuses on constraint from a marketing system perspec- imply excess and at their theoretical core these concepts tive by addressing constraints of access – access to what share a common assumption – people, firms, places, etc. and for whom – within the context of healthcare in the have abundant resources (information, assortments, etc.) U.S. (Mittelstaedt, Mittelstaedt, and Kilbourne 2009). which can be leveraged. The third presentation discusses the philosophies and methods of a deliberative democracy which can be used to What would happen if we threw out assumptions of uncover the diverse interests of stakeholders surrounding bountifulness? What potential does marketing have if we particular marketing activities (Ozanne, Corus, and engage people not typically represented in our research Saatcioglu 2009). Methods and applications of delibera- samples? For example, one in five Americans have a tive democracy are discussed within the context of disability of some sort; 160 million people worldwide sustainability and other marketing and society issues.

192 Marketing and Public Policy Conference Proceedings / 2009 Special Session: Consumption Constraints: Highlighting Perspectives from the Spring 2009 Special Issue of JPP&M with Directions for Future Research

Each presenter co-authored a paper in the Spring (2009) presenters will highlight their approach and then push issue of Journal of Public Policy & Marketing. The forward by discussing future research opportunities.

REFERENCES (2009), “The Philosophy and Methods of a Delibera- tive Democracy: Implications for Public Policy and Mittelstaedt, John D., Charles Duke, and Robert A. Marketing,” Journal of Public Policy & Marketing, Mittelstaedt (2009), “Healthcare Choices and the 28 (Spring), in press. Constrained Consumer: A Marketing Systems Per- Santos, Nicholas and Gene Laczniak (2009), “Marketing spective on Access and Assortment in Healthcare,” to the Poor: An Integrative Justice Model for Engag- Journal of Public Policy & Marketing, 28 (Spring), in ing Impoverished Segments,” Journal of Public Policy press. & Marketing, 28 (Spring), in press. Ozanne, Julie L., Canon Corus, and Bige Saatcioglu

For further information contact: Stacey Menzel Baker Department of Management and Marketing College of Business University of Wyoming 1000 E. University Avenue Laramie, WY 82071 Phone: 307.766.3734 Fax: 307.766.3488 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 193 SPECIAL SESSION

PATIENT CENTERED MEDICAL HOME: PROSPECTS FOR TRANSFORMATION OF HEALTHCARE

Session Chair: Debra L. Scammon, University of Utah, Salt Lake City

Discussant: Kevin Grumbach, University of California, San Francisco

Panelists: Julie Day, University of Utah Community Clinics, Salt Lake City Edwina Rodgers, Patient Centered Primary Care Collaborative, Washington, D.C. Sam Butler, Epic Systems Corporation, Madison

SESSION OVERVIEW will specifically speak to the policy issues that must be addressed to facilitate implementation of medical homes. There is little doubt that healthcare in the United States is in crisis. Patients have little trust in the system, compen- The concept of a medical home is not new. The term was sation schemes have become counter-productive, and first coined in 1967 and further refined in 1992 by the providers are increasingly disenchanted because of time American Academy of Pediatrics (AAP) as an “approach and financial pressures that push them toward shorter to providing comprehensive primary care” for children. visits and mechanistic modes of care. In order to deal with Recently the Association of American Medical Colleges the multitude of current issues nothing short of a transfor- suggested that “every person should have access to a mation of healthcare is required. Medical Home” (AAMC March 2008). Four primary care physician specialty societies – the American Academy of One transformation being considered is implementation Family Physicians (AAFP), the American Academy of of a Patient Centered Medical Home (PCMH) defined as Pediatrics (AAP), the American College of Physicians “a physician-directed practice that provides care that is (ACP) and the American Osteopathic Association (AOA) – accessible, continuous, comprehensive, coordinated, and have adopted a statement of Joint Principles for the delivered in the context of family and community.” Prac- PCMH (2007). The concept has been endorsed by a range tice redesign, payment reform, interoperable electronic of purchasers, labor, and consumer organizations. The medical records, and effective collaboration of payers and Patient Centered Primary Care Collaborative (PCPCC) providers are but some of the challenges facing PCMHs in represents the interests of 300+ organizations that support transforming healthcare. The purposes of this session are implementation of the concept. There is a growing body to: (1) explore the potential for PCMHs to reduce costs of evidence showing that care delivered by primary care and improve quality of healthcare, and (2) highlight the physicians, supported with information systems and ap- administrative and policy issues that must be addressed to propriate reimbursement incentives, can improve the qual- achieve their potential. Panelists will provide background ity of care provided to patients and benefit practice effi- on the medical home initiative and how it fits with the ciency and financial performance (e.g., Deloitte 2008). current agenda for healthcare reform, discuss the practical challenges and potential impacts of practice redesign, The PCMH model could dramatically improve the patient’s introduce new models for reimbursement to compensate experience in receiving healthcare. The model is based on providers, and highlight the role of EMRs in the effective an ongoing relationship between a primary care provider coordination of care. As appropriate, panelists will com- and patient built upon respect for the patient/family’s pare the PCMH model to other conceptualizations of a cultural and religious beliefs. The PCMH is intended to redesigned health care delivery system. The discussant help coordinate care, including that from needed special-

194 Marketing and Public Policy Conference Proceedings / 2009 Special Session: Patient Centered Medical Home: Prospects for Transformation of Healthcare ists as well as community services. A plan of care devel- for meeting these standards performance. Centers for oped by the physician, patient, and family that is shared Medicaid and Medicare Services (CMS) is considering with other providers, agencies, and organizations in- reimbursing practices that qualify. A number of PCMH volved with the care of the patient is integral to effective demonstration projects, based upon collaborative efforts coordination. In order to facilitate access to medical among payers, purchasers, and providers, are evaluating consultation when and where it is needed, the PCMH various pay-for-performance compensation schemes concept encourages non-visit patient interactions such as linked to accomplishments related to these standards. phone or electronic communications. The PCMH should help patients navigate the disconnected parts and often Another essential component of the PCMH model is daunting complexity of the health care system. health information technology to facilitate coordination of care. Well-designed electronic medical records (EMRs) Many of the features of continuity, coordination, and can provide clinical decision support for best practice, communication that are at the heart of PCMH will require links to specialty and community resources, and support the redesign of healthcare delivery processes and the for patient self-management. President Obama’s inclu- introduction of new services not compensated in current sion of stimulus funding for health information technol- reimbursement models. Thus, payment reform will be ogy is evidence for the critical role it plays in comprehen- essential. The National Committee for Quality Assurance sive and coordinated care. The expanded use of EMRs (NCQA) has recently introduced a Practice Recognition raises important policy questions including privacy of and program to certify practices that meet a set of standards security for personal health information and the need for related to the components of a PCMH. Such certification standards to ensure interoperability of information could provide a basis for monetary incentives to practices systems.

For further information contact: Debra L. Scammon University of Utah 1645 Campus Central Drive Salt Lake City, UT 84112 Phone: 801.581.4754 Fax: 801.581.3152 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 195 LOSS OF BALANCE? AN EXAMINATION OF OBESITY, INDIVIDUAL RESPONSIBILITY, AND DEEP CAPTURE IN NEW ZEALAND

Janet Hoek, University of Otago, New Zealand Rachael McLean, University of Otago, New Zealand

SUMMARY Some governments have proposed a compromise that has seen them work jointly with industry to develop obesity- Governments may try to change consumers’ behavior via reduction initiatives. New Zealand’s Food Industry Ac- education, which promotes greater knowledge and aims cord (FIA), in which signatories from food manufacturing to increase self-efficacy, or by regulation, which directly companies declared their commitment to producing com- alters their environments. Public health groups typically mercially viable products that helped reduce obesity, support regulation; however, industry groups favor initia- epitomizes this joint approach. However, the FIA has tives such as education, which promote greater individual concerned the public health community, whose members responsibility; these divergent views have been apparent believe government endorsement of an industry program whenever debate over obesity prevalence measures has designed to pre-empt regulation may reduce its ability to occurred. protect public health.

The ecological approach promoted by public health re- This risk has been described as “deep capture,” and occurs searchers aims to create environments that make “the when governments “pass laws . . . that benefit the entities healthy choice the easy choice” (World Health Organiza- they oversee rather than the public they represent” (Al- tion 1986). This approach is theoretically linked to behav- derman, Smith et al. 2007, p. 102). The New Zealand FIA ior modification and assumes consumers’ choices depend provides a timely vehicle for exploring “deep capture” as largely on the stimuli that precede and reinforce these Food Industry Group (FIG) documents reveal the group (Nord and Peter 1980). Among environmental factors has worked closely with the Ministry of Health. While a thought to contribute to obesity, researchers have noted close working relationship enables integration of industry the lower cost of energy dense foods, differences in food and government initiatives, it may also affect the imple- availability and affordability, and increases in the style, mentation of the government’s agenda by aligning it more sophistication and intensity of marketing used to reach closely with industry interests. While collaboration with consumers (Story, Kaphingst et al. 2008). These changes the food industry (and other industries) may bring ben- have allegedly created an “obesogenic” (Egger and efits, it also exposes regulators to risks that could compro- Swinburn 1997) environment that public health research- mise their role and function. Government thus needs a ers argue could most effectively be changed via regula- framework to assess these risks and whether the potential tion. This reasoning focuses on external determinants of benefits outweigh the possible costs. Such a framework behavior and is the antithesis of the “individual responsi- should consider the benefits collaboration brings, how bility” argument, which suggests obesity could be avoided these will be distributed, and whether similar benefits are if individuals assumed greater responsibility for their available elsewhere. It should also review the opportuni- behavior. ties created by collaboration, how these are shared, and how collaboration could affect other stakeholders. Food manufacturers have promoted individual responsibility by supporting “healthy eating” education programs, Application of this framework to the FIG-government funding physical activity initiatives, and adopting addi- relationship suggests the food industry will derive more tional front-of-pack nutritional labeling. However, sev- benefits, particularly deferment of regulation, an outcome eral researchers have criticized programs such as these; they are otherwise unlikely to realize. However, govern- Jochelson (2006) argued that education could “have only ment will benefit by avoiding direct conflict, although the some impact on knowledge and attitudes, and little impact cost of this appears to be slower change, the cost of which on behavior” (p. 1153). Health promotion models also requires further analysis, and weaker relationships with suggest that, while environments promote unhealthy foods, public health groups, who have argued that industry information will have little influence on behavior. alliances have weakened government’s ability to regulate.

196 Marketing and Public Policy Conference Proceedings / 2009 Specific risks have arisen from the FIG’s opposition to an tioned whether this initiative did more than sanction the external regulatory framework for food advertising, which profits of two companies whose core business is selling food industry spokespeople argued should not be intro- high-fat-salt-sugar food to communities that are unequally duced because it “conflicted” with understandings they affected by deprivation and obesity. had with government. Similar themes are apparent in a voluntary agreement between the NZ government, Coca Although the longer term effects of collaboration are not Cola Amatil New Zealand and Frucor Beverages Group, apparent, our analysis highlights how relationships with where both companies agreed to stop selling full-sugar industry groups may lead regulators to compromise their carbonated soft drinks in New Zealand schools. While role and responsibility. Although the examples outlined described as “World first” in Coca Cola Amatil’s press above appear consistent with an environmental model, release, the withdrawal of sugary carbonated drinks will government’s agenda appears shaped more by industry not take place until 11 December 2009, and, as part of the than consumers’ interests and the long time lines have agreement, the government undertook to pursue alliances deferred change that regulation could have introduced with industry in preference to regulation. “Deep capture” more rapidly. Frameworks that assess the risks and ben- also appears evident in a partnership between Coca-Cola, efits industry-collaboration might deliver will become McDonald’s and a health authority that saw sugar-free increasingly important when potential collaborators are lemonade substituted for full-sugar lemonade in 21 required to pursue profit maximization rather than public McDonald’s restaurants. Public health advocates ques- health objectives. References are available upon request.

For further information contact: Janet Hoek Department of Marketing University of Otago P.O. Box 56 Dunedin New Zealand Phone: +643.479.7692 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 197 WHEN TO SAY WHEN? SELF-REGULATORY RITUALS FOR EATING AND DRINKING

Vanessa Gail Perry, The George Washington University, Washington, D.C. Joonwhan David Lee, The George Washington University, Washington, D.C.

SUMMARY Our analysis revealed that participants used a variety of self-imposed rules to regulate the content and quantity of Nutrition and alcohol consumption are important issues the foods they eat. Participants relied on three main for consumer and public policy researchers due to the approaches in order to regulate their intake of food and significant implications of controlling these behaviors for alcoholic beverages: portion control heuristics; physical individuals and society. Alcohol consumption, referred to signals (e.g., feeling “full”); and consumption rituals. here as “drinking,” has been a longstanding concern due However, when consuming alcohol, participants often to alcoholism and drunk driving, and more recently, binge failed to engage their self-imposed regulatory strategies. drinking. At the same time, there has been a flurry of In order to control portions, many participants used fre- recent attention in the popular media, among policymakers, quency heuristics or cutoffs, such as a certain number of and researchers on nutrition, food choice and consump- drinks. One limitation of frequency heuristics in decision- tion volume (e.g., Wansink 2004; Wansink et al. 2005). making is that this information does not include specific The purpose of this research is to explore what strategies details about the event (Alba and Mamorstein 1987). consumers use to control their consumption of food and These portion control efforts often take the form of alcohol. In addition, we investigate the relationship be- precommitments, which Hoch and Loewenstein (1991) tween alcohol consumption and food consumption. We suggest can be used as a self-control device. These authors draw upon previous research on consumption rituals and claim that these strategies are difficult for consumers to self-control to develop a framework for understanding execute, and are often ineffective. Previous research has consumers’ mechanisms to control their food and alcohol examined the effects of portion control efforts on food intake and reasons for failures of those controls. consumption, and has found that many heuristics used for this purpose are imprecise and unreliable (e.g., Glanz In order to determine what strategies people use to control 1998; Painter et al. 2002). In fact, plate size and design, food and alcohol consumption and how alcohol influ- salience, and convenience all limit the effectiveness of ences food consumption decisions, the two co-authors those self-control techniques (e.g., Painter et al. 2002). and 45 student collaborators interviewed a convenience Thus, educating consumers about the drawbacks of these sample of 89 participants (Spiggle 1994). The participants mechanisms might help consumers improve or enhance included 49 males and 40 females from a variety of their self-control strategies. professional backgrounds, ranging from 18 to 60 years of age. Interviewers were trained to use a standardized in- Eating and drinking behaviors are often ritualistic in depth interview protocol and on probing techniques. nature and individuals often follow specific rules about Statements from interview transcripts were entered into a where, how much, and what to eat and drink (Rook 1985; spreadsheet to facilitate categorization. Interpretations of Treise, Wolburg, and Otnes 1999). Eating behaviors that themes were derived independently by the co-authors occur after drinking also share characteristics of other directly from these data. These analyses were iterative-- consumption rituals. In our interviews, many participants involving independent coding, categorizing, abstracting claimed that they eat foods they normally avoid due to and clustering the data, similar to the approach recom- nutrition and weight concerns when drinking alcohol. mended by Spiggle (1994) and used by Schouten (1991). These participants reported that they believe that alcohol Next, two independent coders categorized each of the consumption causes them to make poor eating decisions. transcript statements. The correlation coefficients for Furthermore, some participants claimed that they control intercoder reliability were assessed by intraclass correla- what they eat by controlling how much alcohol they drink. tion (ICC), and a fixed rater design (3, K) was used in the These interviews also imply that consumers may inten- current study to eliminate rater variance in the denomina- tionally use drinking as a self-justification for unhealthy tor (Glick 1985; Bruton et al. 1994). eating decisions.

198 Marketing and Public Policy Conference Proceedings / 2009 Another prevailing theme in these interviews was that menting self-control strategies. At a result, once alcohol is people crave foods that contain high levels of cholesterol, involved, food choices and even consumption volume fat, and sugar when drinking alcohol, including fried decisions are driven by short-term, impulsive, hedonic foods, carbohydrates, “fast” and “junk” foods. One obvi- urges. ous explanation for this is that alcohol consumption can lead to intoxication, which by definition, results in im- Future research should examine the effectiveness of these paired judgment (Baumeister et al. 1994). However, these self-control strategies, and should test whether these self- interviews suggest that even before individuals become control mechanisms are evident in other types of con- intoxicated they experience ego depletion (Baumeister sumption, such as consumer credit usage. Also, future 1998). According to Baumeister (1999), willpower is a research should further our understanding of simulta- kind of limited resource that gets depleted as used. In other neous efforts to exercise self-control and the effects on words, exercising self-control over both drinking and high-risk behaviors, such as smoking, unprotected sex, or eating depletes the cognitive resources available for imple- other kinds of impulsive consumption.

For further information contact: Vanessa Gail Perry Marketing Department 301 Funger Hall The George Washington University 2201 G Street, NW Washington, D.C. 20052 Phone: 202.994.4190 Fax: 202.994.8999 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 199 VISCERAL CUES AND CONSUMER SKEPTICISM: CONSUMER REACTIONS TO WEIGHT LOSS ADVERTISING

Clinton Amos, Augusta State University Stacy Landreth Grau, Texas Christian University, Fort Worth

SUMMARY decision to purchase the weight loss product, thus consumers should experience higher levels of buying impulse The presence of visceral factors is a phenomenon of recent when visceral cues are present. Second, the visceral cues interest to marketing and public policy (Lynch and should lead consumers to focus more on the reward Zauberman 2006). Visceral factors are rewards that be- components of the advertisement. This effect is expected come progressively more attractive as they appear to draw due to the attention narrowing power of visceral cues nearer in time, appeal to a state of deprivation or despera- (e.g., Loewenstein 1996). Hence, when visceral cues are tion, and appear hedonically salient (e.g., Loewenstein present, consumers should affectively react to the images 1996). In particular, the dissemination of information can and information related to reward obtainment. The abbe affected by the presence of visceral stimuli. The confir- sence of visceral cues should result in a more cognitive mation of such behavior has tremendous implications for assessment of the advertisement including any warnings/ marketing and public policy as society becomes more disclaimers. The number of non-reward thoughts after attuned to its visceral impulses (Bell et al. 2007). viewing an advertisement should be greater when visceral cues are not present. Consequently, it is anticipated that Recent research has posited that persuasive-minded indi- exposure to advertisements with visceral cues will result viduals often employ visceral stimuli to generate rapid- in an increase reward thoughts and a decrease in non- affective responses, thus misleading consumers (e.g., reward thoughts. Langenderfer and Shimp 2001). In particular, the presence of visceral stimuli may result in the complete disre- Furthermore, the anticipated effects of the visceral cues gard of pertinent information, even for individuals con- should be relatively unaffected by ad skepticism. One’s sidered invulnerable. This study contributes to the litera- resistance to visceral cues is determined by one’s own ture by using the theory of visceral influences to assess volition regardless of level of knowledge (Loewenstein reactions to visceral cues (temporal and hedonic cues 1996). As a result, regulators’ efforts to disseminate associated with a reward) in weight loss advertisements in information (i.e., disclaimers, warnings) to counter the an effort to explore whether visceral cues lead to reward- effects of persuasion attempts using visceral cues are fixation by respondents, regardless of ad skepticism lev- likely ineffective (Lynch and Zauberman 2006). els. Method and Results Theoretical Background Two experiments were conducted. In the first experiment, The theory of visceral influences (Loewenstein 1996) the main effect of each manipulated visceral cue was distinctively explains why certain drive states (e.g., sexual examined along with measured levels of ad skepticism desire, fatigue, etc.) are commonly associated with impul- (α = .90). Using a sample of 176 respondents, buying sive behavior, why sensory appeal when short time delays impulse levels were captured on a three-item scale (α = will result in an appetitive response. It predicts that certain .87). Additionally, respondents were asked to list all stimuli can create an attention narrowing effect and inten- thoughts, feelings, or impressions of the ad. Two indepen- sify emotions associated with a reward leading to an dent coders then coded the responses into two categories: impulsive response. Hence, the theory of visceral influ- reward thoughts (e.g., positive thoughts about the visceral ences serves as a predictor of situational impulsivity. cues), and non-reward thoughts (e.g., skeptical thoughts about the ad). Inter-rater reliability between the coders for For weight loss ads, it is anticipated that the ability of the categorical placement of thoughts was .85. MANOVA visceral cues to trigger visceral influence will have the results show that all three visceral cues were significantly following impact. First, the effect of the increased levels (all p-values < 0.02) associated with greater buying im- of visceral influence should lead to a more impulsive pulse. Both the visual prime and vividness of reward were

200 Marketing and Public Policy Conference Proceedings / 2009 associated with fewer non-reward thoughts (p-values < Discussion 0.05) while only the visual prime was associated with more reward thoughts (p < 0.05). Multivariate results The strength of this research stems from its contributions show that the interactions between ad skepticism and the to public policy. First, this study shows that skeptical three visceral cues were not significant (p-values ranged consumers are still susceptible to visceral cues in weight from 0.07 to 0.68). Univariate results confirm that there loss advertising. While vulnerability may be reduced for were no significant interactions across any of the depen- skeptics, the three studies indicate that vulnerability does dent variables (p-values ranged from 0.07 to 0.81). not vanish with higher levels of ad skepticism. While policy initiatives are often placed to protect vulnerable In the second experiment, the combined effects of the consumers, this suggests that policy makers should avoid visceral cues were examined in conjunction with ad focusing only on populations they feel are the most skepticism. Eighty respondents viewed either a visceral vulnerable and instead consider the implications for a weight loss ad (containing all three visceral cues) or a non- wider set of consumers. This research suggests that at- visceral ad (without visceral cues). MANOVA results tempts at increasing consumer knowledge about effective show that respondents viewing ads with visceral cues had means of weight loss may not reduce the vulnerability of a stronger buying impulse (p < 0.001, more reward thoughts consumers. The ability of visceral cues to generate intense (p < 0.001), and fewer non-reward thoughts (p < 0.01). reward-laden emotions appears to override cognition and When examining the interaction between visceral cues increase the vulnerability of even the wariest of consum- and ad skepticism the results were as hypothesized. Mul- ers. Hence, mandatory disclaimer information may be tivariate results indicate that the interaction between the ineffective since both skeptical and non-skeptical respon- visceral manipulation and ad skepticism was not signifi- dents focused less attention toward the non-reward com- cant (p > 0.30). Univariate results confirm that there were ponents of the ad and more attention toward the reward no significant interactions across any of the dependent components of the ad when visceral cues were present. In variables (p-values range from 0.06 to 0.46). other words, the presence of visceral cues reduces the probability that an individual with take notice of disclaimer information. References are available upon request.

For more information: Clinton Amos Hull College of Business Augusta State University 2500 Walton Way Augusta, GA 30904 Phone: 706.667.4546 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 201 AN INVESTIGATION INTO INDIVIDUAL’S REPEATED ATTEMPTS AT BEHAVIOR CHANGE

Courtney Droms, Valdosta State University

SUMMARY attributions also act as input to attitudes toward trying again and as such, hypotheses have been developed to Recent research has shown that individuals make repeated examine several combinations of dimensions that will attempts at behavior change prior to actually being suc- have the most impact on attitudes. cessful. For example, typically individuals try to change their dieting behaviors on 4 to 5 occasions prior to actually Self-esteem is defined as in individual’s subjective ap- being successful (Polivy and Herman 2002). Many of the praisal of him or herself (Sedikides and Gregg 2003), dark side of consumer behavior areas identified by Mick which can be influenced by both goal achievement and (1996), such as smoking, drug use, shopping, and gam- goal failure (Crocker et al. 2003). As such, individuals in bling, are behaviors that individuals try to change on a this study are expected to have lower levels of self-esteem repeated basis. As consumer researchers, however, we do post-failure rather than post-success at trying to change not have a clear understanding of how people interpret behaviors. Hypotheses were also developed and tested to behavior change failures and persist in their efforts to capture the complexity of the interplay between self- change their behaviors after such failures. In an effort to esteem and attributions. Finally, an individual’s self- understand these behavior change attempts, this research esteem is also hypothesized to have a direct effect on that replicates and extends the Theory of Trying (Bagozzi and individual’s attitudes toward trying to change in the Warshaw 1990), to include feedback loops from the future. outcomes of behavior change attempts to attitudes toward trying to change again. MacInnis and de Mello (2005) define hope as a positive emotion attached to a goal. There are three important Theoretical Model dimensions to hope: goal congruency (i.e., the extent to which an environment is conducive to fulfilling a specific The theoretical model is based on the model developed in goal), certainty (i.e., the level of confidence an individual the Theory of Trying (Bagozzi and Warshaw 1990). The has in being able to achieve his or her goal), and impor- theory of trying is employed in this context because it tance of the goal. All three of these dimensions should be includes two variables that seem very relevant to repeated affected by the individual’s success or failure at trying to attempts at changing behaviors (i.e., the frequency and change behaviors, such that failure at trying to change recency of prior attempts at changing behaviors). How- behaviors should lower an individual’s level of hope. ever, the model proposed by Bagozzi and Warshaw (1990) Hope is also hypothesized to affect the ways in which ends at trying to change behaviors. While the author does individuals understand and attribute the causes of their suggest that the outcomes of these behavior change at- behavior change outcomes. Finally, individuals with high tempts feedback into attitudes, they do not explore the levels of hope should have a more positive outlook on the nature of the feedback. In order to explore the nature of the possibilities present in the future, which should enable feedback from outcomes to attitudes, this paper employs them to have more positive attitudes toward changing in three constructs – attributions for outcomes, self-esteem, the future. and hope. Each of these factors will now be briefly discussed. Methodology

Weiner’s (1974) theory of attributions posits that there are An internet survey of 363 people from across the United three dimensions to attributions – locus, stability, and States who are currently trying to diet was completed. controllability – that assist individuals in understanding Modifying the procedure employed by Bagozzi and the causes of their behaviors and behavioral outcomes. In Warshaw (1990), data was collected in two stages, one the model, specific hypotheses, consistent with prior month apart. Participants were asked to complete a series literature on attributions and the self-serving bias, have of measures in the first stage designed to capture their been developed that explicate how outcomes for success attitudes and intentions to try, the frequency and recency and failure are attributed along each dimension. These of prior attempts at trying to change behavior, and their

202 Marketing and Public Policy Conference Proceedings / 2009 current levels of self-esteem and hope. At stage two, one stability and controllability dimensions of attributions month later, individuals completed a second survey in- having a significant positive effect on their attitudes cluding measures of trying, the outcome of trying, and toward trying again in the future. In addition, the results attitudes toward trying again in the future. Participants show that the individual’s attributions mediate the effect will also complete measures capturing their attributions that hope has on their attitude toward changing in the for the outcome of their attempts at behavior change, as future. Moreover, I find that attributions moderate the well as their updated levels of self-esteem and hope. effect that the outcome of their behavior change attempt has on the individual’s level of self-esteem. Such that, Results and Discussion individuals who make stable attributions for failure show a substantial decrease in their self-esteem compared to Overall, this research enlightens the discussion of how those who make unstable attributions for failure. Finally, people repeatedly change their health behaviors. The the results also indicate that the individual’s change in results indicate that individuals attribute the cause of their self-esteem from before to after their behavior change success or failure at their most recent behavior change attempt has a direct effect on their attitude toward trying attempt in such a way that it enables them to have a more again, such that individuals whose self-esteem decreased positive attitude toward trying to change again in the after their behavior change attempt have a higher attitude future. The results from structural equation modeling toward trying to change again in the future than their show that the participants in this study attributed their counterparts whose self-esteem increased post-behavior failure to achieve their behavior change goal to causes that change. There are important implications for both market- were unstable and controllable which resulted in both the ing and public policy as a result of this study.

REFERENCES Mick, David Glen (1996), “Are Studies of Dark Side Variables Confounded by Socially Desirable Re- Bagozzi, Richard P. and Paul R. Warshaw (1990), “Try- sponding? The Case of Materialism,” Journal of ing to Consume,” Journal of Consumer Research, 17 Consumer Research, 23 (2), 106–19. (September), 127–40. Polivy, Janet and C. Peter Herman (2002), “If at First You Crocker, Jennifer, Andrew Karpinski, Diane M. Quinn, Don’t Succeed: False Hopes of Self-Change,” Ameri- and Sara K. Chase (2003), “When Grades Determine can Psychologist, 57 (9), 677–89. Self-Worth: Consequences of Contingent Self-Worth Sedikides, Constantine and Aiden P. Gregg (2003), “Por- for Male and Female Engineering and Psychology traits of the Self,” in Sage Handbook of Social Psy- Majors,” Journal of Personality and Social Psychol- chology, M.A. Hogg and J. Cooper, eds. London: ogy, 85, 507–16. Sage Publications, 110–38. MacInnis, Deborah J. and Gustavo E. de Mello (2005), Weiner, Bernard (1974), Achievement Motivation and “The Concept of Hope and Its Relevance to Product Attribution Theory. Morristown, NJ: General Learn- Evaluation and Choice,” Journal of Marketing, 69 ing Press. (January), 1–14.

For further information contact: Courtney Droms Langdale College of Business Valdosta State University 1500 N. Patterson Street Valdosta, GA 31698 Phone: 229.245.3848 Fax: 229.245.2248 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 203 FAILURE TO UNDERSTAND COMPOUND INTEREST, AND MITIGATING STRATEGIES

Eric M. Eisenstein, Temple University, Philadelphia Stephen Hoch, University of Pennsylvania, Philadelphia

SUMMARY tive framing (e.g., debt, loans, and inflation vs. interest, raises, stocks), and in differences between retrospective Many important financial decisions hinge on a proper and prospective compounding estimation (i.e., what was understanding of compound interest. For example, deci- the cost of a TV in 1970? vs. what will be the cost of a TV sions on how much and when to invest in a 401(k) plan, in 2030?). Furthermore, we examined the effects of exper- whether to refinance a loan, or how to trade off various tise and training on accuracy. The normative theory is attributes of a loan instrument (rate, points up front, simple: given an interest rate, i, and a time horizon, n, the duration, etc.), cannot be made without a proper under- formula governing compound interest is FV = PV (1 + i)t, standing of compounding. Furthermore, some judgments where FV and PV are, respectively, the future value and that enter into assessments of well-being and happiness, present value of the investment, i is the interest rate, and such as whether prices or wages are fair (cf., Bolton, t is the term. Warlop, and Alba 2003), also depend on a proper understanding of compounding. Results demonstrated that the vast majority of people anchor on simple interest (i.e., FV = PV [1 + i t]), and then When consumers approach such decisions, however, their adjust their answer upwards. Reliance on simple interest understanding of this critical concept is frequently quite was demonstrated both by analysis of participants’ open- poor. This problem is exacerbated because the geometric ended responses and by the pattern of coefficients of nature of compounding is counter-intuitive even to those individual-level models that were fit to respondents’ esti- who are intellectually familiar with the underlying theory. mates. Anchoring on simple interest yielded acceptable Furthermore, to accurately choose an option in many of accuracy for short time periods and for low interest rates, the financial domains outlined above would require sub- however, massive underestimation resulted for longer stantially greater spreadsheet modeling expertise than is time horizons and for higher rates. Most of the major possessed by the vast majority of consumers, which financial decisions that people make during their lives means that there will frequently be a substantial intuitive involve either long time periods (e.g., retirement saving, component to consumer decision-making in these do- mortgages) or high interest rates (e.g., taking on credit mains. card debt), which makes anchoring on simple interest a pernicious bias. A limited number of studies (e.g., Wagenaar and Timmers 1967, 1969) have investigated people’s ability to intu- Results of Experiment 2 revealed that financial expertise itively predict the outcome of an exponential growth resulted in only slightly increased accuracy. Furthermore, process. Most of these experiments were conducted in educational level (i.e., years of education) and major had unfamiliar domains (such as the growth of duckweed on no effect on accuracy; undergraduate students in business, a pond) rather than in the more familiar financial domain. MBA students, and a general population of consumers all An unfamiliar domain is likely to reduce the ability of made errors of similar magnitude. The magnitude of error participants to transfer any expertise that they might was also unaffected by negative vs. positive framing (i.e., possess – an effect that has been demonstrated repeatedly debt vs. investments). Additional results demonstrated in other tasks (e.g., the Wason card task). Furthermore, that all respondents found retrospective estimation (i.e., if previous studies have not specifically looked for expertise there are FV dollars in the account today, what PV was effects or examined heterogeneity in the psychological invested n years ago at a constant interest rate, i?) more process used to make decisions. difficult than prospective estimation.

In three experiments, we explored the psychological un- Although most participants used simple interest as an derpinnings of consumer estimation of compound interest anchor and were very inaccurate, a small minority were in a heterogeneous population. In particular, we examined very accurate in their estimates. Open-ended responses differences in processing arising from negative vs. posi- and individual level coefficient estimates revealed that

204 Marketing and Public Policy Conference Proceedings / 2009 these subjects used a completely different estimation In Experiment 3, we trained participants to use the rule of process, the “rule of 72.” The rule of 72 is an accurate 72. Compared to an untrained control group, the trained approximation of the correct financial formula, the core of condition realized enormous improvements in accuracy, which is the realization that y = 72/i, where y is the number with no statistically significant increase in time spent on of years that it takes for money invested (or loaned) at an the task. interest rate of i to double in value. If the time horizon is known, the rule of 72 can be used to approximate the Expertise in estimating compound interest is strongly effects of compounding by computing the number of linked to the use of the rule of 72. If consumers anchor on doublings implied by (i, t), and then doubling the principle simple interest, they are likely to make errors that have that number of times. For example, if PV = $1,000, i = 9 massive impact on their financial well being, including percent, t = 24 years, then an exact computation shows underestimating the cost of long-term loans (such as FV = $7,911. To use the rule of 72, compute 72/9 = 8, mortgages), saving too little and too late for retirement, which demonstrates that money at 9 percent will double in and allowing credit card debt to revolve, rather than approximately 8 years. Since the time horizon is 24 years, aggressively paying it down. Furthermore, the results this represents 24/8 = 3 doublings, which will result in a shed light on the difficulty of learning nonlinear func- 3 FV72 = 2 * $1,000 = $8,000 (error = 1%). Note that simple tions, even for those with substantial experience with interest would provide an anchor of FVSI = $1,000 * 9 them. Consumers can be taught to be considerably more percent * 24 = $2,160 (error = 72%). accurate in their understanding of compounding in a very short amount of time.

REFERENCES Nisbett, R.E. and T.D. Wilson (1977), “Telling More than We Can Know: Verbal Reports on Mental Processes,” Benartzi, S. and R. Thaler (2004), “Save More Tomor- Psychological Review, 84 (3), 231–59. row: Using Behavioral Economics to Increase Em- Wagenaar, W.A. and S.D. Sagaria (1975), “Misperception ployee Savings,” Journal of Political Economy, 112 of Exponential Growth,” Perception & Psychophys- (1), 165–87. ics, 18 (6), 416–22. Bolton, L.E., L. Warlop, and J.W. Alba (2003), “Con- ______and H. Timmers (1978), “Extrapolation sumer Perceptions of Price (Un)Fairness, Journal of of Exponential Time-Series Is Not Enhanced by Consumer Research, 29 (March), 474–90. Having More Data Points,” Perception & Psycho- Jones, G.V. (1979), “Generalized Polynomial Model for physics, 24 (2), 182–84. Perception of Exponential Series,” Perception & ______and ______(1979), “The Pond- Psychophysics, 25 (3), 232–34. and-Duckweed Problem; Three Experiments on the Keren, G. (1983), “Cultural-Differences in the Misperception of Exponential Growth,” Acta Misperception of Exponential-Growth,” Perception Psychologica, 43, 239–51. & Psychophysics, 34 (3), 289–93.

For further information contact: Eric M. Eisenstein Fox School of Business Temple University 527 Alter Hall Philadelphia, PA 19122 Phone: 215.204.7039 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 205 THE EFFECTIVENESS OF REAL ESTATE FORECLOSURE AUCTIONS: A CASE STUDY AND POLICY IMPLICATIONS

Ekaterina V. Karniouchina, Chapman University, Orange Debra L. Scammon, University of Utah, Salt Late City William L. Moore, University of Utah, Salt Lake City Dan A. Fuller, Weber State University, Ogden

ABSTRACT auctions (Hagerty 2007). Several methods for selling a large number of lender foreclosed properties in a short The rapid deterioration of the housing market in 2007 and period of time have been observed, including “foreclosure 20081 brought about a wave of property foreclosures and bus tours” organized by real estate agents2 and large the failure of several prominent financial institutions. This auctions held at convention centers around the country. paper uses observational and secondary data to investi- Property disposition auctions are taking place all over the gate the process and outcome of a real estate foreclosure United States (see [http://www.ushomeauction. com] for auction, one method banks are using to dispose of fore- examples of upcoming auctions and property listings); closed properties. these auctions are the primary focus of this study. The case described in this paper presents an example of an auction INTRODUCTION that sold a large number of properties in a short period of time. Real estate markets are highly cyclical and often inefficient. Case and Shiller (1998) argue that periods of excess Previous research (e.g., Pennington-Cross 2006) finds demand and supply may arise for various reasons includ- that foreclosed homes sell at a discount. Discounts may be ing “notions of fairness, intrinsic worth, popular theories based on a number of factors. Owners at risk of default about prices, coordination problems, and simple mistakes may spend less on maintaining their properties. Owners (p. 29). The current downturn, attributed to multiple” threatened with foreclosure may also damage homes factors including financial innovations in risk manage- before eviction. Pennigton-Cross (2006) finds that the ment, the growth of fee-based mortgage brokerages, mar- longer a lender holds onto a piece of property after default, keting of alternative mortgages terms (such as no money the larger the foreclosure discount, but he does not specifi- down, adjustable rates, interest only payments), and lack cally examine the auction as a means of clearing fore- of regulatory oversight, has deepened due to rapid defla- closed homes. tion of housing prices. Accompanied by a general recession, the rate of mortgage delinquencies is soaring. Banks are likely rational when deciding which properties go to auction. Previous research (Vandell and Riddiough As borrower delinquencies of over 90 days mount, the 1990) examined the use of auctions in real estate disposi- role of lenders has changed from delinquency manage- tion and concluded that traditional ascending bid English ment to loss mitigation. There are several options avail- type auctions may be unsuccessful due to a lack of buyer able including loan workouts, loan modifications, pre- interest. On the other hand, auctions provide an opportu- foreclosure sales, and foreclosures. Although the pre- nity to target entry level buyers. The Colorado Indepen- ferred alternative depends upon individual circumstances, dent3 quoted a Wells Fargo representative saying: “We do foreclosure generally results in the largest loss to the sell some of our foreclosure homes through auctions. It’s lender. However, in the current housing cycle, wide- a pretty effective way to reach potential homebuyers, and spread perceptions of negative equity have led an increas- we’ve found that real estate auctions are a pretty good way ing number of borrowers to opt for foreclosure. of attracting first-time homebuyers.”4 To help stimulate sales, the federal government is currently considering a In the current housing market, banks are flooded with an $15,000.00 tax incentive to homebuyers. oversupply of foreclosed assets (Palmeri 2008) and need tools for their disposition. Banks with foreclosed proper- Governments at all levels are increasingly concerned with ties are increasingly inclined to discount properties sold in the growing stock of unoccupied housing. At the local the traditional real estate channels or sell properties through level, not only are taxes adversely impacted, but the

206 Marketing and Public Policy Conference Proceedings / 2009 perceived attractiveness, and hence economic viability, of across the country. At this event (which was a part of a localities likely suffer as well. At the national level, the broader five-day Northern California event), 222 homes drop in the value of assets relative to the liabilities held by were sold to the highest bidder in a little over eight hours banks (outstanding loans) has precipitated a credit crunch (from 9:30 AM to 5:40 PM). A total of 277 sales were that has thrown the economy into a recession, raised the completed including 44 re-sales (described later). Two potential liabilities of various government insurance pro- researchers participated in observational research. The grams, and generated large and increasing budgetary information about the homes came from the promotional deficits. Thus, the processes that the private sector uses to brochures printed for this auction and the information draw down the excess inventory of unoccupied housing is provided during the auction itself. of significant interest. Auctions are unique because of the large numbers of properties that can be sold in a short In traditional real estate markets, agents attempt to secure period of time. the maximum transaction price given that a buyer can be found in a reasonable amount of time (Yavas and Yang This research relies upon observational methods and 1995). With auctions, the auction house wants to extract secondary data sources to examine the process and effi- the highest price in order to maximize its commission and ciency of a real estate foreclosure auction. We conducted also to improve the odds of the lender approving a given participant observations throughout various steps of the transaction.5 At the same time the auction house wants foreclosed property auction process, from touring the every property to be sold which explains very low starting properties offered for sale to attending the actual auction bids (some starting bids were as low as $1,000). event. Grove and Fisk (1992) synthesize previous research on observational methods and note several advan- The event was handled by the Real Estate Disposition tages of this method that are pertinent to this research. Corporation (REDC).6 The company sells homes for Observational methods (1) complement more traditional different lenders. It is responsible for carrying out the data collection methods, (2) serve as means of discovery advertising campaign (television and radio spots, and research question formulation, (3) offer an “up close” infomercials, telemarketing, distribution of printed mate- view of the phenomena, and (4) provide information rials) to ensure bidder turnout. The company charges a 5 about phenomena that are gleaned in their natural setting. percent commission, refered to as a “premium,” that is added to the winning bid and thus paid by the buyer. This Observations reveal that for a buyer, a foreclosure auction is somewhat unusual since in most auctions the commis- sale is very different from a conventional real estate sion is paid by the seller. If the sale does not go through for transaction. Specifically, the auction is very much struc- any reason, the commission is returned to the buyer. tured as a “caveat emptor” transaction. Information for the Financing, for most properties, was available from a buyer is often limited and/or costly to gather. Analysis of number of financial companies (e.g., Countrywide, Wells sales data indicates that the concentration of foreclosures Fargo, etc.) who had representatives at the auction. Initial in particular neighborhoods has important social and credit checks were performed at the onsite escrow sta- economic implications for communities leading to declin- tions. About 14 percent of the properties were sold on a ing home values. “cash only” basis, often the case with low value properties which fall under the minimum value for bank mortgages. The next section of the paper presents an overview of a lender foreclosure event. Following that, analyses of the Most properties offered in the auction event were either transactions for over 200 properties offered at one auction distressed or properties in markets with extremely low are presented. We examine selling prices, returns com- demand, i.e., areas hit exceptionally hard by economic pared to previous values, and price formation dynamics. malaise. Better properties in more attractive markets are We conclude with a discussion of policy implications and sold by conventional means. In the San Mateo County future research directions. event, there were no properties from Palo Alto, Milpitas, Santa Clara or any other prime real estate markets. The OBSERVATIONAL ANALYSIS very few properties located in more attractive markets (e.g., two properties located in San Francisco metro area, We present a summary of the North San Mateo Area one in Berkley) were run down condos or small houses Foreclosed Home Auction held at San Mateo County with virtually no land, in need of serious repairs. How- Events Center on July 19, 2008. Although some of the ever, in recently overbuilt areas (i.e., commuter towns specific requirements for buyers and sellers vary from built in anticipation of workforce growth in major metro- venue to venue, this auction is similar to those being held politan areas), one could find surprisingly nice, relatively

Marketing and Public Policy Conference Proceedings / 2009 207 new homes in the action pool. All properties were sold “as steps back because the winning bidder thought he/she was is,” the only guarantee the new homeowner gets is the title bidding a different amount compared to what the auction- insurance. eer had in mind. Selling prices ranged from $40,000 to $695,000 while “previous values” ranged from $110,000 As opposed to exploring carefully staged properties smell- to $1.326 million. Descriptive statistics are provided in ing of flowers and home-baked cookies in the company of Table 1. accommodating real estate agents willing to show the properties at virtually all hours, prospective buyers in an The auction line up changed significantly between the auction face higher information costs and uncertainties. time the houses were listed in the promotional brochures We visited several properties listed for auction. Often and the day of the auction. Over one quarter (27%) of the properties were available for inspection only during very auctioned homes were new additions to the auction line short open house hours. It is likely some buyers resorted up, while 15 percent of the homes originally scheduled to to external inspection only. Information was often lim- be auctioned off were removed from the auction. This is ited. There was often only a temporary worker on site and common in distressed markets where a large number of no one who could answer questions regarding the prop- new properties become available due to new foreclosures erty. Properties showed signs of decay (e.g., dying lawns, while some homes tentatively scheduled for auction are water stains, signs of recent fumigation, dead mice, evi- sold via alternative channels. These updates to the auction dence of looting, etc.). Prospective buyers are encouraged list create additional issues for prospective buyers as to bring along professional appraisers when they inspect information on new additions is only available through homes but they bear the cost of doing so. REDC provides the auction house website. a “previous value,” based on the higher of (a) the ap- praised value of the property in connection with the most Contrary to common belief that investors are speculating recent mortgage on the property, (b) the most recent on foreclosures, according to REDC (as well as our asking price, (c) the assessed value, or (d) the most recent observations) most buyers (over 80%) are people who broker price opinion. intend to live in the purchased properties.7 Our observations revealed that having a lot of people making major Many attendees were new to auctions: by the show of spending decisions in a short period of time contributed to hands before the event about one third of the participants making this event very emotional: from bickering spouses, had never attended an auction of any kind prior to the to disappointed outbid buyers, to tears of joy in the eyes of event. While REDC offers an optional orientation semi- first time home buyers who did not envision owning the nar prior to the auction, there were occasional hiccups in California properties under the pre-meltdown prices. the bidding process that were quickly and efficiently resolved by the award winning REDC auctioneers. For In a credit crunch, it is harder and harder for people with instance, people were often confused by the bidding lower credit scores to obtain financing. Although REDC increments, and could not follow the auctioneer “chant.” encourages loan pre-approval, a significant portion of In fact several times the auctioneer had to take several successful bidders was unable to close the transaction.

TABLE 1 Descriptive Statistics

Minimum Maximum Mean Std. Deviation

Square Ft. 494 3976 1600.80 710.159 Cash Only 0 1 .14 .343 Bedrooms 1 8 3.25 1.104 Bathrooms 1 4 1.98 .779 Previous Market Value $110,145 $1,326,000 $458,657.79 $165,188.43 Sale Price $40,000 $695,000 $199,664.41 $112,712.90 New Offering 0 1 .27 .447

208 Marketing and Public Policy Conference Proceedings / 2009 The winning bidder was either unable to prove the ability efficiency of the auction as a mechanism for bringing to come up with the cash or was denied a loan during the prospective buyers and sellers together. onsite credit checks. As a result, many properties were sold several times during the event. In these instances, the ANALYSIS OF SALES DATA re-bid sales price was almost always lower with only 8 out Examining Returns. We define “return” as the ratio of of 44 homes in a second sale fetching higher prices than sales price to the ‘previous value’ estimate provided by the first sale. Furthermore, over 21 percent of these second REDC. On average condominiums fetched a lower return sale properties failed in the (post-auction) credit applica- than single family homes. Single family residences (SFR) tion stage. These properties are typically scheduled to be were sold at 44 percent of previous value; while condos resold at an “insider event” (according to REDC, typically were sold at 36 percent (the difference is significant at between 18 and 20%), where people who attended the 95% confidence level). For the handful of multi-unit auction but did not get a property of their choice get a properties, the difference between their mean return and second chance to bid on unsold properties. that of other housing types was not significant (see Tables 2 and 3). In addition to our observational data, we collected data on sale prices, estimates of previous value provided by the “Location, location, and location” is a popular real estate auction house, descriptors of the property (e.g., square motto. Other research has suggested that location is im- footage, location, property type), and some factors related portant in determining the discount associated with fore- to how the property was auctioned off (cash only, repeat closed properties. Discounts on foreclosed properties are sale, etc.). Turning now to these data, we look at the (1) usually higher in areas where inventories of unsold homes auction returns, (2) sales price dynamics throughout the have soared (Hagerty 2007). Local job and population auction (returns over the course of the auction), as well as growth also help explain the severity of the price declines (3) failed deals and repeat sales to gain insights into the (Hagerty 2007).

TABLE 2 Mean Returns by Property Type

Type Mean Std. 95% Confidence Interval Error Lower Bound Upper Bound

Condo 36.578 2.956 30.753 42.404 Multi-Unit 37.775 5.329 27.272 48.277 SFR 44.179 1.108 41.995 46.363

TABLE 3 Comparisons for Auction Returns vis-à-vis Previous Market Value by Property Type

Mean 95% Confidence (I) Type (J) Type Difference (I-J) Std. Error Sig. Interval for Difference

Condo Multi-Unit -1.196 6.093 .845 -13.207 10.814 SFR -7.600(*) 3.157 .017 -13.822 -1.379 Multi-Unit Condo 1.196 6.093 .845 -10.814 13.207 SFR -6.404 5.443 .241 -17.131 4.323 SFR Condo 7.600(*) 3.157 .017 1.379 13.822 Multi-Unit 6.404 5.443 .241 -4.323 17.131

* The mean difference is significant at the .05 level.

Marketing and Public Policy Conference Proceedings / 2009 209 To reflect the nesting of the data by location characteristic overall reliability of level-1 coefficients across the 48 while identifying the drivers of foreclosure home returns, level-2 units is equal to 0.747. Table 4 shows that location we specify a hierarchical linear model (Raudenbush and effect is highly significant (p < .001) and is responsible for Bryk 2002) where the properties are grouped within their 58 percent of the difference in returns.8 respective zip codes. At level-1 (or house level) we specify an equation describing the return on house i in zip In addition, we found several significant fixed effects. code j: Larger homes sold at a higher return while homes with P higher previous values sold at a lower return. At a foreclo- yij 0 j B papij ij (1) sure event, it is likely that many people are trying to get p1 into affordable homes on the low value end but they also try to buy as much space as they can for their dollar. β 0j Intercept or average return for house within zip code j Cash only transactions were significantly discounted. This could be traced to the fact that banks often decline to β p coefficient for house characteristic p finance these properties due to their deteriorated state. In addition there are some straightforward demand side α pij value of characteristic p for house i within dynamics, there is only a limited group of people who can zip code j afford to pay for these properties without additional financing and out of those who can, many would prefer to ε σ2 ij ~ N(0, ) Level-1 within cell random effect that pay more and buy a property in better condition. We got represents the deviation from the predicted a negative coefficient for new offers indicating that less outcome based on the level-1 descriptors. demand was generated for houses that had been added to the auction late (less awareness and less time to examine At level-2 (or zip code level) we specify the equation for the property) but this difference was not significant (see the zip code mean: Table 5). 0 j 00 u0 j (2) Examining “Declining Price” Anomaly. The declining price anomaly is concerned with the observation that in γ Where 00 is the mean overall return which could be multiple item auctions, prices tend to decline over time interpreted as grand intercept and u0j is the random effect (Ashenfelter 1989). Generally auction scholars have found associated with zip code j assumed ~N (0,τ). The final evidence of a price decline anomaly during various auc- mixed model can be written as: tions such as wine auctions and auction-based condo sales P (Ashenfelter and Genesove 1992; Ashenfelter 1989). It is ij 00 0 j B pauy pij ij (3) not surprising to find a price decline anomaly during the p1 auction sales of new condos since these sales often operate Estimation of equation (3) is based on 219 units in 48 on the “right to choose” or the pooled principle, where the different zip codes that had sufficient data (only the first highest bidder gets to pick his/her property out of the sales are included in estimation, re-bids are analyzed later remaining pool of condos. In essence, the earlier bidders in the paper). All continuous variables were mean cen- buy the right to choose. tered prior to entering them into the model. Estimation was carried out using the restricted maximum likelihood Unlike the case of wine sold through auctions (in which method; conversion was achieved in 11 iterations. The the items sold are identical) described in Ashenfelter

TABLE 4 Summary of Variance Components for Random Zip Code Effects

Random Effect Std. Deviation Var. Comp. D.f. Chi-Squared p-value

ZIP CODE 11.62390 135.11506 47 373.02177 0.000

ERROR 9.88956 97.80342

210 Marketing and Public Policy Conference Proceedings / 2009 TABLE 5 Final Estimation of Fixed Effects

Fixed Effect Coefficient Standard Error T-ratio d.f. P-value

INTERCEPT 53.6856 2.0929 25.6 47 0.000 SQ. FT. 0.0145 0.0015 9.9 213 0.000 CASH ONLY -10.6923 2.2392 -4.8 213 0.000 PREVIOUS VALUE -0.000035 0.000007 -5.2 213 0.000 NEW OFFER -1.9388 1.7350 -1.1 213 0.265 CONDO -6.6905 2.5980 -2.6 213 0.011

(1989), homes are different and one needs to control for trend toward either return increase or reduction; instead it house level heterogeneity. Even when it comes to hetero- seems that prices become slightly less predictable (at least geneous products, some studies find evidence of price as described by our model).9 Regression analysis shows declines from early to late in the auction which could be that the absolute value of the Empirical Bayes (EB) attributable to, among other factors, the decline in demand residual increases .02 percent with each additional sale throughout the auction. For example Beggs and Graddy (p < .001).10 This could be caused by some buyers becom- (1997) established the existence of the “declining price ing “desperate” to get one of their desired properties as the anomaly” in art auctions (also see Ashenfelter and Graddy auction nears its completion. At the same time as more and 2003) while controlling for art piece characteristics. When more potential buyers do get their properties, the pool of it comes to real estate auctions Ong, Lusht, and Mak potential bidders gets smaller. Some properties may fail to (2005) find that bidder turnout influences auction out- generate interest and may consequently be picked up by comes tying the number of interested parties to the price investors of the property. There is some evidence of price reduction if we look at the In the case of distressed home auctions, controlling for 44 properties that were re-sold at the auction. Table 6 location is critical since auction order is tied to location. shows that on average the price fell by over ten thousand Therefore, instead of simply looking for trends in sale dollars from first to second sale resulting in a significant prices or return estimates, we use the residuals from model pair wise comparison test. The price drops were in excess (3) that capture location and home-specific variation to of what would be expected if the second bidder got the study the possible existence of the price-decline effect property at his/her initial reservation price since the last after controlling for these salient factors. We do not see a price increment for most houses was $5,000.

TABLE 6 Paired Samples Test for 44 Properties That Were Sold Twice

Paired Differences T Df Sig.

Std. Std. Error 95% CI of the Mean Deviation Mean Difference

Lower Upper

FIRST SALE – SECOND SALE 10227.27 18737.93 2824.85 4530.42 15924.12 3.62 43 .001

Marketing and Public Policy Conference Proceedings / 2009 211 Examining Supply and Demand Dynamics. There is a overall level of foreclosures in the area. The Foreclosure strong relationship between the number of properties Index variable was introduced at Level-2. This factor offered for the auction sale in the same zip code and the captures the increased supply of houses due to foreclo- returns. Figure 1 indicates that with a larger number of sures, overall decline in economic conditions in the area, homes in a particular zip code, average returns experience and housing values. We obtained the number of foreclo- a sharp decline. sures per 1000 homes for Bay Area zip codes. The data were obtained from SF Gate/San Francisco Chronicle.11 This could be indicative of the discount associated with Not all zip codes are covered in their database, therefore zip codes hit the hardest by the real estate meltdown. the analysis was performed on 213 homes in 43 zip codes Therefore, we introduced a new measure to control for the (see Table 7).

FIGURE 1 Empirical Bayes Estimates of Average Returns for Specific Zip Codes vs. Number of Homes Offered from the Same Zip Code12

80.000 San Fracisco Daily City

70.000 Martinez Penngrove

Novato Rodeo 60.000 Rohnert Park Sonoma East Palo Alto Kelseyville Vacaville Santa Rosa 50.000 Concord Concord Oakley Santa Rosa Vacaville Middleton Richmond Vallejo Vallejo Antioch Richmond Brentwood 40.000 Fairfield Antioch Bay Point

EB Estimate of Zip Code Specific Effect Specific Code Zip of Estimate EB Vallejo

Richmond 30.000

0.00 5.00 10. 00 15.00 20.00 25.00 Number of Offerings

Introducing this variable reduces the explanatory power the sale price, the greater the chance that the sale will not of the zip code effect to 25 percent12 indicating that the be finalized. There could be two reasons for this phenom- presence of other foreclosed properties in the area ex- enon. Low income buyers willing to occupy cheap homes plains about half of the between neighborhood price as a could have problems obtaining financing. Only initial proportion of previous value differences.13 screening is conducted at the event and subsequent credit and reference checks may result in lenders’ decisions to Examining Failed Deals. We also collected data on the decline the loans. While banks are often reluctant to property sales that did not materialize. We examined the finance low value housing, they may also be reluctant to experience on those properties that came back on the approve the transaction if their loss on the resale is too market a month after they originally sold. The only great. Therefore, one needs to consider the return vis-à-vis variable linked to the odds of the house not being sold was the previous home value. Additional analysis revealed the sale price. The results of a logistic regression that that this ratio is not informative. Therefore, it is more explains about 7% of the variation indicates that the lower likely that the deals fell through because of potential

212 Marketing and Public Policy Conference Proceedings / 2009 homeowners’ inability to qualify for a loan rather than foreclosure, job gains/losses, quality of life metrics, school banks’ refusal to accept the sale. This issue needs further ratings, etc. for a given neighborhood. This sort of index investigation. could help banks meet the requirement that they carry assets (loans) on their books at a reasonable market value MAJOR FINDINGS, IMPLICATIONS, AND (“mark to market”). It could also help local governments FUTURE RESEARCH as they attempt to forecast local housing market conditions. Our data indicate that the return or discount to an auctioned property is sensitive to location as measured by zip Our research has implications for the consumer price code. In addition, by introducing a simple foreclosure judgment and reference pricing literatures as we uncov- index that captures the proportion of foreclosed properties ered a price decline effect for the resale of individual in an area, we found that the foreclosure index accounted properties. Our data suggest that there may be an anchor- for about one-half of the explained variance in returns. ing and adjustment effect such that initial closing bids The development of a more sophisticated distress index provide a ceiling for future bids. A property may be resold should improve the ability to forecast property values. if the seller does not accept the initial bid or if the buyer Such an index should capture the number of homes in does not qualify for financing or was not able to demon-

TABLE 7 Model with Foreclosure Index at Level-2

Final Estimation of Fixed Effects:

Fixed Effect Coefficient Std. Error T-ratio d.f. P-value

INTERCEPT 54.390097 1.531538 35.513 41 0.000 FORECLOSURE INDEX -1.369276 0.170112 -8.049 41 0.000 SQ. FT. 0.014964 0.001409 10.620 204 0.000 CASH ONLY -10.374223 2.135690 -4.858 204 0.000 PREVIOUS VALUE -0.000039 0.000007 -5.729 204 0.000 NEW OFFER -1.592835 1.678386 -0.949 204 0.344 CONDO -7.206146 2.449535 -2.942 204 0.004

Final Estimation of Variance Components:

Random Effect Std. Dev. Variance Component D.f. Chi-Squared p-value

ZIP CODE 5.72001 32.71854 41 103.78901 0.000 ERROR 9.85119 97.04597

TABLE 8 Sale Price as Determinant of House Coming Back to the Auction

B S.E. Wald D.f. Sig. Exp(B)

PRICE ($K) -.006 .002 9.017 1 .003 .994 Constant -.271 .350 .597 1 .440 .763

Nagelkerke R2 .075

Marketing and Public Policy Conference Proceedings / 2009 213 strate ability to pay the “cash only” price. The former We present data about one real estate foreclosure auction. situation could be avoided if the seller was required to set Future studies should look at the longitudinal impact of a minimum acceptable bid prior to the start of the auction. real estate auctions, e.g., it would be interesting to see if, The latter problem could be averted if buyers were re- over time, previous auction prices in a given zip code had quired to be pre-qualified for financing. Regardless of an impact on subsequent auction prices. Future studies what might explain the price decline, it is important to should also look at a cross-section of different auctions. anticipate the negative impact on the value of a property Currently individual auction houses set their auction that is resold. One way to minimize the potential for an rules, subject to state regulations and laws. A fruitful anchoring effect would be to offer properties on which the avenue for future research would be analysis of the results deal failed at an auction held at a later point in time. of auctions with different rules. Such comparative analyses could help inform public policy. Clearly, the economic Our observations and analyses of the data from the auction efficiency of auctions is impacted by the information suggest that there may be several obstacles to buyers available. The REDC auction was conducted primarily as making informed decisions that may impact the allocative a caveat emptor transaction. Not only was minimal infor- efficiency of the auction market. These include buyer- mation available about the condition of the properties, but seller information asymmetry, pressures of an unfamiliar “previous value” lacks objectivity when market prices are bidding environment, and severe time pressure in making rapidly changing. It is curious that previous values were the final purchase and financing decisions. Buyers are often adjusted upward when we compared the original likely to have incomplete information about the condition promotional brochures to the information presented at the of properties since professional inspections of all proper- auction.14 One possibility to reduce information ties of interest may be cost prohibitive. REDC does assymetries is to supply prospective buyers with recent, provide a real estate agent ‘walk though report’ for some independent, third-party inspection reports and apprais- properties but these reports are very superficial and are als. This would be especially helpful for homes that furnished by the lender. They seem to be very “kind” to the cannot be inspected prior to the auction because they are properties using terms such as “overgrown landscaping” tenant occupied at the time of the sale. A related question, to describe a backyard that will require complete renova- then, is who would bear these additional costs. REDC tion. Understandably, buyers may be reluctant to pay for rules specify that winning bidders pay 5 percent of the professional inspection for each property of interest since purchase price to cover auction expenses. The provision they do not know if they are going to be the highest bidder. of the additional information may justify an increase in As a result, a lot of defects are uncovered ex post (e.g., this payment since the winning bidder presumably would house extension was done without a permit, the plumbing capture the benefits of the extra information. needs to be replaced or there is a termite infestation). During the auction, the auctioneers stress previous value In order to minimize market deterioration and the social and overwhelm bidders with the “auctioneer’s chant” but and economic impact of foreclosures,15 thought should be avoid making any substantive disclosures about the prop- given to crafting a government homeowner incentive erties. This emphasis on previous value may increase package to assist those purchasing homes through fore- another source of sellers’ power, that is, the sale is final on closure auctions. For instance the universal first time the buyer’s but not on the seller’s end. The bank has the home buyer tax credit (part of the Housing and Economic right to say “no” but the buyer does not have such a right Recovery Act of 2008) could be structured so that it without losing the earnest money. Some states have imple- supports purchases of properties in particularly depressed mented cooling-off-periods to handle such problems. housing markets. Bidders eligible for the tax credit could receive a pre-qualification verification enabling them to Clearly, the issue of failed sales is problematic to both bid on properties up to some limit price. Developing a seller and buyer. As noted, 20 percent of the auction sales more sophisticated distress index reflective of local con- in this event did not result in completed transactions; these ditions (as described above) could aid such efforts. Such properties are later offered at an “insiders’ event.” Future an index should be very informative in trying to predict research should investigate the returns obtained on such real estate prices for foreclosed as well as regular proper- delayed sales. If returns are significantly lower, it may be ties. beneficial to require prospective buyers to pre-qualify for credit and sellers to set minimum acceptable sales prices Beyond fine-tuning the real estate auction to improve the prior to the auction. efficiency of the transaction, it is important to understand

214 Marketing and Public Policy Conference Proceedings / 2009 the circumstances that contributed to the stock of fore- tributed to the foreclosure problem. Questionable lending closed properties currently being offered at auction. Fu- practices including lax scrutiny of customer credit-wor- ture research should investigate the terms of the loans thiness and speculation during periods of rising real estate originally made on foreclosed properties. Loan options values, as well as economic factors such as falling prop- such as those with no money down, interest only repay- erty values and job losses have been tied to the current ment requirements, or adjustable interest rates, intro- financial crisis. Future research should carefully examine duced as product innovations by lenders may have con- the role that these factors may have played.

ENDNOTES rules may also discourage excessive speculation by requiring a higher percentage of the purchase price 1 Palmeri (2008) states that Standard & Poor’s/Case- (15%) as earnest money for subsequent homes pur- Shiller home price index reported sharp declines in chased in one auction. home prices in 2007 and 2008, the steepest decline in 8 Estimated as suggested in Raudenbush and Bryk (2003) the two-decade history of the index. by dividing the variance component associated with 2 See example at http://www.repohometour.com/ (last the random effect by the total variance. downloaded on August 2, 2008). 9 Additional analysis shows that including the auction 3 http://coloradoindependent.com/view/foreclosed-home order variable in model 3 does not produce a signifi- (last accessed August 14, 2008). cant coefficient. 4 2005 America Housing Survey (AHS) data suggests 10 Adjusted R-squared =.06. that first-time buyers are on average, about 33 years 11 Online portal provides a searchable database of Q2 old, have a household income of almost $64,100, buy 2008 information. Database can be found at [http:// homes with a median price of $150,000 and account www.sfgate.com/webdb/quarterlyforeclosures/ for 43 percent of homes sold (Eisenberg 2008). 2008q1/?Search+Again=] (last accessed on August 5 The price has to be high enough for the lender who owns 14, 2008) the loan to approve the sale. If the bank feels the offer 12 Total variance is the variance attributable to the zip is too low, they reserve the right to refuse the sale. All code level (32.72) and error (97.05). This indicates properties were sold subject to lender approval. The that 25 percent or 32.72/129.77 of the overall vari- right of the seller to refuse a bid raises questions about ance lies between neighborhoods. the fairness of the process. At one recent auction in 13 No other changes are detected, the increase in bidding Salt Lake City, successful bidders were informed variance toward the end of the auction persists in this more than two weeks following the auction that the smaller dataset. lender-owner of the property did not accept their bid. 14 The list of changes to the original brochures was Neither the commission on the sale nor the buyer’s printed and distributed by REDC at the auction. earnest money had been returned and there was no These changes included significant upward adjust- indication as to when these funds would be returned ment to previous values of several properties. (KSL Nightly News, October 14, 2008). 15 See [http://stlouisfed.org/RRRSeries/event_5.html]. 6 This is a private company although its name is quite 16 Posterior means are much higher for zip codes with similar to the government agency that handled real only a few homes offered at the auction and are much estate transactions following the savings and loan lower for zip codes with a large number of offerings. crisis in the 1980, the Resolution Trust Corporation 17 The data were analyzed by zip code but the labels (RTC). where created using the city or town names for easier 7 Some of these individuals may still be speculating in the interpretation, therefore some names repeat. Some sense of planning to remodel or repair while living labels are suppressed for easier reading. there and resell in the future for a gain. REDC auction

Marketing and Public Policy Conference Proceedings / 2009 215 REFERENCES Street Journal Online, November 30, 2006. Ong, Seow Eng, Kenneth M. Lusht, and Chee Yong Mak Ashenfelter, Orley (1989), “How Auctions Work for (2005), “Factors Influencing Auction Outcomes: Wine and Art,” Journal of Economic Perspectives, 3, Bidder Turnout, Auction Houses and Market Condi- 23–36. tions,” Journal of Real Estate Research, 27 (2), 177– ______and David Genesove (1992), “Testing for 92. Price Anomalies in Real-Estate Auctions,” The Ameri- Palmeri, Christopher (2008), “They Can Go Home Again: can Economic Review, 92 (2), 501–5. Mortgage Bankers Who Contributed to Housing Mess ______and K. Graddy (2003), “Auctions and the Are Now Cleaning it Up,” Business Week, updated Price of Art,” Journal of Economic Literature, 41, 10:43 a.m. PT, Wed., March 26, 2008, [http:// 763–87. www.msnbc.msn.com/id/23812916/], last accessed Beggs, A. and K. Graddy (1997), “Declining Values and on October 22, 2008. the Afternoon Effect: Evidence from Art Auctions,” Pennington-Cross, Anthony N. (2006), “The Value of Rand Journal of Economics, 23, 544–65. Foreclosed Property,” Journal of Real Estate Re- Case, Karl E. and Robert J. Shiller (1988), “The Behavior search, 28 (2), 193–214. of Home Buyers in Boom and Post-Boom Markets,” Raudenbush, S.W. and A.S. Bryk (2002), Hierarchical New England Economic Review, Federal Reserve Linear Models: Applications and Data Analysis Meth- Bank of Boston, (November), 29–46. ods, 2nd ed. Newbury Park, CA: Sage. Eisenberg, Elliot (2008), “Characteristics of First-Time Vandell, Kerry and Timothy Riddiough (1990), “On the Home Buyers,” National Association of Home Build- Use of Auctions as a Disposition Strategy for the ers, Special Studies, (January 23). TRC Real Estate Assets: A Policy Perspective,” Gau, George, Daniel Quan, and Theodore Sternberg Housing Policy Debate, 3 (1), 117–41. (1990), “The Performance of Real Estate Auctions: A Wotapka, Dawn (2008), “Builders help buyers to help Case Study,” Working Paper, University of Texas at themselves,” Wall Street Journal Online, October 18 Austin. [http://online.wsj.com/article/ Grove, Stephen J. and Raymond P. Fisk (1992), “Obser- SB122428402382246349.html, last access on Oct vational Data Collection Methods for Services Mar- 22, 2008]. keting: An Overview,” Journal of the Academy of Yavas, Abdullah and Shiawee Yang (1995), “The Strate- Marketing Science, 20 (3), 217–24. gic Role of Listing Price in Marketing Real Estate: Hagerty, James (2007), “Distressed Real-Estate: Priced to Theory and Evidence,” Real Estate Economics, 23 Sell in 2007,” WSJ Real Estate Archives, The Wall (3), 347–68.

For further information contact: Ekaterina Karniouchina Argyros School of Business and Economics Chapman University One University Drive Orange, CA 92866 Phone: 714.289.2068 Fax: 714.532.6081 E-Mail: [email protected]

216 Marketing and Public Policy Conference Proceedings / 2009 RISK-BASED PRICING OR EXCUSE-BASED PRICING? AN ANALYSIS OF PENALTY REPRICING IN THE CREDIT CARD INDUSTRY

Joshua M. Frank, Center for Responsible Lending, Durham

SUMMARY People who referred to documents are found to report a significantly higher APR across a variety of loan types Penalty rates on credit cards have become more common including credit cards, car loans, and home equity loans. in recent years. Borrowers can be repriced to a penalty rate Credit card had the highest bias with a 3.3 percentage of they are late or have their credit score decline, with their point difference in reported non-teaser APR. Since the rate typically jumping substantially. This study examines SCF data does not specify whether somebody is at a the prevalence of penalty repricing, the level of consumer penalty rate, three different APR cutoffs were used as the awareness, and the amount of penalty rate increases. Data most likely proxies for a penalty rate: 22 percent, 24 on penalty repricing is estimated from the Survey of percent, and 25 percent. People who referred to docu- Consumer Finances (SCF). The SCF is a national tele- ments were significantly more likely to report a penalty phone survey of about 4,500 conducted every three years APR using all three cutoff levels. For the 22 percent APR by the Federal Reserve intended to provide highly de- cutoff, 8 percent of document-users reported a penalty tailed information on the financial status of individuals. It rate in the combined surveys while only 5 percent of those includes data on both assets and liabilities, including who answered “off-the-cuff” reported a penalty rate. credit card loan balances and prices. However, credit card balances have previously been found to be substantially Using the reduced population of those who have no other underestimated in SCF data (Zinman 2007) and general loans, people were 2.2 times as likely to report being at a evidence of “wishful thinking bias,” where people have penalty price when they referred to documents. This been found in other situations to remember information in implies that 54 percent of consumers who did not look at a more positive light than reality (Gordon et al. 2005; their statement did not know they were at a penalty rate. Bahrick et al. 1996; Greenwald 1980), suggests people When weighted by balances, 58 percent of penalty bal- may also underestimate their APR. ances were held by consumers who if they did not look at their statement would not know they were at a penalty rate. Within the SCF is a variable notated by the interviewer indicating whether the respondent referred to documents The extent to which borrowers are deceived by penalty in their answers that is used here but has not been utilized repricing is partially controllable by lenders through dis- in prior research. We compare the reported level of closure. Additional information on this topic was ana- penalty repricing between those who referred to their lyzed based on data collected by Mintel’s Compermedia, credit card statement and those who did not to estimate a database that contains direct mail from credit card how frequently people mistakenly believe they are not at issuers, including both new solicitations and mail sent to a penalty price. existing accounts. We examined data from 2006 through August 2008. In each of the 32 months in the study period, Unfortunately, the data in the SCF for whether a person 9,700 households are surveyed. From that entire period, referred to documents uses general categories that indi- only a single mailing was found that informed a customer cate the respondent referred to some document that might they were receiving an increased penalty price. This be a credit card statement. Using these more global flags suggests that penalty repricings are generally done with- will tend to underestimate the level of bias. To eliminate out notification. This in turn suggests that issuers seek to some of this underestimation effect, in some data consum- maximize the APR in ways that will receive the least ers with car loans or mortgages are excluded to obtain a consumer attention rather than seeking to reduce moral group of consumers who more likely referred to credit hazard. card documents in particular when referring to a loan document. While this could be a biased population, we The penalty rates have significantly increased in preva- only use it to extract a multiplier for “wishful thinking” lence both in terms of existing balances (using SCF data) bias to apply to the general population, and we do not find and terms in solicitations (using Comperemedia data). In evidence of bias in the level of the multiplier calculated. 2007 and 2008 through August, over 90 percent of solici-

Marketing and Public Policy Conference Proceedings / 2009 217 tations included a penalty rate. Furthermore, the penalty distress likely will cause default in at least some house- APR in solicitations has generally been rising over time, holds. both in terms of rate and spread over prime. There is also a growing disparity between the penalty APR charged Using the bias multipliers from the prior section times the when borrowers are repriced and the normal purchase percent of people in the survey who say they are at a APR. The difference between the penalty APR and the penalty rate without referring to documents, it is estimated normal purchase APR shown in solicitations has in- that between 9.5 percent and 12.1 percent of revolving creased in every year since 2000. In 2008, the difference balances are at a penalty rate. between the penalty APR and the normal (non-teaser) purchase APR in solicitations that show both rates was While we have tried to account for socioeconomic factors 16.9 percentage points. This is more than double the 8.1 in our analysis, it is still possible that some of the differ- percentage point difference that was observed in 2000. ences reported are from these factors. However, the evi- The difference is also higher than the purchase rate itself, dence appears strong that: (1) most people appear to be with the APR increasing 141 percent on average when unaware they are at a penalty rate, (2) issuers behavior is borrowers are put into the penalty box. For a household consistent with minimizing awareness, (3) penalty rates that has the average level of credit card debt, being penalty have become both more common and more severe in repriced on all their balances would result in an additional terms of rate jump. Combined, the evidence suggests that $1,800 a year coming out of that family’s pocket in penalty rates are more of an excuse for a poorly monitored interest costs. The increasing rate shock from penalty rate increase than motivated by risk-based pricing. Some repricing along with its large potential impact on a family’s rule changes on this issue by the Federal Reserve are budget also raises the question of whether penalty repric- scheduled to occur in July 2010, but given the research ing actually causes credit risk to rise. An $1,800 budget results further policy changes may be warranted. Full shock to a family that may already be in financially references are available upon request.

For further information contact: Joshua M. Frank Center for Responsible Lending 302 West Main St. Durham, NC 27701 Phone: 919.313.8525 Fax: 919.313.8592 E-Mail: [email protected]

218 Marketing and Public Policy Conference Proceedings / 2009 SPECIAL SESSION

BLIND MEN AND AN ELEPHANT: TRIANGULATION OF RESEARCH METHODS TO DOCUMENT THE DISCRIMINATION OF MINORITY CONSUMERS AND MINORITY ENTREPRENEURS

Session Co-Chairs: Jerome D. Williams, University of Texas at Austin Sterling A. Bone, Brigham Young University, Provo

SESSION OVERVIEW The purpose of this special session is to describe various methodological approaches employed by academics and Despite great strides, racial discrimination remains a law-enforcement agents (e.g., the courts) to document and serious problem in the United States. Efforts to eradicate enforce equality and fair treatment of racial minorities in discrimination have concentrated on education, housing, business-to-consumer (B2C) and business-to-business and employment. Discrimination is however equally per- (B2B) marketplace exchanges. The presenters of this vasive in many other marketplace situations, including in session represent marketplace discrimination researchers the consumption experiences of hotels, restaurants, gas from traditional academia and industry. They will de- stations, department stores, and in banks. Consumer racial scribe novel research methods including, experimental, profiling (CRP) is one form of marketplace discrimina- qualitative depth-interview, and field-tests using mystery tion. It has been defined as the differential treatment of shoppers. consumers in the marketplace based on race/ethnicity Due in part to the serious social and legal ramifications of that constitutes denial or degradation in the products or marketplace discrimination, recent research in this do- services that are offered to the consumer and has been main has had to adopt various methodological approaches recognized as actual and perceived consumer discrimina- to accurately expose these stereotypical attitude, perception in marketplace settings (Harris, Henderson, and Wil- tions, and behaviors. Furthermore, to date, the research liams 2005). focus has been on exposing the marketplace discrimination that occurs in business-to-consumer (B2C) settings. CRP and other forms of marketplace discrimination impact Bone et al. (2008) proposes that future marketplace dis- members of minority groups beyond those classified as crimination research be directed toward disentangling the Black/African American, such as Hispanics, Asians, Native expressions and external forces of vulnerability among and Arab Americans. In fact, since September 11, 2001, small business owners as they transact as consumers in the there has been heightened interest and concern about CRP marketplace. Lost in the wide spectrum of much larger as it applies to anyone perceived as Middle Eastern, “small business,” entrepreneurs have attenuated business including South Asians, Latinos, and even Jews (Nakao acumen and limited resources, and therefore behave more 2001). From a vulnerable population perspective (Baker like “consumers” than they do “commercial” entities. et al. 2005; Hill 2006), the feelings and expressions of minority consumers that they feel out of place, unwelcome, We hope this session stimulates new thought and research and against all odds when attempting to transact business directions to help academics, practitioners, and policy- in the marketplace may be due to the interaction of makers eradicate race discrimination in both B2C and individual (e.g., a person’s race) and situational B2B marketplace exchanges. References are available characteristics (e.g., structural or societal barriers). upon request.

Marketing and Public Policy Conference Proceedings / 2009 219 Special Session: Blind Men and an Elephant: Triangulation of Research Methods to Document the Discrimination of Minority Consumers and Minority Entrepreneurs

For further information contact: Sterling A. Bone Marriott School of Management Brigham Young University 660 TNRB Provo, UT 84602 Phone: 801.422.3113 Fax: 801.422.0108 E-Mail: [email protected]

220 Marketing and Public Policy Conference Proceedings / 2009 Special Session: Blind Men and an Elephant: Triangulation of Research Methods to Document the Discrimination of Minority Consumers and Minority Entrepreneurs

AN EXPERIMENTAL SCENARIO-BASED APPROACH TO ANALYZING CONSUMER RESPONSE TO MARKETPLACE DISCRIMINATION

Jerome D. Williams, University of Texas at Austin Geraldine R. Henderson, University of Texas at Austin Anne-Marie G. Hakstian, Salem State College Sophia R. Evett, Salem State College

SUMMARY The purpose of the present study was twofold: first, to determine what characteristics of the discriminatory situ- Euphoria related to the recent significant gains of histori- ation influence people’s perceptions of the severity of cally subordinate groups in the United States has led many cases of marketplace discrimination and second, to ascer- to discount the inequalities that still exist in many domains tain what actions they deem are appropriate following the of society. To pinpoint this contrast, on the day after discrimination incident. Approximately 1100 participants President Barack Obama’s historic victory, Reuters pub- were randomly presented with one of 32 conditions of a lished data from U.S. government agencies highlighting marketplace discrimination scenario that varied in the racial inequality covering the four broad categories of nature of the discrimination and the ethnicity of the health, the economy, criminal justice, and education. Our victim. Participants rated the severity of the treatment, research suggests that continuing economic inequality in their emotional responses to the scenario, and the re- the form of consumer discrimination may be a continuing sponses they thought the victim should have. Results are and significant problem in the American marketplace. discussed in terms of the theories of aversive racism and The litigation that has taken place during the past decade liberation psychology. reveals that the marketplace is not immune from the discrimination that exists in other sectors of society. To address the research questions, we conducted a sce- Therefore, marketers and public policymakers must ad- nario-based experiment embedded in a survey using web- dress the perception and/or the reality that they engage in based data collection. This method has been used in other discriminatory conduct based on their customers’ race studies where respondents are presented with scenarios and ethnicity. More research is needed to understand the for which they answer survey questions, with the factor of causes, the consequences, and potential solutions for interest manipulated in the vignettes (e.g., Wirtz and eliminating this form of unfair treatment in the market- Bateson 1999). We manipulated the type of marketplace place. discrimination based on actual court cases of consumer racial profiling and marketplace discrimination, using the Marketplace discrimination is a form of discrimination dimensions of discrimination developed by Harris, against customers engaged in shopping or purchasing Henderson, and Williams (2005): the type of discrimina- goods or services. According to Harris, Henderson, and tion (overt or subtle), the level of service (degradation or Williams (2005), the discriminatory treatment may in- denial), suspicion of criminal treatment (present or not). volve overt or subtle prejudice and may result in outright denial or degradation in goods or services. Furthermore, Our preliminary results indicate that overall participants customers may or may not be treated with criminal suspi- felt that the customer was very badly treated across the cion. While such discrimination experiences are preva- experimental conditions. Compared to the Low Discrimi- lent according to surveys of African- and Hispanic-Ameri- nation Scenario, participants who read the High Discrimi- can shoppers, lawsuits against retailers and other com- natory Scenario (1) rated the treatment to be more un- mercial establishments are largely unsuccessful and shop- pleasant, (2) disagreed that the customer should keep pers may feel limited in the action they can take. shopping or should dismiss the treatment because the

Marketing and Public Policy Conference Proceedings / 2009 221 Special Session: Blind Men and an Elephant: Triangulation of Research Methods to Document the Discrimination of Minority Consumers and Minority Entrepreneurs clerk was having a bad day, and (3) agreed that the can. These results suggest that participants must: (1) customer should take actions such as contacting the Better already perceive that consumer discrimination is a perva- Business Bureau, talking to an attorney, or organizing a sive problem in society and (2) be exposed to the “worst boycott. Also, participants who perceived that higher case scenario” in order to respond that the customer in the levels of race-based marketplace discrimination exist in scenario should do something as drastic as contacting an the U.S. supported stronger reactions (contacting the attorney or organizing a boycott. We will discuss the attorney, organizing a boycott) in response to the more implications of our results for marketers and public severe scenario and were less likely to dismiss the dis- policymakers. References are available upon request. crimination when the customer was White or Asian Ameri-

For further information contact: Jerome D. Williams Department of Advertising/Center for African and African American Studies University of Texas at Austin 1 University Station A1200 Austin, TX 78712–1092 Phone: 512.471.7302 Fax: 512.471.7018 E-Mail: [email protected]

222 Marketing and Public Policy Conference Proceedings / 2009 Special Session: Blind Men and an Elephant: Triangulation of Research Methods to Document the Discrimination of Minority Consumers and Minority Entrepreneurs

THE PLIGHT AND PILGRIMAGE OF MINORITY SMALL BUSINESS ENTREPRENEURS: EXPLORING THE EXPERIENCE OF VULNERABLE POPULATIONS WITH ZMET

Sterling A. Bone, Brigham Young University, Provo Glenn L. Christensen, Brigham Young University, Provo Jerome D. Williams, University of Texas at Austin Elise Riker, Arizona State University, Tempe

SUMMARY do the conditions and expressions of vulnerability among White business owners differ from the conditions and In our previous work (Bone et al. 2008), we conclude that expressions of African American and Hispanic business small entrepreneurial business banking customers are owners in capital-seeking marketplace exchanges? Sec- inherently vulnerable to prejudicial discrimination be- ond, what are the factors that increase the likelihood of cause regulatory enforcement and policy oversight of vulnerability among African American and Hispanic en- bank lending practices rest on the erroneous assumption trepreneurs? Third, what types of appraisal, defense, and that retail bank consumers and business banking custom- coping mechanisms do minority entrepreneurs employ to ers are completely discrete and disparate consumers of counter vulnerability? bank products. This undergirding, yet erroneous assumption limits efficacy of public policy safety nets and may Thirty-four informants who owned their business were restrict access to capital financing for entrepreneurs. In recruited for this study. Eighteen of these informants were our earlier work we proposed entrepreneurs of racial and/ White, eight were Hispanic, and eight were African Ameri- or ethnic minority status might experience exacerbated can. We employed the Zaltman Metaphor Elicitation effects of restricted market access to commercial funding. Technique (ZMET; see Zaltman 1997), a semi-structured, The present research employs in-depth interviews with in-depth interview format that focuses on uncovering the minority business owners to understand the racial and informants’ emotions, beliefs, and values. This approach ethnic profiling of small business owners. is recommended for exploring the deeply held and shared cultural models (Keller 1993). Informants were asked to Minority entrepreneurship is pressing issue for federal, prepare for their 90-minute interview by writing down at state, and economic stakeholders. Recent U.S. Senate least five thoughts and feelings about seeking financing subcommittee hearings on minority entrepreneurship re- for their businesses. Informants were also asked to find a ported census and economic data which suggests that picture that represented each of their thoughts or feelings. minority entrepreneurs are most vulnerable to market- The combination of the approaches elicited rich and place discrimination (Snowe 2007; Wainwright 2007). descriptive insights into the informants’ perceived vul- Where is discrimination among small business owners nerability. Interviewers were matched to the informant most likely to manifest its ugly head? A report from the based on their race and language of origin. This step was U.S. Chamber of Commerce (2005) concludes that lack of taken to enhance conversational flow and to ensure cul- access to capital and credit is among the most cited tural sensitivity. The data were analyzed using the grounded obstacles for African American and Hispanic business theory framework (Strauss and Corbin 1990). owners. Furthermore, Wainwright (2007) reported finding quantitative evidence that minority-owned businesses The coding process produced more than 200 unique are substantially and significantly more likely to be denied meaning constructs, of which 50 were unique to minority credit than are white-owned firms. business owners. A careful comparative analysis within and across the race/ethnic groups revealed insights into We employed a qualitative depth-interview method to the internal and external forces of vulnerability and the investigate the following research questions. First, how stark differences in appraisal, coping and defense

Marketing and Public Policy Conference Proceedings / 2009 223 Special Session: Blind Men and an Elephant: Triangulation of Research Methods to Document the Discrimination of Minority Consumers and Minority Entrepreneurs mechanisms that non-minority and minority entrepreneurs that policy-makers may take to promote and strengthen employ to respond to situational and race-related minority business ownership. References are available vulnerability. The results are presented as action-steps upon request.

For further information contact: Sterling A. Bone Marriott School of Management Brigham Young University 660 TNRB Provo, UT 84602 Phone: 801.422.3113 Fax: 801.422.0108 E-Mail: [email protected]

224 Marketing and Public Policy Conference Proceedings / 2009 Special Session: Blind Men and an Elephant: Triangulation of Research Methods to Document the Discrimination of Minority Consumers and Minority Entrepreneurs

MARKET RESEARCH FOR DETECTING AND ELIMINATING RACIAL PROFILING AND DISCRIMINATION IN LENDING

Paul C. Lubin, Informa Research Services, New York

SUMMARY tion, a direct comparison of the experience encountered by minority (e.g., Black or Hispanic) and non-minority The paper describes the application of market research for (White) classes of testers is made. Post application tele- detecting and alleviating consumer vulnerabilities in lend- phone surveys measure the assistance and treatment en- ing. Discrimination and unfair sales practices, limit the countered by loan applicants and whether it is of a nature consumers’ ability to make sound credit decisions. These that maximizes the opportunity for loan approval at terms vulnerabilities impact the consumers that need the infor- based on the financial circumstances of the borrower. mation the most, minorities and those less knowledgeable Similar to mystery shopping, direct comparisons of the about financial products. attitudes and experiences recalled by recent approved and denied (some surveys include withdrawn and approved What Are the Benefits of Using Market Research? not accepted applicants) minority and non-minority loan applicants are made. Market research is a powerful tool for detecting and resolving problems consumers encounter when applying for credit. These problems and vulnerabilities can mani- Both approaches measure disparate treatment. In con- fest themselves in racial profiling and disparate treatment sumer credit, this may manifest itself in discouraging and/ during the loan process. For the consumer seeking credit or providing different information to a potential Black he or she may not be able to obtain accurate information applicant than to a similarly situated White applicant or to make appropriate credit decisions. Racial profiling and rejecting a Hispanic or Latino and accepting the similarly discrimination magnify this problem and impact the very situated White applicant or even providing different and consumers that need the information the most. For the unfavorable terms to the Black female applicant than the lender it can result in business practices that violate the similarly situated White female applicant. law, reduce revenue, and increase risk. For the nation these problems can result in inefficient and unsound credit Differential or disparate treatment may be classified as markets. These problems ultimately affect wealth at the overt or subtle in nature. Overt differences are those household and national level. actions which are evident to the prospective borrower. Examples include a prospective borrower told to “go to Time-tested research methods, for example mystery shop- another lender,” or told that he or she “will not qualify” for ping and match pair testing, monadic testing and post the loan. Subtle differences are those actions which are application consumer telephone surveys, can help guide not noticeable to the borrower but nevertheless affect the government policy to ensure the allocation of credit is prospective borrower’s access to credit and ability to based on consumes receiving understandable and sound make an appropriate decision about credit. Examples information about credit products. include telling the Black potential borrower approval time will take 60 days while the White potential borrower is Match pair testing and mystery shopping call for the use told 30 days, taking longer to approve the Black loan of testers or mystery shoppers posing as potential or actual applicant than the potential White applicant or making buyers. Unlike statistical procedures which require out- available to the potential Hispanic borrower only loans comes (loan approval, loan denial, pricing) and rely on offering lower monthly payments but higher variable abstract arguments and statistical principals, testing pro- interest while offering the potential White borrower a vides a record of the treatment or experience encountered variety of loans including Fixed Rate and Adjustable by the mystery shopper or tester. In tests for discrimina- loans.

Marketing and Public Policy Conference Proceedings / 2009 225 Special Session: Blind Men and an Elephant: Triangulation of Research Methods to Document the Discrimination of Minority Consumers and Minority Entrepreneurs

Changes in the Credit Marketplace (1991–2008) ers to a variety of complex product scenarios. Ancillary products, for example credit insurance were often in- Underwriting terms were less flexible in the early to cluded in the monthly payment. At the same time lenders mid -1990’s than they were in the late 1990’s and 2000– using risk based pricing pulled credit scores and credit 2007. In 2008 with the restricted availability of funds information in real time. Potential borrowers now re- underwriting terms have become less flexible and in some ceived almost instant approval limiting the ability to sense returned to the practices of the 1990’s. There were evaluate and compare products. Then in late 2007 and into fewer products offered in the early – mid 1990’s and fewer 2008 owing to collapse of credit markets, the consumer channels where the consumer could apply for a loan. The was faced with more rigid loan terms and fewer lenders approval process was not instantaneous and allowed con- and product alternatives. At times, lender capacity issues sumers more time to search for credit. Complementary limited the consumers’ ability to obtain information. products, e.g., credit protection, were far fewer and sales tactics were more constrained and reflected the more rigid underwriting criteria and more experienced sales person- Summary of Market Research and Self-Testing Results nel of the time. While loan denial rates and denial disparities between Starting in the very late 1990’s and accelerating into the minorities and non-minorities declined in 2000–2007, new century, commensurate with the growth of secondary differences in treatment between protected (minorities) market funding and risk-based pricing, consumers were and non-protected classes (non-minorities) of consumers faced with an increasing number of lenders, sales person- increased. Minority and non-minority potential loan cus- nel, delivery channels and complex product alternatives tomers more often received different information and to choose from. Underwriting polices became less rigid assistance more often than in the mid to late 1990’s. reflecting the lenders’ ability to source increasingly lower Consumers also faced more aggressive sales tactics in the cost funds. Sub-prime lending to consumers that carried form of frequent refinancing and consolidation of secured greater risk of default increased dramatically. Sub-prime and non-secured debt, diminished emphasis on ability to lenders extended credit at higher rates, longer loan terms pay and emphasizing reduced monthly payments without and lower payments to those who would have not nor- regard to loan principal and interest payments over the life mally obtained credit. During this time government regu- of the loan. During late 2007 and through 2008 overt latory agencies turned their attention to privacy, money differences in treatment between minorities and non- laundering and the Bank Secrecy Act and were less minorities while few in number increased. Subtle differ- concerned with lending sales practices and adherence to ences in terms of information provided and explanation of fair lending laws. products and rates and terms provided to minorities and Sales personnel, often with little experience and using non-minorities also increased. technology and “what-if” simulations exposed consum-

For further information contact: Paul C. Lubin Informa Research Services 80 Norman Road New Rochelle, NY 10804 Phone: 914.548.1124 Fax: 818.880.8788

226 Marketing and Public Policy Conference Proceedings / 2009 A METHOD TO TRACK VIEWING BEHAVIOR: DEMONSTRATING THE EFFECT OF A CONCURRENT DISCLOSURE IN DTC ADVERTISING

Jack Swasy, American University, Washington, D.C. Anu Mitra, American University, Washington, D.C.

SUMMARY sessing mandated messages; they conclude that non significant differences across the conditions suggest respon- In this paper, we present a method that may enhance our dents either did not read the disclosure message or did not ability to understand the effects of mandated disclosures. comprehend it. However, because their study included no We demonstrate this in the context of a field study using indicators of processing, it may be problematic to draw a concurrent disclosure to correct a misleading inference conclusions about the potential effects of such disclo- created by a market leadership claim (MLC) presented in sures. Mixed and null findings create an especially diffi- a direct-to-consumer (DTC) print ad for a prescription cult choice for regulators. Some have suggested that medication. courts may place an inappropriate emphasis on factors believed to affect attention to a disclosure (size or place- Introduction ment) rather than implementing a potentially more effective and flexible performance standard for disclosures FDA requires that an informational type of disclosure (Mason and Scammon 2000). However, without proce- (Foxman, Muehling, and Moore 1988) accompany a dures to distinguish between cognitive efforts such as market leadership claim (MLC) such as most prescribed. attention and processing (see Stewart and Martin 1994), it This is typically done as a footnote which identifies the may be unrealistic to implement or even propose perfor- independent research study and a timeframe for the sales mance standards. In this research, we use a behavioral leadership claim. This requirement raises the possibility indicator of processing to help clarify null findings in a of an interesting, but perhaps untended, consequence. test of the effects of a disclaimer that accompanied a sales This information footnote may enhance the credibility of leadership claim. a leadership claim. Thus, those subjects who attend to and encode the informational footnote may be more likely to Method form comparative inferences about a non comparative attribute (such as proven effective) featured after the Adult participants (N = 130) 80 percent white, all female, (comparative) leadership claim (copy by copy (see Baron and equal split between age groups (18–34 and 35–54) and Miniard 1999). This sets a context for examining the were recruited in a shopping mall in Virginia and ran- ability of a simple concurrent disclosure (a restrictive domly assigned to one of two versions (nMLC = 64; nNoSE = footnote) to correct the tendency of those subjects who are 66) of a direct-to-consumer (DTC) print ad for Retin-A, a most at risk of forming invalid comparative efficacy prescription acne medication. Respondents were given beliefs. We examine a restrictive disclosure that directly instructions that they would be viewing an electronic addresses (only) the inferred comparative belief (com- magazine and that parts of each page would be covered by parative effectiveness) – There is no scientific evidence grey boxes, and the content beneath could be revealed by that (brand) is more effective than other prescription moving a cursor over the box. The magazine contained a treatments. Such a disclosure (NoSE) may be indicative of cover, editorial content, and three ads including the target the type of restrictive disclosures (simple; factual; placed ad. The Retin-A ad was the last in the sequence of ads. in close proximity to the informative disclosure; and Each of the magazine page and filler ads had several grey narrow in scope) a firm may wish to use or the FDA might boxes. As subjects browsed, their viewing behavior was consider as a remedy. recorded (e.g., dwell time on each box). After browsing, subjects completed a series of structured questions (e.g., We focus on consumer processing of disclosures in print effectiveness and comparative effectiveness). The pro- advertisements in situations where viewers are not forced motional claim of sales leadership was presented as fol- to fully attend to and process the advertisement. Jacoby lows: “. . . Based on independent research, doctors pre- and Szybillo (1994) commented on the difficulty of as- scribe Retin-A more than any other brand for the treat-

Marketing and Public Policy Conference Proceedings / 2009 227 rd ment of acne – about 75 percent of all prescriptions were least deliberation (lowest 1/3 , index = .23), a modest level rd for Retin-A. The MLC claim was supported with an (the middle 1/3 , index = .84) and the most deliberation rd informational footnote referencing data from the IMS (top 1/3 , index = 3.0). As expected, blocking subjects by Health National Prescription Audit for 2007. Two ver- deliberation revealed a statistically significant ad by block- sions of the ad – qualified or unqualified -- varied on ing interaction effect on effectiveness (F (2, 87) = 3.77 whether (or not) the comparative sales leadership claim (p < .05)) with perceived effectiveness increasing with was accompanied by a qualifying restrictive disclosure more deliberation in the MLC ad (3.71, 4.07, and 4.33), (There is no scientific evidence that Retin-A is more but decreasing with increasing deliberation in the NoSE effective than other prescription treatments). The word- ad (4.11, 3.79, and 3.73). Similar results were observed on ing in the covered bottom disclosure was the only aspect comparative effectiveness (F (2, 87) = 3.44 (p < .05). that varied between the two ad conditions. Other portions Comparative effectiveness increased with more delibera- of the advertisement (major statement, side effects, prod- tion in the MLC ad (3.49, 3.64, and 3.81) but decreased uct graphics), whether covered or uncovered, were held with increasing deliberation in the disclosure ad (4.06, constant. 3.56, and 3.62).

Results Implications for Public Policy

As expected, no difference between the NoSE and MLC We present an approach to track self-paced viewing conditions was observed (effectiveness: 3.78 v 3.93 re- behavior in a field setting and demonstrate its value in spectively, F (1,126) < 1; comparative effectiveness: detecting disclosure effects where it may be problematic 3.67 v 3.52, F (1,126) = 1.14, ns)). Based on an inspection to assume claim and disclosure processing. Results sug- of viewing behavior patterns, we identified twenty four gest that processing of an FDA-mandated informational individuals who exhibited questionable viewing behav- disclosure that accompanies a market leadership claim iors, two subjects who had extreme reading times on may actually enhance the likelihood of unwarranted infer- instruction pages, and eleven individuals who did not ences about comparative effectiveness. We offer prelimi- open the masked ML claim. These subjects were omitted nary evidence that a qualifying disclosure, if processed, from the analysis, yielding ninety five with verified expo- could be useful in partially correcting unwarranted beliefs sure to the ML claim (43 in NoSE and 52 in MLC). caused by leadership claims. We suggest that this method of revealed exposure may have application for studying a Reading speed was estimated over six instruction pages variety of placement, visual and text issues in advertise- (mean of 186 milliseconds per word; alpha = .74). An ments. index of dwell time was calculated using viewing time on the bottom disclosure divided by an expected time (using KEYWORDS: Qualifying disclosures, market leadership the participant’s reading speed and the number of words claims, comparative claims, prescription drug advertis- in the disclosure of that condition). This index was used to ing, DTC. References are available upon request. form three groups in each condition – those exhibiting the

For further information contact: Jack Swasy American University Washington, D.C. 20016–8044 Phone: 202.885.1974 Fax: 202.885.2691 E-Mail: [email protected]

228 Marketing and Public Policy Conference Proceedings / 2009 DOES PRESENTATION MAKE A DIFFERENCE TO RISK PERCEPTION: TESTING DIFFERENT FORMATS FOR COMMUNICATION OF CANCER RISKS

Sandra C. Jones, University of Wollongong, Australia

ABSTRACT Hoeffler and Ariely 1999; Simonson et al. 2001). Tradi- tional theories of decision-making, such as the theory of Evidence suggests that the presentation format of risk rational choice (Savage 1954; vonNeuman and information can affect people’s perceptions of risk and Morgenstein 1947), assume that people make choices that influence health-related decisions. In these studies we maximize their well-being or utility in a given situation, investigated the impact of four different risk presentation within available resources and constraints. That is, people formats: standard presentation, risk ladder, different base are assumed to be rational and to make rational choices rates and visual representations on women’s perceptions using common knowledge. The foundation of theories of of developing breast cancer or lymphoma. We found that rational choice is invariance (Slovic 1995); that is, ratio- the different presentations had virtually no impact on the nal choices must comply with the assumptions of: (a) participant’s risk estimates. Only in the second study description invariance – preferences should not depend relating to risk perceptions for lymphoma was there a on the description of options; and (b) procedure invari- significant difference between conditions for estimated ance – preferences should not depend on the method of 10-year risk, with those in the ladder present condition elicitation. reporting a lower estimated risk. The implications of these findings for future research on consumer risk perceptions However, empirical investigations of decision-making are discussed. have shown many violations of both description invariance and procedure invariance. The application of this INTRODUCTION research to health-behavior has been limited to a few studies, although there is a large body of research on It has been consistently shown that women tend to over- message framing, i.e., presenting logically equivalent in estimate their own risk of developing breast cancer (e.g., semantically different ways (Krishnamurthy, Carter, and Black, Nease, and Tosteson 1995; Smith et al. 1996), and Blair 2001). Studies of framing effects on health-related the population incidence of breast cancer (e.g., Barratt intentions and behavior have been conducted in numerous et al. 1997). It has been argued that this overestimation is areas, including breast self-examination (Lalor and Hailey due in large part to the sensationalist and misleading 1990; Meyerowitz and Chaiken 1987) and mammogra- coverage of breast cancer in the mainstream media phy (Banks et al. 1995). Within this paradigm, researchers (NHMRC National Breast Cancer Centre 1999; Jones have also considered the effects on risk perceptions of 2004; Blanchard et al. 2002). There is considerable evi- interactions between framing and other variables, such as dence that risk estimations, and processing of risk infor- source credibility (Grewal, Gotlieb, and Marmorstein mation, are influenced by the extent to which specific 1994). risks are “available,” in people’s minds; for example, that people are exposed to via media coverage or personal While there are a limited number of studies focusing experience (e.g., Kuran and Sunstein 1999). specifically on message presentations effects in health behavior communications, the evidence from the existing Presenting Risk Information studies suggests that the presentation format of risk information can affect people’s perceptions of risk and influ- Attitude theory assumes that the mode of presentation of ence health-related decisions, including decisions about attributes should not matter, as long as the same informa- whether to accept medical treatment (e.g., Kaplan et al. tion is presented (e.g., McGuire 1985; Petty and Cacioppo 1985; Feldman-Stewart et al. 2000). However, studies of 1986). Behavioral Decision Theory (BDT), on the other particular formats (such as “risk ladders” or “1 in x” hand, posits that task characteristics (such as the mode of scales) are limited in number and have produced inconsis- presentation format and response format) can have content findings, with evidence for the superiority of a pre- siderable impact on the evaluation of outcomes, and thus sentation format in one study being absent or reversed in on intentions and behavior (e.g., Cox and Grether 1994; another.

Marketing and Public Policy Conference Proceedings / 2009 229 The purpose of the present study was to investigate the Risk Ladder. The use of a risk ladder (i.e., presenting the effect of four different format manipulations on women’s risk of a disease or condition in the context of other, perceptions of their risk of developing breast cancer. First, known, diseases) provides people with a context in which a brief overview of the literature on women’s perceptions to evaluate the risk, and thus should lead to more accurate of breast cancer risk is provided, followed by a review of risk perceptions. Loomis and duVair (1993) found that a previous studies of the four risk format manipulations that risk ladder was more effective in conveying information were the focus of the current research. An experimental about a risk in comparison to other risks. The use of risk study was conducted to investigate these manipulations – ladder presentations of risk information is common in lay this study is described and the findings presented and publications. For example, USA Today has used risk discussed. Due to the unexpected nature of these findings, ladders to communicate the likelihood of being murdered a second experimental study was conducted using a dif- with a gun, compared to the likelihood of dying from a ferent cancer type and this second study is described and range of other non-natural causes (Memmott 2002). It is the findings presented and discussed. Finally, the implica- reasonable to assume that presenting a risk in the context tions of the two studies are examined. of other known risks (with which people are likely to have first- or second-hand experience) should increase the Risk Communication Formats accuracy of risk perceptions (particularly relative to other risks). Thus: It has been suggested that some of the differences in proportions of women over-estimating their risk of breast H2: The inclusion of a risk ladder will lead to more cancer – e.g., approximately one-third of U.K. women accurate estimates of breast cancer risk compared to two-thirds of U.S. women – may be due to differential exposure to risk information (Hopwood 2000), H3: The inclusion of a risk ladder will lead to more both in terms of the amount of exposure and the accuracy accurate estimates of breast cancer risk in relation to and nature of such information. There are few studies that other risks which are included on the risk ladder have actually compared the effectiveness of different risk information presentation formats, and most have only Different Base Rates. A commonly manipulated variable varied one aspect of the information. Further, as discussed in the presentation of risk information is the size of the above, the results of these studies are frequently conflict- denominator (base rate). That is, a risk can be presented in ing or inconsistent (Vernon 1999). The following section terms of the incidence per 10, 100, 1,000 or any other reviews the four information presentation features to be number of people, and the numerators will be progres- investigated in the current studies. sively larger (for example, “one in ten” can be presented as “100 in 1,000"). Several studies have demonstrated that Natural Frequencies (1 in X). Gigerenzer and Hoffrage people tend to display “base-rate neglect,” which is to (1995) provided two groups of participants with hypo- focus on the numerator and ignore the denominator. One thetical information about breast cancer incidence and study (Yamagishi 1997) found that 1,286/10,000 was mammography diagnostics and found that almost three assessed as a greater risk than 24.14/100. That is, people times as many of the participants given the information in tended to focus on the total number of affected people natural frequency format were able to provide the correct (1,286 compared with 24.14) rather than the proportion answer (46% correct compared to 16%). Similar effects (12.86% compared with 24.14%). The existence of base- have been demonstrated for communicating information rate neglect means that a larger base will result in higher about the accuracy of mammograms to laypeople (Hoffrage estimates of breast cancer risk. Given that, in general, 2003). Based on this and several other studies, Hanoch women overestimate their risk of breast cancer, it is and Pachur (2004) have advocated that nurses use natural reasonable to assume that the use of a lower base will frequencies to communicate statistical information to result in more accurate (i.e., lower) risk perceptions. Thus: patients. The “standard” breast cancer message in Austra- lia is “1 in 8 Australian women will develop breast H4: The use of a smaller base (100) will result in more cancer.” Based on the research on the use of natural accurate (i.e., lower) estimates of breast cancer risk frequencies, we propose: than the use of a larger base (1,000).

H1: The use of the “standard presentation” (1 in X) format Visual Representation. It is generally believed that visual will lead to more accurate (i.e., lower) estimates of information communicates more effectively than verbal breast cancer risk. information (“a picture is worth a thousand words”); and

230 Marketing and Public Policy Conference Proceedings / 2009 that the most effective communication combines the two was determined based on the results of the Lipkus et al. forms. This strategy is used in the media to convey (2000) study, and the estimation of a meaningful effect comparative statistical information; for example, USA size. Today used gender symbols (each representing 10 individuals) to demonstrate the small number of women who Stimuli. As shown in Table 1, there were nine conditions are executed relative to men (Ahrens 2002). Graphic in the study (that is, women were randomly allocated to presentation of information has been found to assist in reading one of nine versions of the information sheet). people’s processing of numerical information (Cleveland and McGill 1984). However, risk information is generally All versions included the same basic information on presented in numerical form (i.e., percentage, proportion, breast cancer incidence and risk factors, and a photograph risk ratio etc.) or verbal-numerical (i.e., using verbal of Sara Henderson (an Australian celebrity who is the descriptors of probability such as “often” or “rarely”). spokesperson for the national breast cancer screening Kaplan et al. (1985) found that using graphical displays to campaign). present extremely small probabilities of side effects increased participants’ willingness to take a hypothetical Estimated Risk. Participants were asked to estimate their vaccine; suggesting the tendency to overemphasize small risk of developing breast cancer (a) in the next 10 years probabilities may be reduced by a visual illustration of and (b) in their lifetime on a 0–100 scale, where 0 = “No how small they really are. However, Schapira and col- chance of me getting breast cancer” and 100 = “I will leagues (2001) found that the use of human-like visual definitely get breast cancer.” They were then asked to representations increases perceived risk, perhaps due to estimate their risk (c) compared to their risk of developing the increased salience and “personalization” of the infor- heart disease and (d) compared to other women their age mation. It appears logical, given the processing differ- on a five-point scale, where 1 = a lot higher, 3 = about the ences discussed above, that the addition of visual informa- same, and 5 = a lot lower. tion would result in higher perceptions of risk.1 Thus: Exposure to Breast Cancer. Participants were asked H5: The addition of visual representation of risk to verbal whether any of their family members or friends had ever information will lead to less accurate (i.e., higher) had breast cancer. Response options were “None that I estimates of breast cancer risk. know of,” “1 (family member/friend),” “2–3 (family members/friends),” and “more than 3 (family members/ STUDY 1 friends).”

The purpose of this study was to investigate the effect of Health Literacy and Numeracy. Health literacy was manipulating the four information presentation variables measured with a subset of items from the S-TOFHLA discussed above on women’s estimates of their own (Baker et al. 1999); the S-TOFHLA and the longer absolute and comparative risk of developing breast can- TOFHLA are the recommended instruments for measur- cer. ing functional health literacy (Ad Hoc Committee on Health Literacy for the Council on Scientific Affairs Methodology 1999). Health numeracy was measured using the scale developed by Lipkus, Samsa, and Rimer (2001). Participants. The study participants were 182 women aged 18–81 years (mean = 43.5 years). Quotas were Results established to ensure even representation of participants across three pre-determined age groupings: < 35 years, Previous Exposure to Breast Cancer. Sixty-nine percent 35–49 years, and 50 years and over – with the final (125) of the respondents reported that they had no family proportions in the three groups being 34.6 percent (63), 33 members with breast cancer, 23 percent (41) had one percent (60), and 32.4 percent (59) respectively. Ninety- family member with breast cancer, and 8 percent (15) two eight percent of the participants had completed some high or more family members. Forty-four percent (79) of the school, with 86.7 percent having completed at least 10 respondents reported that they had no friends with breast years of education (i.e., three years of secondary), 60.8 cancer, 26 percent (47) had one friend with breast cancer, percent at least 12 years (high school graduates), and 51.9 and 30 percent (54) two or more friends. percent a post-secondary qualification. All of the participants were fluent in English, with 98 percent speaking Health Literacy and Numeracy. The mean health literacy English at home. Current or previous breast cancer was an score was 5.2 (SD = 1.8), within a possible range of zero exclusion criterion for this study. Required sample size to seven. For example, 86 percent correctly interpreted a

Marketing and Public Policy Conference Proceedings / 2009 231 TABLE 1 The Nine Information Conditions

Standard Risk Format Base Rate 100 Base Rate 1,000

Risk ladder present 123 Visual present 456 Ladder & visual absent 789

prescription label that stated “finish all this medication” to The mean response to the question “Compared to your mean that they should take all of the tablets. The mean chance of developing heart disease, do you think your numeracy score was 5.1 (SD = 1.8), within a possible chance of developing breast cancer is . . .” was 2.9 (sd = range of zero to seven. Across each of the seven items, 1.2) on a five-point scale (where 1 = a lot higher, 3 = about there were similar proportions of correct responses to the the same, and 5 = a lot lower). That is, on average, the Lipkus, Samsa, and Rimer (2001) study. women rated their risk of developing breast cancer to be approximately equivalent to their risk of developing heart Estimated Risk (All Participants). The mean response to disease – when, in fact, their average risk of developing the question “How likely is it that you will get breast heart disease is four times higher than their risk of devel- cancer in the next 10 years?” was 30.4 percent (sd = 22.8). oping breast cancer. Whilst this value is considerably higher than the objective mean risk for this population – less than 2 percent for the The mean response to the question “Compared to other group as a whole, given the mean age of the participants2 women your age, do you think your chance of developing (Merrill et al. 1999) – it is remarkably consistent with breast cancer is . . .” was 2.8 (sd = 0.8) on the same scale results from similar studies. For example, using the same as the previous question. That is, on average, the women 0 to 100 scale, Lipkus et al. (2000) found a mean estimate rated their risk of developing breast cancer as marginally of 30.2 percent (sd = 21.7) for perceived risk of breast higher than other women their age. cancer in the next 10 years. Effect of Presentation Manipulations The mean response to the question “How likely is it that you will get breast cancer in your lifetime?” was 35.1 Effect of Standard Presentation. There were no signifi- percent (sd = 22.5). As for the 10-year risk estimate, this cant differences in risk estimates between those partici- value is higher than the objective risk – interestingly, it is pants in the standard presentation and other presentation only fractionally higher than perceived 10-year risk – but conditions, for any of the four dependent variables. The is again consistent with previous findings. The mean mean 10-year risk estimate was 0.4 percent lower for estimate of the lifetime risk in the Lipkus et al. (2000) those in the standard presentation condition (CI = -7.5% study was 34.4 percent (sd = 22.3). to 6.6%); the mean lifetime risk estimate was 0.1 percent higher (CI = -6.8% to 7.1%). The mean difference in It is apparent from a comparison of the responses to the relative risk scores was less than 0.1 on a 5-point scale for two questions that a majority of women demonstrate both risk compared to heart disease (0.05, CI = -0.3 to 0.4) difficulty in distinguishing between 10-year and lifetime and risk compared to other women (-0.02, CI = -0.3 to risk, or at least in articulating this distinction. Only 33.7 0.2). Thus, H1 was not supported. percent (61) of the respondents correctly provided a lifetime risk estimate higher than their 10-year risk esti- Effect of Risk Ladder. There were no significant differ- mate; 57.5 percent (104) provided equal estimates for the ences in risk estimates between those participants in the two risks (which would only be accurate if they had risk ladder present and risk ladder absent conditions, for exactly 10 years left to live); and 8.8 percent (19) gave a any of the four dependent variables. The mean 10-year lifetime risk estimate lower than their 10-year risk esti- risk estimate was 1.0 percent higher for those in the risk mate. Again, these findings are remarkably consistent ladder condition (CI = -6.2% to 8.1%); the mean lifetime with Lipkus et al. (2000) who reported corresponding risk estimate was 2.0 percent higher (CI = -5.0% to 9.0%). figures of 32 percent, 61 percent, and 7 percent. The mean difference in relative risk compared to other

232 Marketing and Public Policy Conference Proceedings / 2009 women was – 0.1 on a 5-point scale (CI = -0.3 to 0.2). Having a family member with breast cancer was also the Thus, H2 was not supported. Further, the mean difference only significant predictor of perceived comparative risk – in relative risk compared to heart disease (which was that is, risk compared to women of the same age (t = 3.7, clearly shown on the ladder to be four times the risk of p = .000), and compared to risk of developing heart breast cancer) was -0.1 on a 5-point scale for risk (CI = disease (t = 2.4, p = .02). -0.5 to 0.3). Thus, H3 was not supported. Discussion Effect of Different Base Rates. There were no significant differences in risk estimates between those participants in Contrary to all of the hypotheses, risk information presen- the base 100 and base 1,000 conditions, for any of the four tation format was found to have no impact on perceived dependent variables. The mean 10-year risk estimate was risk. This is an important finding, particularly in the 1.0 percent higher for those in the base rate 100 condition context of inconclusive and contradictory results from compared to those in the base rate 1,000 (CI = -7.5% to previous studies. While the sample sizes were relatively 9.5%); the mean lifetime risk estimate was 4.2 percent small, there was sufficient power to detect meaningful higher (CI = -4.0% to 12.4%). The mean difference in differences, thus small sample size is not sufficient to relative risk scores was -0.01 on a 5-point scale for risk explain the lack of significant differences for any of the compared to heart disease (CI = -0.5 to 0.4) and 0.2 for risk presentation formats. The confidence intervals were con- compared to other women (CI = -0.1 to 0.5). Thus, H4 was sistently narrow enough to allow us to conclude that there not supported. were no meaningful differences in risk estimates between the conditions. Effect of Visual Representation. There were no significant differences in risk estimates between those partici- It was very clear from this study that the primary predictor pants in the visual representation present and visual rep- of perceived breast cancer risk is having had a family resentation absent conditions, for any of the four depen- member with breast cancer. Previous studies have found dent variables. The mean 10-year risk estimate was 2.7 that women’s perceptions of their own breast cancer risk percent lower for those in the visual representation condi- are strongly influenced by family history. For example, a tion (CI = -9.8% to 4.5%); the mean lifetime risk estimate study of perceived risk among employees in an oncology was 1.1 percent lower (CI = -8.1% to 5.9%). The mean Centre (Helzlsouer et al. 1994) found that perceived risk difference in relative risk scores was -0.1 on a 5-point of breast cancer was approximately double among women scale for risk compared to heart disease (CI = -0.5 to 0.3) with a first-degree relative with breast cancer for both 20- and -0.1 for risk compared to other women (CI = -0.4 to year risk (63% versus 29%, p < .005) and 40-year risk 0.2). Thus, H5 was not supported. (73% versus 40%, p < .005). Further, perceived risk was considerably higher among women who had both a friend Predictors of Risk Estimates and a relative (any relation) with breast cancer (40-year risk 51% versus 29%, p < .001). There were only two significant predictors of perceived 10-year risk of developing breast cancer: having a family One possible explanation for the unexpected absence of member with breast cancer resulted in a higher risk presentation effects is that, due to women’s exposure to a estimate (t = 3.8, p = .000); and having a low numeracy wide range of information on breast cancer risk, the score (t = 2.3, p = .02). In total, the model predicted 13 participants had already established estimates of their percent of the variance in risk estimates (p = .002). None own risk and thus were not influenced by the information of the other variables entered into the model (presentation presented to them. In order to exclude this explanation, a format, friend had breast cancer, age, language spoken at second study was conducted using information about the home, education level, and literacy score) were signifi- risks of developing a cancer to which participants would cant predictors of estimated 10-year risk. have had little, if any, exposure.

There was only one significant predictor of perceived STUDY 2 lifetime risk: having a family member with breast cancer again resulted in a higher risk estimate (t = 4.1, p = .000). Study 2 was designed to exactly replicate Study 1 in terms This final model predicted 11 percent of the variance in of methodology and risk presentation formats. However, risk estimates, and again none of the other variables were the context chosen was the risk of developing lymphoma significant predictors. (we did not specify whether Hodgkins or non-Hodgkins).

Marketing and Public Policy Conference Proceedings / 2009 233 This cancer was chosen following a pre-test with 20 Health Literacy and Numeracy. Health literacy was women in the target age group, which confirmed that they measured with the same subset of seven items from the S- had minimal knowledge of lymphoma, and no pre-exist- TOFHLA (Baker et al. 1999). Numeracy was measured ing personal risk estimates. using the same seven-item numeracy scale developed by Lipkus et al. (2001). The hypotheses were the same as those in study 1. Results Methodology Previous Exposure to Lymphoma. The majority of the Participants. The study participants were 99 women aged respondents (86%) reported that they had no family mem- 18–82 years (mean = 38.3 years). Quotas were again bers with lymphoma, 11 percent (11) had one family established to ensure even representation of participants member with lymphoma, and 3 percent (3) two or more across three pre-determined age groupings, although in family members. Similarly, 76 percent (75) of the respon- this study there was a smaller than expected proportion in dents reported that they had no friends with lymphoma, 16 the oldest group: < 35 years, 35–49 years, and 50 years and percent (16) had one friend with lymphoma, and 8 percent over – with the proportions in the three groups being 44.9 (8) two or more friends. percent (44), 35.7 percent (35), and 19.4 percent (19) respectively. Ninety-six percent of the participants had Health Literacy and Numeracy. The mean health literacy completed some high school, with 83.7 percent having score was 5.6 (SD = 1.8), within a possible range of zero completed at least 10 years of education (i.e., three years to seven. That is, the mean health literacy score for this of secondary), with 61.2 percent at least 12 years, and 41.8 group was marginally higher than for the Study 1 partici- percent a post-secondary qualification. All of the participants, and again consistent with previous studies. The pants were fluent in English, with 97 percent speaking mean numeracy score was 5.1 (SD = 1.9), within a English at home. Current or previous lymphoma was an possible range of zero to seven. That is, the mean numeracy exclusion criterion for this study. score for this group was the same as the mean score of the Study 1 participants. Stimuli. As for study 1, there were nine conditions in the study (that is, women were randomly allocated to reading Estimated Risk (All Participants). The mean response to one of nine versions of the information sheet). All ver- the question “How likely is it that you will get lymphoma sions included the same basic information on lymphoma in the next 10 years?” was 20.5 percent (SD = 21.2). While incidence and risk factors; what was varied was: the this value is considerably lower than the mean estimated numerical presentation of the risk (i.e., “one in 92,” “one risk for breast cancer, it is much higher than the objective out of 100” or “10 out of 1,000”); the inclusion or mean risk for this population (4%). exclusion of the risk ladder; and the inclusion or exclusion of visual representation. The mean response to the question “How likely is it that you will get lymphoma in your lifetime?” was 26.2 Measures percent (SD = 23.4). As for study 1 (breast cancer), the lifetime risk estimate is not a lot higher than perceived 10- Estimated risk. Participants were asked to estimate their year risk. risk of developing lymphoma (a) in the next 10 years and (b) in their lifetime on a 0–100 scale, where 0 = “No The mean response to the question “Compared to your chance of me getting lymphoma” and 100 = “I will chance of developing heart disease, do you think your definitely get lymphoma.” They were then asked to esti- chance of developing lymphoma is . . .” was 2.4 (SD = 1.1) mate their risk (c) compared to their risk of developing on a five-point scale (where 1 = a lot lower, 3 = about the heart disease and (d) compared to other women their age same, and 5 = a lot higher). That is, on average, the women on a five-point scale, where 1 = a lot higher, 3 = about the rated their risk of developing lymphoma to be slightly same, and 5 = a lot lower. lower than their risk of developing heart disease – when, in fact, their average risk of developing heart disease is Exposure to Lymphoma. Participants were asked whether approximately 30 times higher than their risk of develop- any of their family members or friends had ever had ing lymphoma. lymphoma. Response options were “None that I know of,” “1 (family member/friend),” “2–3 (family members/ The mean response to the question “Compared to other friends),” and “more than 3 (family members/friends).” women your age, do you think your chance of developing

234 Marketing and Public Policy Conference Proceedings / 2009 lymphoma is . . .” was 2.6 (SD = 0.9), on the same scale Discussion as the previous question. That is, on average, the women rated their risk of developing lymphoma as marginally The results of Study 2 are remarkably consistent with lower than other women their age. those of Study 1, in the virtual absence of any effect of risk information presentation variables on participants’ esti- Effect of Presentation Manipulations mates of own risk. Although the sample sizes were smaller than for Study 1, the confidence intervals were again Effect of Standard Presentation. As for Study 1, there consistently narrow enough to allow us to conclude that were no significant differences in risk estimates between there were no meaningful differences in risk estimates those participants in the standard presentation and other between the conditions. Thus, the lack of effects on risk presentation conditions, for any of the four dependent perceptions of the different presentation formats cannot variables. Thus, H1 was not supported. be attributed solely to small sample sizes. The only effect detected in this study was a shift toward more accurate Effect of Risk Ladder. As for Study 1, there were no estimates with the presence of a risk ladder, although this significant differences in risk estimates between those effect was only found for 10-year risk estimates, not for participants in the risk ladder present and risk ladder lifetime or comparative risk estimates; a counter-intuitive absent conditions for lifetime risk or risk compared to result as it would be expected that the presence or absence other women. There was, however, a significant differ- of a clear statement of the relative position of lymphoma ence between conditions for estimated 10-year risk, with risk compared to heart disease risk would increase the those in the ladder present condition reporting a lower accuracy of responses to the specific “compared to your estimated risk (15.9% versus 26.1%; a difference of 6.8%, risk of heart disease . . .” question. CI = -15.9 to 2.2; t = 2.1; p = .04). Thus, H2 was partially supported. Further, the mean difference in relative risk CONCLUSIONS compared to heart disease (which was clearly shown on the ladder to be many times the risk of breast cancer) was Both Study 1 and Study 2 found that the inclusion of -0.3 on a 5-point scale for risk (CI = -0.8 to 0.2). Thus, H3 presentation elements posited to reduce overestimations was not supported. of risk, such as risk ladders, and those posited to increase overestimations, such as visual representation, had no Effect of Different Base Rates. As for Study 1, there were impact on participants’ estimated risk of developing breast no significant differences in risk estimates between those cancer or lymphoma. Across the two studies, only one participants in the base 100 and base rate 1000 conditions, significant effect was noted; given that a total of 32 for any of the four dependent variables. Thus, H4 was not comparisons were made, and we used the criterion of a .05 supported. level of significance, it is not unexpected that we would find at least one apparently significant difference purely Effect of Visual Representation. As for Study 1, there by chance. were no significant differences in risk estimates between those participants in the visual representation present and It was suggested that the reason for the absence of effects visual representation absent conditions, for any of the four in Study 1 may have been due to participant’s awareness dependent variables. Thus, H5 was not supported. and exposure to high levels of media coverage relating to breast cancer, and thus predetermined perceptions of risk. Predictors of Risk Estimates Thus Study 2 focused on lymphoma, a far less well- known cancer and one which the majority of participants There was only one significant predictor of perceived risk had no personal experience. However, with one minor of developing lymphoma: having a family member with exception, the absence of effects of presentation elements lymphoma. Those with a family member with lymphoma on risk perceptions remained. provided significantly higher 10-year risk estimates (t = 3.3, p = .002) and lifetime risk estimates (t = 3.5, p = .001). Both Study 1 and Study 2 found that the women reported Having a family member with lymphoma was also the risk estimates for 10-year and lifetime risk that were only significant predictor of perceived comparative risk – considerably higher than their actual mean risk, contrary that is, risk compared to women of the same age (t = 3.3, to the principle of defensive optimism (where people p = .001) and compared to risk of developing heart disease perceive their own chances of negative consequences to (t = 1.9, p = .05). be lower than those of other people) which appears to

Marketing and Public Policy Conference Proceedings / 2009 235 apply to a wide range of negative health events. Although breast cancer than developing heart disease (Blanchard their risk estimates were not as high for lymphoma (Study 2) et al. 2002). as for breast cancer (Study 1), the actual risk of lymphoma is much lower than that of breast cancer – thus the degree Future studies should consider the impact of health lit- of overestimation is similar. eracy and numeracy, as well as the impact of prior exposure to medical conditions or information about these These findings appear to be in conflict with much of what conditions. It is also essential to undertake further exami- has been found in previous consumer research, particu- nation of the effects of different risk information presen- larly in the area of behavioral decision theory, which tation formats on perceived risk. Further studies are needed suggests that presentation format should have an impact to determine which formats have the greatest impact on on risk perceptions. The current study’s results have three risk perception; and which have no, or minimal, impact. important implications for future consumer research. First Ideally, this should include a meta-analysis of all previous and foremost, we need to examine the extent to which findings – including null findings (many of which are findings in consumer research are generalizable across likely to be unpublished studies) – to determine the extent topic areas, particularly in terms of the impact of to which current beliefs about the effects of different consumer’s preconceived estimates of disease risk and the formats are actually supported by the research evidence. extent to which these are resistant to change in the face of accurate statistical information. For example, previous Finally, as pointed out by Fischoff et al.(1998), it is studies have demonstrated that women’s inaccurate breast important to bear in mind that consumer’s risk decisions cancer risk perceptions often remain even after personal- are extremely complex: they are not just “waiting for a ized genetic counseling (Hopwood 2000). It is estimated missing number – upon receipt of which they will run the that the average woman’s risk of developing heart disease calculations needed to identify the best course of action” is approximately 32 percent (Lloyd-Jones et al. 1999); (p. 669). Thus, any research evidence on the effects of that is, three times her risk of developing breast cancer different formats needs to be considered – and, ideally, (Blanchard et al. 2002). However, several studies have conducted – in the context of consumer’s understanding shown that women are more concerned about developing of the issues, information needs, beliefs and values, and other influences in their environment.

ENDNOTES J. Rothman, Anne Moyer, John Beauvais, and Elissa Epel (1995), “The Effects of Message Framing on 1 As the risk of breast cancer, unlike the hypothetical Mammography Utilization,” Health Psychology, 14 vaccine in the Kaplan (1985) study, is not a very small (2), 178–84. risk. Barratt, Alexandra, Jill Cockburn, John Lowe, Christine 2 Ranging from 1.5 percent for a 40 year old woman to Paul, Janice Perkins, and Sally Redman (1997), Re- 10.7 percent for a 70 year old. port on the 1996 Breast Health Survey. Woolloomooloo NSW: NHMRC National Breast REFERENCES Cancer Centre. Black, William C., Robert F.J. Nease, and Anna N. Ad Hoc Committee on Health Literacy for the Council on Tosteson (1995), “Perceptions of Breast Cancer Risk Scientific Affairs (1999), “Health Literacy: Report of and Screening Effectiveness in Women Younger the Council on Scientific Affairs,” Journal of the Than 50 Years of Age,” Journal of the National American Medical Association, 281 (6), 552–57. Cancer Institute, 87, 720–31. Ahrens, Robert W. (2002), “Men Executed Far More Blanchard, D., Joel Erblich, Guy H. Montgomery, and Than Women,” USA Today, 1A. Dana H. Bovbjerg (2002). “Read All About It: The Baker, David W., Mark V. Williams, Ruth M. Parker, Over-Representation of Breast Cancer in Popular Julie A. Gazmararian, and Joanne Nurss (1999), Magazines.” Preventive Medicine, 35 (4), 343–48. “Development of a Brief Test to Measure Functional Cleveland, William S. and Robert McGill (1984), “Graphi- Health Literacy,” Patient Education and Counseling, cal Perceptions: Theory, Experimentation, and Ap- 38, 33–42. plication to the Development of Graphic Methods,” Banks, Sara M., Peter Salovey, Susan Greener, Alexander Journal of the American Statistical Association, 70,

236 Marketing and Public Policy Conference Proceedings / 2009 531–54. Cascades and Risk Regulation,” Stanford Law Re- Cox, James C. and David M. Grether (1994), The Prefer- view, 51, 683–768. ence Reversal Phenomenon: Response Mode, Mar- Lalor, K. and Barbara J. Hailey (1990), “The Effects of kets and Incentives. California: California Institute of Message Framing and Feelings of Susceptibility to Technology, Department of Economics. Breast Cancer on Reported Frequency of Breast Self- Feldman-Stewart, Deb, Nancy Kocovski, Beth A. Examination,” International Quarterly of Commu- McConnell, Michael D. Brundage, and William J. nity Health Education, 10 (3), 183–92. Mackillop (2000), “Perception of Quantitative Infor- Lipkus, Isaac M, Maggie Kuchibhatla, Colleeen M. mation for Treatment Decisions,” Medical Decision McBride, Hayden B. Bosworth, Kathryn I. Pollak, Making, 20, 228–38. Ilene C. Siegler, and Barbara K. Rimer (2000), “Re- Fischhoff, Baruch, Donna Riley, Daniel C. Kovacs, and lationships Among Breast Cancer Perceived Abso- Mitchell Small (1998), “What Information Belongs lute Risk, Comparative Risk, and Worries,” Cancer in a Warning?” Psychology and Marketing, 15 (7), Epidemiology, Biomarkers & Prevention, 9 (9), 973– 663–86. 75. Gigerenzer, Gerd and Ulrich Hoffrage (1995), “How to ______, Greg Samsa, and Barbara K. Rimer (2001), Improve Bayesian Reasoning Without Instruction: “General Performance on A Numeracy Scale Among Frequency Formats,” Psychology Review, 102, 684– Highly Educated Samples,” Medical Decision Mak- 704. ing, 21 (1), 37–44. Grewal, Dhruv, Jerry Gotlieb, and Howard Marmorstein Lloyd-Jones, Donald M., M.G. Larson, A. Beiser, and D. (1994), “The Moderating Effects of Message Fram- Levy (1999), “Lifetime Risk of Developing Coro- ing and Source Credibility on the Price-Perceived nary Heart Disease,” Lancet, 353 (9147), 89–92. Risk Relationship,” Journal of Consumer Research, Loomis, John B. and Pierre H. duVair (1993), “Evaluating 21 (June), 145–53. the Effect of Alternative Risk Communication De- Helzlsouer, Kathy J., Daniel Ford, Robert Hayward, vices on Willingness to Pay: Results From a Dichoto- Melinda Midzenski, and Helen Perry (1994), “Per- mous Choice Contingent Valuation Experiment,” ceived Risk of Cancer and Practice of Cancer Preven- Land Economics, 69, 297. tion Behaviors Among Employees in An Oncology McGuire, William J. (1985), “Attitudes and Attitude Centre,” Preventive Medicine, 23, 302–8. Change,” in Handbook of Social Psychology, Gardner Hoeffler, Steve and Dan Ariely (1999), “Constructing Lindzey and Elliot Aronson, eds. New York: Ran- Stable Preferences: A Look Into Dimensions of Ex- dom House, 3rd ed. 2, 233–346. perience and Their Impact on Preference Stability,” Memmott, Mark (2002), “Fear May Is Overwhelming, Journal of Consumer Psychology, 8 (2), 113–39. But So Are the Odds,” USA Today, 7A. Hoffrage, Ulrich (2003), “Risk Communication in the Merrill, Ray M., Larry G. Kessler, Joshua M. Udler, Early Identification of Breast Cancer and Hormone- Gloria C. Rasband, and E.J. Feuer (1999), “Compari- Replacement Therapy,” Zeitschrift fur son of Risk Estimates for Selected Diseases and Gesundheitspsychologie, 11 (3), 76–86. Causes of Death,” Preventive Medicine, 28, 179–93. Hopwood, Penelope (2000), “Breast Cancer Risk Percep- Meyerowitz, Beth E. and Shelly Chaiken (1987), “The tion: What Do We Know and Understand?” Breast Effect of Message-Framing on Breast Self-Examina- Cancer Research and Treatment, 2, 387–91. tion Attitudes, Intentions, and Behavior,” Journal of Jones, Sandra C. (2004), “Coverage of Breast Cancer in Personality and Social Psychology, 52 (3), 500–10. the Australian Print Media – Does Advertising and NHMRC National Breast Cancer Centre (1999), “Don’t Editorial Coverage Reflect Correct Social Marketing Ask for An Opinion – Ask for the Scalpel; Print Messages?” Journal of Health Communication, 9 Media Coverage of Breast Cancer in Australia in (4), 1–17. 1995,” NHMRC National Breast Cancer Centre, 14/ Kaplan, Robert M.B., Bonnie Hammel, and Leslie E. 12/00, 2000. Schimmel (1985), “Patient Information Processing Petty, Richard E. and John T. Cacioppo (1986), “The and the Decision to Accept Treatment,” Journal of Elaboration Likelihood Model of Persuasion,” Ad- Soc Behav Pers, 1 (1), 113–20. vances in Experimental Social Psychology, 19, 123– Krishnamurthy, Parthasarathy, Patrick Carter, and Ed- 205. ward Blair (2001), “Attribute Framing and Goal Savage, Leonard J. (1954), The Foundations of Statistics. Framing Effects in Health Decisions,” Organiza- New York: Wiley. tional Behavior and Human Decision Processes, 85 Schapira, Marilyn M., Ann B. Nattinger, and Colleen A. (2), 382–99. McHorney (2001), “Frequency or Probability? A Kuran, Timur and Cass R. Sunstein (1999), “Availability Qualitative Study of Risk Communication Formats

Marketing and Public Policy Conference Proceedings / 2009 237 Used in Health Care,” Medical Decision Making, Family History of Breast Cancer,” Surgery, 120, (November/December), 459–67. 297–303. Simonson, Itamar, Ziv Carmon, Ravi Dhar, Aimee Drolet, Vernon, Sally W. (1999), “Risk Perception and Risk and Steve Nowlis (2001), “Consumer Research: In Communication for Cancer Screening Behaviors: A Search of Identity,” Annual Review of Psychology, Review,” Journal of the National Cancer Institute 52, 249–75. Monographs, 25, 101–19. Slovic, Paul (1995), “The Construction of Preferences,” vonNeuman, John and Oscar Morgenstein (1947), Theory American Psychologist, 50 (May), 364–71. of Games and Economic Behavior. Princeton, NJ: Smith, Barbara L., Michele A. Gadd, C. Lawler, D.J. Princeton University Press. MacDonald, S.C. Grudberg, F.S. Chi, K. Carlson, A. Yamagishi, Kimihiko (1997), “When a 12.86 Percent Comegno, and W.W. Souba (1996), “Perception of Mortality is More Dangerous Than 24.14 Percent: Breast Cancer Risk Among Women in Breast Center Implications for Risk Communication,” Applied Cog- and Primary Care Settings: Correlation with Age and nitive Psychology, 11, 495–506.

For further information contact: Sandra C. Jones Centre for Health Initiatives University of Wollongong Northfields Ave. Wollongong NSW 2522 Australia Phone: +61.2.4221.5106 Fax: +61.2.4221.3370 E-Mail: [email protected]

238 Marketing and Public Policy Conference Proceedings / 2009 THE EFFECT OF FRONT-OF-PACKAGE NUTRITION INFORMATION AND PRODUCT CLAIMS ON CONSUMERS’ ATTITUDINAL EVALUATIONS AND CHOICE BEHAVIOR

Ninya Maubach, Massey University, New Zealand Janet Hoek, University of Otago, New Zealand Philip Gendall, Massey University, New Zealand Duncan Hedderley, Institute for Plant and Food Research, New Zealand

SUMMARY The Elaboration Likelihood Model (ELM, Rucker, and Petty 2006) provides a theoretical basis for expecting that There is global debate over the antecedents of unhealthy more detailed and complex information, such as con- weight gain and how these can be modified to improve tained in PDI nutrition labels, require effort and ability to population health. In particular, interventions that will read and comprehend, and thus only a proportion of prevent children and adolescents from becoming over- consumers will access this information. Conversely, the weight and obese are sought, as this is correlated with simple nature of TLLs suggests that most consumers will long-term adult health (Ferraro et al. 2003). While both be able to interpret this format, and this it should affect energy intake and expenditure influence weight, evidence their product evaluations. Therefore, this research tests suggests that reducing energy consumption should have a the hypothesis that TLLs should improve consumers’ greater effect on children’s weight (Swinburn et al. 2006). ability to discriminate between nutritionally superior and Several authors have outlined how food marketing prac- inferior food products. tices encourage over consumption of energy dense food, and have suggesting restrictions on marketing to children Manufacturers also print on their packaging claims about (Zimmet and James 2006). However, industry representa- the nutrients present in products and the health benefits tives argue regulatory restrictions on business activities these have. Consumers are thought able to access his are inappropriate (Irwin 2005), and the food industry has information quite readily also, as verbal information has proactively developed self-regulatory initiatives intended a greater effect on assessments of product healthiness than to combat rising obesity rates. numerical information, even when the two types of information provide conflicting evidence (Viswanathan 1996). One action taken by industry in Australia and New Zealand Thus, this research tested the hypothesis that product is voluntary inclusion of Percent Daily Intake (PDI) claims have a greater influence on consumers’ evalua- nutrition labels on the front-of-packages (FOP). Industry tions when only numeric nutrition information is avail- argues that PDI labels will enable consumers to recognize able. foods’ energy content and thus to manage their weight. However, health lobbyists claim the PDI format is too Two online experiments were conducted with adult New complex, and have called for regulators to mandate an Zealand consumers in 2008. The first study, informed by alternative FOP format, the Traffic Light Label (TLL) the ELM, was a 2 x 3 x 2 factorial experiment manipulat- developed by the British Food Standards Agency (FSA ing the FOP nutrition information format (PDI, TLL) and 2006). Food Standards Australia New Zealand regulates three levels of product claim (no claim, nutrition-content the manufacture and sale of foods in both Australia and claim, health claim), across two nutrition profiles (better, New Zealand, and has begun reviewing regulatory op- worse). The main dependent variable was consumers’ tions for FOP nutrition labels (Australia and New Zealand attitudinal responses. A three-way ANOVA analyzed the Food Regulation Ministerial Council 2008). Similarly, main effects of, and interactions between, the three IVs on the recent United States Government Accountability Of- respondents’ attitudes, followed by a series of indepen- fice (GAO 2008) report into food labeling recommends dent-samples t-tests to explore the main effects in more the FDA investigate options for a simplified FOP labeling detail. A final set of t-tests explored how the FOP label system. formats influenced consumers’ perceptions of the cereals’

Marketing and Public Policy Conference Proceedings / 2009 239 macronutrient content (sugar, sodium, total fat and fiber). The results from the choice research differed slightly. The The second study was a stated preference discrete choice major variation between findings in the two studies related modeling experiment, designed to estimate the effect of to the addition of nutrient and health claims; these had the nutrition format and claim types on choice behavior strongest effect on consumers’ behavior in the choice (Keller et al. 1997). Participants saw ten pairs of cereal experiment, particularly when nutrition information was images and were instructed to select the one they would presented numerically. However, similar to the pattern in buy for their children, with “Neither” as a third option to the first study, the choice experiment showed that the TLL make the task more realistic and improve data quality format led to the greatest change in utility when the (Dhar and Simonson 2003). The choice data were ana- product nutritional profile improved, suggesting that this lyzed using a multinomial logit model. FOP format is most easily understood by consumers.

Study one found that the two FOP labels had differential Overall, traffic light nutrition labels appear to help con- effects on respondents’ attitudes to breakfast cereals with sumers evaluate energy-dense products more accurately diverse nutrition profiles. Specifically, attitudes toward that PDI labels. TLLs also lowered stated preferences for the two cereal profiles (better, worse) were equivalent nutritionally inferior products more than when a PDI label when shown the PDI label, suggesting this format had was used, and TLLs increased preferences for nutrition- little influence on attitude formation. However, attitudes ally superior products substantially more than when nutri- differed in the TLL conditions, being significantly lower tion information was presented in numeric formats. Con- for the nutritionally worse cereal. Furthermore, respon- sumers were also more influenced by nutrient claims dents’ perceptions of macronutrient levels varied accord- when they were also presented with numeric nutrition ing to the FOP label format. The TLL format had a greater information. Given the ubiquitous nature of nutrient claims, effect on consumers’ ability to identify and recall the which may divert attention away from products’ overall quantity of sugar and total fat, in particular. The addition nutritional values, this is a further rationale for using of a nutrient content or health claim did not significantly easily understood nutrition information formats. Refer- change consumers’ product evaluations over a control ences are available upon request. condition. This finding was not anticipated.

For further information contact: Janet Hoek Department of Marketing University of Otago P.O. Box 56 Dunedin New Zealand Phone: 0064.3.479.7697 Fax: 0064.3.479.8172 E-Mail: [email protected]

240 Marketing and Public Policy Conference Proceedings / 2009 PERCEPTIONS OF NUTRITION SYMBOLS BY SOME U.S. CONSUMERS

Chung-Tung Jordan Lin, U.S. Food and Drug Administration, College Park

SUMMARY symbol; and (5) a front panel showing the United Kingdom’s Multiple Traffic Lights (MTL) symbol. Par- Introduction ticipants were asked to rank the labels in terms of how well each label conveys nutrition characteristics, and then An emerging trend of food labeling in the United States discuss each symbol concept. Section 3 repeated the and many other countries is the use of nutrition symbols second section with the addition that each of the symbol on the front panels of packaged food labels. Manufactur- concepts was accompanied by a brief explanation of the ers and retailers claim that symbols can help consumers nutrition criteria. The explanations were developed partly make healthier food selections by providing them a con- using materials available on industry Web sites. For venient point-of-purchase snapshot of the nutrition pro- instance, the explanation for the GC symbols states “Com- file of a product. Yet each symbol system has its own pany B has developed 26 symbols based on criteria format, includes different nutrient information, and has its established by the U.S. Food and Drug Administra- own nutrient criteria. FDA believes there is a need to tion . . . ,” with the Italicized text taken from General examine how consumers understand and respond to these Mills’ Web site. (Note: Most of the company’s standards symbols, but relevant and nonproprietary consumer infor- are either not consistent with current food labeling regu- mation is sparse. FDA therefore commissioned a focus lations or not specified by any current regulations.) The group study on a sample of existing and alternative groups ended with brief discussions on other topics re- symbols. The study examined: (1) exposure, understand- lated to symbols, such as standardization and restaurant ing, and use of existing symbols; (2) confusion caused by labeling. Participants were instructed to consider the five various symbols; and (3) perceived merits and deficien- concepts as how five different companies may choose to cies of various symbols. The information will help inform present product information. The symbols were shown the design of subsequent quantitative evaluation of the side-by-side on a sheet of 11" by 14" paper. Participants effects of symbols on consumers. were reminded that a Nutrition Facts (NF) label for the product was available on the back of the paper. Methodology

In April 2008, the same moderator conducted two focus Major Findings groups in Greenbelt, MD and two in Albany, NY. At each location, one group consisted of college-or-more edu- Many participants considered nutrient-specific symbols, cated adults and another group adults with high school i.e., the United Kingdom’s Multiple Traffic Lights (MTL) diploma or GED. Thirty-seven adult consumers, mixed by symbol and General Mills’s Goodness Corner (GC) sym- age, gender, and race/ethnicity, participated in the discus- bols, more informative than summary symbols (the Healthy sion. The moderator led the discussions through three Check (HC) and Guiding Stars (GS) symbols) in provid- sections. Section 1 covered participants’ general attitudes ing nutrition information about the product. Many partici- and practices about foods, diets, and nutrition. Section 2 pants also saw the MTL as “more honest” because it focused on participants’ perceptions of and reactions to shows nutrient contents and both positive (e.g., “low total different symbols presented on the front panel of a single fat”) and negative (e.g., “high sugars”) attributes of the unbranded raisin bran package. Five concepts were pre- product. Participants appreciated the potential helpful- sented: (1) a front panel without a symbol; (2) a front panel ness of symbols in grocery shopping, but many of them showing three General Mills Goodness Corner (GC) also recognized the usefulness of the NF label as well as symbols [“GOOD SOURCE – WHOLE GRAIN,” “GOOD the limitations of symbols. Some participants considered SOURCE – CALCIUM,” and “EXCELLENT SOURCE – the GC and HC symbols a screener or attention grabber IRON”;] (3) a front panel showing a fictitious “Healthy that would lead them to look for further product informa- Check” (HC) symbol that was created for the purposes of tion, particularly the NF label. Except for the MTL and the the study; (4) a no-symbol front panel with a shelf tag GS symbols, symbols were similar to advertising to some underneath showing a Hannaford Guiding Stars (GS) participants.

Marketing and Public Policy Conference Proceedings / 2009 241 When asked, many participants interpreted the GC, HC, product packages; mentioned reasons included more trust and GS symbols to indicate a product is “healthier,” in manufacturers, suspicion of retailer motives, and suspi- “nutritious,” or “good for you.” A symbol-absent product, cion of misplacement of tags for undeserving products. however, was not seen as necessarily less nutritious than Many participants thought the presence of various differ- a product that carries a symbol. Even though participants ent symbols in the market would be confusing and they perceived symbols as indicating better product choices, would prefer a standardized symbol. how symbols may affect intended or actual purchase was less clear. Some participants noted they would select a Conclusions symbol-carrying product if they were in a hurry. Others, said symbols would trigger a search for product informa- The focus group research provides some preliminary and tion. Many participants mentioned taste and price as qualitative observations about possible ranges and varia- important purchase considerations. tions in consumer reactions to various types of nutrition symbols. This information will help fill the gaps in exist- Explanation of criteria behind the symbols raised the level ing literature regarding this emerging type of label infor- of confidence that some participants felt toward certain mation on the nutritional characteristics of food products. symbols. Many participants considered nutrient criteria The study will also help the development of quantitative established by FDA more trustworthy than that by “na- research to investigate the relationships between features tional and international health organizations.” Many par- of symbols and other label information and product per- ticipants also preferred symbols that take account of ceptions and attitudes. Finally, the study will help provide multiple nutrients to symbols that consider only indi- a useful reference point for dialogues between interested vidual nutrients. Some participants felt the shelf-tag sym- parties. bol (Guiding Stars) was not as trustworthy as a symbol on

For further information contact: Chung-Tung Jordan Lin U.S. Food and Drug Administration 5100 Paint Branch Parkway, HFS–020 College Park, MD 20740 Phone: 301.436.1831 Fax: 301.436.2637 E-Mail: [email protected]

242 Marketing and Public Policy Conference Proceedings / 2009 BEYOND THE WATER COOLER: GENDER DIFFERENCES IN PRIVACY-RELATED MEASURES FOR FACEBOOK PROFILE USAGE BEYOND ITS ORIGINAL PURPOSE

Mariea Grubbs Hoy, University of Tennessee, Knoxville George R. Milne, University of Massachusetts, Amherst

SUMMARY information beyond its original purpose (including behavioral advertising)? Background RQ3: How do men and women differ in terms of privacy- At an estimated 18.6 million, adults ages 18 to 24 are the related behaviors including SNS privacy policy largest age cohort on Social Networking Sites (SNSs). readership, pro-active self-protective behaviors and The expected transparency among those who create SNS self-protective behaviors that require others’ com- profiles implies that individuals who communicate and pliance. share information within social networks have expectations that their personal information will be identifiable Method Overview and Sample and accessible to their friends. However, the line between what is private – that is intended for a limited, user- This study used an online survey that by virtue of recruit- determined audience – and what is “public” and usable by ing through an invitation to join a Facebook group at- third parties, including advertisers, is increasingly blurred. tracted users from a variety of universities/colleges and Additionally, while students disclose extensive personal states. In essence, because the very instrument studied (an information on their SNS profiles they often are unaware SNS) was the mechanism used to recruit respondents and of the extent of those profiles’ visibility (Acquisiti and direct individuals to the online means of collecting data, Gross 2006; Gross and Acquisiti 2005; Kolek and Saunders this was more akin to a field experiment as the information 2008). was collected in the real environment.

SNSs provide an online environment for forming and One thousand-thirty individuals clicked the link to begin sustaining a sense of community and interpersonal rela- the survey. After restricting the sample to 18–24 year olds tionships. Women tend to be more interpersonally-ori- residing in the U.S., the final analysis sample of suffi- ented than men (Jackson et al. 2001) and are described as ciently completed surveys was 589 (72.7% female and “enthusiastic online communicators” (Fallows 2005). Not 27.3% male.) Nearly three-fourths of the sample was 19 to surprisingly, women use SNSs more than men (Hargittai 21. The sample was overwhelmingly Caucasian (90.7%). 2008). Prior research shows that women express more Although the respondents were concentrated in South- concern about online privacy compared to men and en- eastern states, 31 states and the District of Columbia were gage in noticeably different self-protective behaviors represented. (Sheehan 1999). However, the context of these findings was in Web 1.0 from a decade ago. We extend this Results research into Web 2.0 by examining gender differences relative to privacy measures regarding SNSs’ profile First, we found that although women and men had similar information. This study reports the results of an online number of friends and usage of SNSs, women were more survey among 589 18–24 year old Facebook users that concerned about their privacy being invaded while using addresses the following research questions: SNSs. Second, with respect to using SNS profiles for behavioral advertising, both women and men were against RQ1: How do men and women differ in terms of privacy this. However, almost half of both genders were not concern, attitudes and beliefs with respect to Social informed about how this information is being used. Addi- Networking Sites? tionally, both genders indicated a strong desire to be alerted when a company is tracking them. Third, women RQ2: How do men and women differ in terms of attitudes are more likely than men to take proactive self-protective and beliefs regarding usage of their SNS profile behaviors. Interestingly, this finding is in opposition to

Marketing and Public Policy Conference Proceedings / 2009 243 research a decade prior (Sheehan 1999). This study’s industry as a signal regarding “the FTC’s aggressiveness results provide an updated view of women protecting their in online privacy, behavioral marketing and in enforce- privacy in a Web 2.0 environment. ment” (Teinowitz 2009, p. 1). Conspicuously notifying online consumers of behavioral advertising activity, rather Recommendations than burying notice in a privacy policy, may be viewed as a stronger commitment to self-regulation. The FTC’s Report (2009) Self-Regulatory Principles for Online Behavior underscores the importance of transpar- Getting more women, as well as men, aware of the ency and control. Not only is the Commission calling for potential dangers of social networking information dis- companies to provide “meaningful disclosures,” but the closure and encouraging them to take privacy protection present study found that young adult SNS users are into their own hands is important. SNSs, as well as woefully unaware of how their profile information might companies that engage in behavioral advertising, have be used and strongly desire to be informed when they are unique insight into social network users and their online being tracked. President Obama’s appointment of Jon behavior that could be used to engage in consumer educa- Leibowitz as FTC Chairman is viewed by the advertising tion efforts in this area.

For further information contact: Mariea Hoy 476 College of Communication and Information University of Tennessee Knoxville, TN 37996 Phone: 865.974.3048 Fax: 865.974.2826 E-Mail: [email protected]

244 Marketing and Public Policy Conference Proceedings / 2009 WHAT CALIFORNIANS UNDERSTAND ABOUT PRIVACY OFFLINE

Chris Jay Hoofnagle, University of California, Berkeley Jennifer King, University of California, Berkeley

ABSTRACT of Acquisti and Turow, this paper proposes an explanation for the gulf between privacy concern and privacy Americans express high levels of privacy concern and yet behavior: consumers falsely believe that privacy law reveal information to marketers with alacrity. This survey strongly protects personal information in common trans- of Californians may explain this tension: consumers mis- actions. To test this explanation, we asked a representative takenly believe that privacy law protects marketing data, sample of California consumers about rules for commer- and those opposed to new privacy legislation are more cial use of personal information in nine contexts. likely to falsely believe that privacy protections already exist. In six of commercial contexts (pizza delivery, donations to charities, product warranties, product rebates, phone INTRODUCTION numbers collected at the register, and catalog sales), a majority either didn’t know or falsely believed that the If consumers care about protecting information privacy, law prohibited merchants from selling their personal why do they so frequently give away personal data to information to others. In one – grocery store club cards – marketers with alacrity? Public opinion polls show that a majority did not know or thought information could be Americans’ concerns over information privacy have in- sold when California law prohibited the sale. Only in two creased, despite, (or perhaps because of) the 9/11 terrorist contexts – newspaper and magazine subscriptions and attacks (Best et al. 2006). In a review of legislative debates sweepstakes competitions – did a majority of our sample on privacy, Gandy (2003) found that many attempting to of Californians understand that personal information col- influence legislation cited to high levels of privacy con- lected by the company could be sold to others. cern in their testimony. Privacy concern remains high, and a highly examined aspect of public policy concerning Professor Alan Westin (2001) has proposed a three-group marketing. Yet U.S. consumers participate in marketing segmentation of the American public on privacy, where surveys by the millions. They readily provide their phone consumers are typed as having high, mid, or low-level numbers to cashiers upon request. And they participate in privacy concern. Westin has argued that individuals with loyalty card programs without reservation. American mid to low-level privacy concern favor industry self- consumers seem to betray their privacy attitudes by ex- regulation over the imposition of privacy laws. We seg- pressing different preferences in the marketplace. The mented our sample of consumers according to the Westin persistence of privacy-risky behavior has undermined the model, and found that individuals with high levels of findings that privacy is an important concern. Many have concern answered correctly significantly higher than indi- concluded that consumers say they want privacy, but in viduals with lower levels of concern. This result indicates fact do not. This gulf between behavior and concern has that opposition to new privacy laws among low and mid- become a rationale for not expanding protections for concern consumers may be the product of a mistaken personal data. belief that privacy laws already exist and protect them in marketplace transactions. Acquisti and Grossklags (2005) argue that privacy decision making appears irrational because consumers face BACKGROUND several barriers in pursuing optimal strategies. First among these barriers is incomplete information: consumers do Online Privacy Problems Find Roots in Offline Rules not know how firms will use personal information, nor can they easily identify the risks and consequences of Much attention is focused upon the privacy problems privacy decisions. In 2003 and 2005, Turow et al. exposed by new technologies and information-intensive plored Americans’ privacy knowledge in nationally rep- business models online. For instance, social networking resentative polls, finding significant misunderstandings website Facebook.com’s “Beacon” system, an applica- of privacy law among consumers. Building upon the work tion which informs Facebook users’ friends about pur-

Marketing and Public Policy Conference Proceedings / 2009 245 chases made and activities on other websites, inspired ing Enterprises, a Canadian firm that was investigated widespread protests among Facebook users and others multiple times but never charged with wrongdoing.” when it launched in 2007 (Story and Stone 2007). Some users objected vigorously to the Beacon application be- In earlier work, to learn more about information selling we cause their activities were reported on an opt-out basis, used a California privacy law to make requests to 86 meaning that the user had to take affirmative action to companies for a disclosure of their information sharing prevent others from learning about their activities. practices. We found that while many companies have Facebook responded to these objections by changing its voluntarily adopted a policy of not sharing personal policy to obtain express approval before activities per- information with third parties, many still operate under an formed on other websites would be shared with friends. opt-out model, and others simply did not respond to the request (Hoofnagle and King 2007). In many cases, these information sharing problems ex- isted in the offline world long before consumers embraced Consumer Preferences for Opt-In Rules e-commerce. Facebook’s Beacon simply lifted the veil on the already common business practice of selling informa- As explained above, an opt-out standard requires the tion about consumers and their purchases. consumer to take some affirmative action to prevent the sale of personal information. Opt-in rules shift the default In the offline world, collection and disclosure of personal so that the company bears the burden to obtain affirmative information is widespread, but less efficient and generally consent from the consumer before information is sold. unaccompanied by privacy notices. Companies are free to This consent is not presumed when a consumer merely sell personal information about their customers unless buys a product or service; opt-in is a separate process from they promise not to under a self-adopted privacy policy or the decision to purchase a product. when state or federal law restrains information sale. Many companies quietly sell customer lists and other informa- Generally speaking, companies prefer opt-out rules to tion, as evidenced by the DirectMag Listfinder, a collec- opt-in rules, because of transaction costs associated with tion of over 60,000 “datacards,” advertisements for the obtaining affirmative consent. In some cases, companies sale of personal information (Direct Magazine). Catalog favor “no-opt,” meaning that the consumer has no option sales companies often provide customer lists to “coopera- to limit sharing of personal information for marketing. tive databases” where many businesses trade personal information. When a company enters into a cooperative In his review of the use of public opinion polls to influence database relationship, hundreds of other businesses have legislative policy on privacy laws, Gandy (2003) noted access to the company’s customer list. Similarly, ordinary that “. . . none of the surveys cited in congressional consumer behavior such as ordering a pizza for delivery, testimony inquired specifically about whether the public subscribing to a magazine or newspaper, or signing up for preferred that ‘opt-in’ would be the default. Moreover, a sweepstakes contest often results in contact information there was no mention of any surveys in which consumers being sold to third parties. were asked to express their preferences for one option over the other.” Nevertheless, there is evidence that third In addition to subjecting individuals to potentially un- party information sharing based on an opt-out standard is wanted marketing communications, customer list selling strongly opposed by the public. As Joanne McNabb, also exposes individuals to fraud (Duhigg 2007). Many Chief of California’s Office of Privacy Protection, ex- lists characterize customers pejoratively, with euphe- plains, “Consumers are increasingly very unhappy with misms such as “spontaneous” and “impulsive” used to sharing of their information for marketing purposes” mark easy sells and people less skeptical of advertising (Trager 2005). Few public opinion polls directly ask claims. One such list obtained by the New York Times respondents about opt-in preferences. A meta study of 43 read: “These people are gullible. They want to believe that opinion polls on privacy by Sheehan found that only four their luck can change” (Duhigg 2007). The same article touched upon opt-in issues (Sheehan 2004). goes on to recount multiple incidents where InfoUSA, a direct marketing company, sold information to fraudsters: Commercial firms that rely upon information exchange “Between 2001 and 2004, InfoUSA also sold lists to have been the primary sponsors of public opinion research World Marketing Service, a company that a judge shut in privacy (Gandy 2003). That sponsorship allows such down in 2003 for running a lottery scam; to Atlas Market- firms to control what questions are asked, the framing of ing, which a court closed in 2006 for selling $86 million the debate, and perhaps even the outcome of those debates of bogus business opportunities; and to Emerald Market- (Gandy 2003). When the opt-in issue is polled, large

246 Marketing and Public Policy Conference Proceedings / 2009 majorities of Americans indicate that they support requir- may lead policy makers to come to a different conclusion ing businesses to obtain affirmative consent before selling on the appropriate role for government interventions to personal information to third parties. Sheehan (2004) protect privacy. notes that of the polls she reviewed that asked respondents about opt-in, support for affirmative consent ranged be- How Well Do Californians Understand the Rules of tween 78 percent – 88 percent. Offline Information Sharing?

The Westin Taxonomy Applied to Offline Privacy In this paper, we assess Californians’ understanding of information privacy rights concerning ordinary, offline In 1995, Professor Priscilla Regan concluded that: “Both purchases. historical examples and the results of public opinion surveys demonstrate that people value privacy as an ideal We asked Californians about the sale of personal informa- that is important in the abstract and as a factor that is tion in nine contexts where companies commonly collect understood in real-life situations . . . a great deal of latent personal data and sell it to other businesses. These con- public policy concern about privacy exists” (Regan 1995). texts are all common commercial experiences for most Professor Alan Westin has pioneered a popular “segmen- Americans, they usually occur offline, and a privacy tation” of privacy attitudes among the American public to policy is typically not available to the consumer before or gauge this concern and to make policy recommendations during the transaction. They are: subscribing to a newspa- (Kumaraguru and Cranor 2005). In Westin’s segmenta- per, ordering a pizza, donating to a charity, participating tion, Americans are divided into three groups: “Privacy in a sweepstakes, registering a product warranty, giving a Fundamentalists,” who represent about 25 percent of the phone number to a cashier at checkout, registering a public, and place a high value on privacy and favor product rebate, ordering from a catalog, and using a passage of strong privacy laws; “Privacy Pragmatists,” “loyalty card” at a grocery store. We were also interested represent about 55 percent of the public, and see the in how attitudes toward privacy are linked to knowledge relative benefits of information collection and favor vol- about these offline business practices. Thus, we divided untary standards for privacy protection; and the “Privacy our data subjects into groups using Professor Westin’s Unconcerned” represent about 20 percent of the public, segmentation of Privacy Fundamentalists, Privacy Prag- and have low privacy concern and have little objection to matists, and the Privacy Unconcerned. giving government or businesses personal information. METHODS While it has not received extensive treatment in the academic literature, the Westin segmentation has been The survey questions were asked as part of the 2007 influential politically and is often used to frame Ameri- Golden Bear Omnibus Survey, a telephone-based survey cans’ privacy attitudes. Generally, Westin (2001) has of a representative sample of California residents con- emphasized that public policy and business approaches ducted by the Survey Research Center of University of should be targeted to the privacy pragmatist, explaining California, Berkeley. The Samuelson Law, Technology & that: “In the politics of privacy, the battle is for the hearts Public Policy Clinic funded the privacy portion of the and minds of the Privacy Pragmatists. If most of them feel Golden Bear survey from general operating funds; no that their personal information is being used fairly and outside organization sponsored the survey. The dual frame properly by businesses, especially online, they join the sample used random digit dialing of both cell phones and Privacy Unconcerned to make up a 75 percent level residential landline telephones, with one respondent per support for the existing rules and practices. But if most of landline household selected. English and Spanish speak- the Privacy Pragmatists feel that information practices are ers over the age of 18 were eligible. 1,186 respondents intrusive or their information is being misused, they join completed the telephone interview, conducted from April the Privacy Fundamentalists to make up a majority seek- 30th to September 2nd 2007, for a response rate of 15.9 ing legislative or regulatory measures, or consumer boy- percent. However, in order to include more questions in cotts.” the survey than could be administered to all respondents in a reasonable period of time, the sample was divided into Westin places a high importance on the public’s attitudes six randomized parts or units. All respondents were asked toward government and private sector use of personal certain basic demographic and background questions, but information. Ours examines a different question: how most questions were administered only to 5/6th of the well do individuals understand the rules surrounding complete sample. This reduced the number of respon- information use? Their understanding may inform pri- dents who answered our questions to 991. Weights were vacy sensitivity, and lack of knowledge of privacy rules applied to compensate for probabilities of selection and to

Marketing and Public Policy Conference Proceedings / 2009 247 match certain demographic distributions. This weighting other companies, unless I give them explicit permission.” ensures that the results reflect a representative sample of Fifty point nine percent correctly answered false. A simple Californians by age, education, ethnicity, and gender, and majority understands that newspapers and magazines are compensates for differences in probabilities of selection free to sell subscription lists without subscriber consent. based on use of landline versus mobile phone. But a sizable minority – 49.1 percent – either does not know or believes that opt-in consent is required. These Responses to nine questions provided the basis of analysis results are similar to those obtained by Turow et al. (2005) for this paper. The offline questions were prefaced with in a national sample of adults who use the internet. Turow the following statement to convey that respondents should found that 48 percent correctly answered false to the answer them in the context of ordinary, offline transac- statement, “When I subscribe to a magazine, by law that tions: “Now, I would like to ask a series of questions about magazine cannot sell my name to another company unless shopping in the ‘real world,’ such as in grocery stores or I give it permission.” Of the remaining respondents in shopping malls.” Turow’s survey, 36 percent incorrectly answered true and 16 percent didn’t know. Finally, reviewers have frequently asked why we only polled California residents. We research and teach at the Ordering a Pizza for Delivery University of California. A California-focused poll seemed especially relevant, in part because so many privacy laws Next, we asked consumers about a popular treat: ordering adopted at the federal level are first passed in California’s a pizza for delivery. Pizza delivery companies, since they legislature. And frankly, we Californians consider our are called so frequently by consumers, are a hub for State to be a sovereign nation. Nevertheless, we plan a collecting personal information. A delivery company can national poll exploring similar issues in 2009. collect and aggregate caller identification information (typically name and phone number), ask the customer for Our Research Standards their phone number (which may be different than what is displayed by caller identification), and in order to process We hold ourselves to high standards in conducting public the order, acquire the delivery address. Pizza delivery polls. We encourage the reader to compare our methods to information is used by private investigators and by gov- the best practices articulated in 20 Questions A Journalist ernments to track individuals. In the marketing context, Should Ask About Poll Results, published by the National pizza delivery databases have been discussed as source Council on Public Polls. Furthermore, we go beyond these for phone numbers for wireless 411 databases. standards by, first, guaranteeing that we publish all the questions asked and responses received; and second, We asked Californians true or false (N = 341): “When I sharing our results so that others can inspect them. order a pizza to be delivered to my home, the pizza company is prohibited from selling my address and phone RESULTS AND DISCUSSION number to other companies, unless I give them explicit permission.” Fifty-four point seven percent incorrectly Newspaper and Magazine Subscriptions answered true and 5.8 percent said they didn’t know. Thirty-nine point five percent correctly answered false. Newspapers and magazines may sell their subscriber lists unless they promise not to in a privacy policy. No federal Donations to Charities or California sectoral law addresses subscription list sales. Searches of the DirectMag Listfinder, a search engine for Our next question asked about charitable donations. Many direct marketing databases, reveal that many prominent organizations that solicit charitable donations sell lists of newspapers and magazines sell their subscriber lists, members and donors. A search in the DirectMag Listfinder including the New York Times, and the Wall Street Jour- shows that even lists of members and donors to pro- nal, as well as California-based publications such as privacy causes are sold, including lists of Congressman Sacramento Magazine, Los Angeles Magazine, and Ron Paul’s supporters, former members of the Sovereign Oakland’s Diablo Magazine (Direct Magazine). Society, and supporters of NARAL Pro-Choice America. Many other California political candidates, museums, We asked Californians true or false (N = 309): “When I non-profit advocacy groups, and public radio and televi- subscribe to a newspaper or magazine, the publisher is sion stations appear in the database, offering lists of prohibited from selling my address and phone number to members and donors for sale.

248 Marketing and Public Policy Conference Proceedings / 2009 We asked Californians true or false (N = 339): “When I Often, these warranty cards solicit information from the make a donation to a charity, the charity is prohibited from consumer that is wholly unrelated to the purchase of the selling my address and phone number to other companies, product and registration of the warranty (such as income unless I give them explicit permission.” Forty-three point information, number of persons in the household, marital six percent thought they could not, and 13.9 percent said status, ages of children, interests, hobbies, ailments, etc.). they didn’t know. This means that 57.5 percent falsely This information is sold in bulk for direct marketing believe or do not know that their personal information is purposes. protected from sale when they make a donation to a charity. We asked Californians true or false (N = 365): “When I send in a product warranty card, the product manufacturer Turow et al. (2005), in their national sample of internet is prohibited from selling my address or phone number to users, found that 47 percent incorrectly answered true to other companies, unless I give them explicit permission.” the statement, “When I give money to charity, by law that Only 38.9 percent understand that this information can be charity cannot sell my name to another charity unless I sold without consent, while 50.3 percent falsely believe it give it permission.” Of the remaining respondents, 28 cannot, and 10.8 percent answered that they didn’t know. percent answered false and 25 percent didn’t know. What’s Your Phone Number? Sweepstakes Remember when Radio Shack would ask for your name In many sweepstakes contests, individuals are encour- and address, even for a cash purchase? The company aged to provide personal information for the remote stopped that practice in 2002, but in recent years, many chance of winning a prize. These sweepstakes are mar- other retailers have started asking customers for their keted as “free” opportunities to win prizes, but they come phone number. Often, no reason is given for this request; at the cost of the sale of the individual’s personal informa- the consumer may assume that providing the phone num- tion. Often, the individuals who appear on marketing lists ber is required or is related to fighting fraud. Retailers are of sweepstakes players are characterized as “impulsive” free to sell the information, unless they have promised not and become the targets for even more targeted messaging. to in a privacy policy. (Radio Shack is one of the few retailers that have privacy policies at the register.) We asked Californians true or false (N = 292): “When I enter a sweepstakes contest, the sweepstakes company is There are privacy risks associated with revealing a phone prohibited from selling my address or phone number to number. First, many people used to share a single phone other companies, unless I given them explicit permis- number, but the advent of wireless phones has made it sion.” Most Californians understood that sweepstakes more likely that a number belongs to a single individual. offers can result in the sale of personal information Second, because of local number portability, increasingly without consent. A majority – 54.7 percent – correctly consumers are retaining the same phone number for many answered false to a question asking whether sweepstakes years. These two developments have converted the phone companies can sell contact information without explicit number into a new form of persistent identifier, one that permission. A minority, 45.3 percent answered true incor- can track individuals over time. Third, because of a rectly or didn’t know the answer. loophole in the Do-Not-Call Telemarketing Registry, businesses can call individuals with whom they have a Product Warranty Cards business relationship. Thus, providing a phone number to a cashier facilitates telemarketing, even to those who have Many consumer products are accompanied by product opted out of telemarketing. Most individuals are unaware warranty cards, which request that the consumer register of this implication of sharing their phone number, as it is their warranty with the manufacturer. While warranty not widely publicized, and retailers in the U.S. also do not registration is not necessary for enjoying the protections provide a privacy policy at the point of purchase. of the warranty (keeping a copy of the receipt suffices for proving when the product was purchased), many consum- We asked Californians (N = 333): “When I give my phone ers complete registration cards. Completing the card may number to a store cashier, the store is prohibited from help to learn about product recalls, especially with baby selling my address or phone number to other companies, products and other important purchases. unless I give them explicit permission.” Fifty-six point

Marketing and Public Policy Conference Proceedings / 2009 249 nine percent thought the practice was prohibited and 4.2 five different American cities and found that “In all five of percent said they didn’t know. Only 38.9 percent correctly our comparisons, we wound up spending less money in a answered false. supermarket that doesn’t offer a card, in one case 29 percent less.” In addition, Turow et al. (2005) found that Product Rebates the public was concerned with supermarket purchase tracking. In a national survey of 1,500 internet-using Consumer protection concerns surrounding mail-in prod- adults, 66 percent disagreed or disagreed strongly with the uct rebates focus mainly on the problem of non-payment statement, “It’s OK with me if the supermarket I shop at of rebate rewards. We are all familiar with the hurdles keeps detailed records of my buying behavior.” involved in obtaining a product rebate – often complex, easy-to-violate rules result in disqualification. Often over- Nevertheless, club cards remain popular. In California, looked in the debate over mail-in rebates is that informa- substantive privacy regulation limits collection of infor- tion submitted on the forms, which like product warran- mation and sale of data collected through the club pro- ties, often include requests for irrelevant personal lifestyle grams; for example, under California Civil Code § 1749.64, data. This data is sold to third parties. supermarkets are prohibited from requesting social security numbers or driver’s license numbers in order to enroll We asked Californians true or false (N = 388): “When I in a loyalty club. Furthermore, State law specifies that “no complete a form for a rebate on a product, the product club card issuer may sell or share a cardholder’s name, manufacturer is prohibited from selling my address and address, telephone number, or other personal identifica- phone number to other companies, unless I give them tion information.” explicit permission.” A majority, 50.8 percent, thought the practice was prohibited, and an additional 12.1 percent We asked Californians true or false (N = 293): “When I answered that they didn’t know. Only 37.2 percent knew purchase groceries using a loyalty or club card, the gro- the correct answer: false. cery store is prohibited from selling my address and phone number to other companies, unless I give them explicit Catalog Sales permission.” Forty-nine point eight percent agreed that grocery stores could not sell personal information without Catalog companies have long sold personal information explicit permission. The question is somewhat problem- and data about purchases that customers have made. The atic, because California law prohibits all sale of club card DirectMag Listfinder returns over 5,000 results when one information, even with consent, and waivers of the prohi- searches for “catalog.” In addition, “cooperative data- bition are unenforceable. Therefore, a true answer isn’t bases” allow catalog merchants to share personal infor- exactly correct, but a false answer, which was chosen by mation among hundreds of companies. We asked Califor- 42.6 percent is less accurate, and 7.6 percent said they nians true or false (N = 308): “When I order something didn’t know. from a catalog, the catalog company is prohibited from selling my address and phone number to other companies, Turow et al. (2005), in their national survey, found that 36 unless I give them explicit permission.” Forty-eight point percent answered “true” to the statement, “My supermar- five percent falsely thought the practice was prohibited ket is allowed to sell other companies information about without express consent, and 3.7 percent said they didn’t what I buy.” Another 36 percent answered false, and 28 know. percent didn’t know.

Grocery Store “Loyalty” Club Cards Westin’s Fundamentalists, Pragmatists, and the Un- concerned Many grocery stores offer loyalty or club cards to track individuals’ purchases with the promise of special dis- In comparing our California-specific population to counts and coupons to customers. There is some debate Westin’s general population numbers (Equifax-Harris whether cards really offer discounts. A 2003 Wall Street 1995, 1996; IBM-Harris 1999; Kumaraguru and Cranor Journal study concluded that, “most likely, you are saving 2005), it is clear that Californians have even stronger no money at all [from supermarket shopping cards]. In privacy concerns; while Fundamentalists are slightly lower fact, if you are shopping at a store using a card, you may than Westin’s 2001 numbers (23% in CA compared to be spending more money than you would down the street 25% nationally), Pragmatists are over 10 points higher at a grocery store that doesn’t have a discount card” (74% in CA compared to 63% nationally), and the Uncon- (McLaughlin 2003). The Wall Street Journal study sur- cerned nine points lower (3% in CA compared to 12% veyed grocery stores that did and did not use club cards in nationally).

250 Marketing and Public Policy Conference Proceedings / 2009 Since our survey gauged Californians’ knowledge of Similarly, privacy Pragmatists may be providing answers privacy rights in nine different contexts, it offered the relative to their belief that existing laws and regulations opportunity to compare how often privacy Fundamental- fairly protect data, and that businesses are responsible in ists, Pragmatists, and the Unconcerned correctly identi- their stewardship of personal information. fied their privacy rights. We found that in almost every case, Privacy Fundamentalists answered correctly signifi- These questions should open one of Westin’s (2001) core cantly higher than Pragmatists and the Unconcerned (p = policy assumptions to more scrutiny – the idea that “In the .047). Privacy Fundamentalists, as a group, were more politics of privacy, the battle is for the hearts and minds of likely to be correct than Pragmatists and the concerned on the Privacy Pragmatists.” Turow (2003) concluded that eight of the nine questions. Only with respect to grocery Americans “. . . cannot really engage in the kinds of club card sales did Pragmatists and the unconcerned informed cost-benefit analyses that writers such as Alan answer correctly more frequently than Fundamentalists. Westin suggest take place when consumers ‘pragmati- cally’ give up information about themselves.” In some cases, the Unconcerned answered correctly more than the Pragmatists, but because our sample included CONCLUSION only 30 of the individuals in our 991-person sample, this finding is less significant. Prevailing studies of public opinion on privacy focuses on how concerned individuals are about losses to privacy, In his review of Westin’s work, Gandy (2005) found that and classify the public into high, mid, and low-level the privacy segmentation was highly dependent upon concern segments. These studies have limited use in knowledge about privacy problems: “I discovered that the policy analysis. They make obvious conclusions: privacy extent to which people had read or heard about the concern is high, but different people feel differently about “potential use or misuse of computerized information it. These conclusions have lended support to the idea that about consumers” was a powerful explanatory factor. The the majority of Americans is comfortable with opt-out more they had heard or read, the more they were con- rules, and industry self-regulation instead of more en- cerned about threats to their privacy, the more concerned forceable privacy laws. they were about the sale of personal information by the list industry. And consistent with a view that sees mediated Our study builds on the work of Turow et al. (2003, 2005) experience as a source of social opinion, the more you to show that consumers are generally unaware of whether heard or read, the less you trusted organizations that privacy law protects them in commercial transactions. collected and used information about consumers.” Simi- More importantly, segments of the population who have larly, Turow et al. (2005) found using regression analysis mid or low level concern and are happy with the status quo that having more years of education correlated with an- are more likely to be ignorant of the lack of privacy law swering questions about online and offline privacy rules than high-level concern consumers, who support the correctly. passage of new privacy laws.

Privacy Fundamentalists’ correct answers may be driven Of course, it is intuitively appealing to frame a public by a general pessimism or skepticism concerning privacy policy approach as serving the interests of a balanced, protections, rather than knowledge of California law. reasonable group – the mid-level concern Privacy Prag- Under Westin’s (2001) segmentation, a Privacy Funda- matists. But this survey shows that this group is either mentalist agrees or strongly agrees with the statement, misinformed or overly optimistic about rules concerning “Consumers have lost all control over how personal use of their data in many ordinary, offline transactions. If information is collected and used by companies.” This one’s membership in a segment is explained by knowl- skepticism could have caused Fundamentalists, as a group, edge and experience, as Gandy argues, then Westin ap- the reject the idea that there are legal controls to govern proach does not serve privacy Pragmatists. Instead it consumer data, especially in light of how many funda- manipulates them, by relying upon their ignorance of mentalists incorrectly answered our grocery club card rules and practices to support a policy outcome that they question, where California law offers strong protections. would likely oppose, if better informed.

Marketing and Public Policy Conference Proceedings / 2009 251 REFERENCES ticle/0,,SB1043006872628231744,00.html]. Kumaraguru, Ponnurangam and Lorrie Faith Cranor Acquisti, Alessandro and Jens Grossklags (2005), “Pri- (2005), “Privacy Indexes: A Survey of Westin’s vacy and Rationality in Individual Decision Mak- Studies,” [available at http://reports–archive.adm. ing,” IEEE Security and Privacy, (January/Febru- cs.cmu.edu/anon/isri2005/CMU-ISRI-05-138.pdf]. ary), 24–30. Regan, Priscilla M. (2005), Legislating Privacy, Technol- Best, Samuel J., Brian S. Krueger, and Jeffrey Ladewig ogy Social Values, and Public Policy. Chapel Hill. (2006), “Trends: Privacy in the Information Age,” Sheehan, Kim Bartel (2004), “How Public Opinion Polls Public Opinion Quarterly, 70 (Fall), 375–401. Define and Circumscribe Online Privacy,” First Duhigg, Charles (2007), “Bilking the Elderly, With a Monday, [available at http://www.firstmonday.org/ Corporate Assist,” New York Times, (May 20), [avail- issues/issue9_7/sheehan/index.html]. able at http://www.nytimes.com/2007/05/20/busi- Story, Louise and Brad Stone (2007), “Facebook Retreats ness/20tele.html]. on Online Tracking,” New York Times, (November Equifax-Harris, Mid Decade Consumer Privacy Survey 30) [available at http://www.nytimes.com/2007/11/ (1995). 30/technology/30face.html]. ______, Consumer Privacy Survey (1996). Trager, Louis (2005), “Cal.’s Unique, Broad New Info- IBM-Harris, Multi-National Consumer Privacy Study Sharing Law Largely Under the Radar, Says State (1999). Privacy Chief,” Washington Internet Daily, (October Listfinder, Direct Magazine, [available at http:// 7). directmag.com/resourcecenter/listfinder/]. Turow, Joseph (2003), “Americans & Online Privacy, Gandy, Oscar (2003), “Public Opinion Surveys and the The System is Broken,” Annenberg Public Policy Formation of Privacy Policy,” Journal of Social Center, (June), [available at, http://www. Issues, 59, 283–99. annenbergpublicpolicycenter.org/Downloads/ ______(2005), “The Role of Theory in the Policy Information_And_Society/20030701_America_ Process, A Response to Professor Westin,” in To- and_Online_Privacy/20030701_online_privacy_ ward an Information Bill of Rights and Responsibili- report.pdf]. ties, C. Firestone and J. Schement, eds. Aspen. ______, Lauren Feldman, and Kimberly Meltzer Hoofnagle, Chris Jay and Jennifer King (2007), “Con- (2005), “Open to Exploitation: American Shoppers sumer Information Sharing: Where the Sun Still Online and Offline,” Annenberg Public Policy Don’t Shine,” Samuelson Law, Technology & Public Center (June), [available at http://www. Policy Clinic Report, [available at http:// annenbergpublicpolicycenter.org/NewsDetails. www.law.berkeley.edu/samuelsonclinic/privacy/ aspx?myId=31]. 219]. Westin, Alan K. (2001), “Opinion Surveys: What Con- McLaughlin, Katy (2003), “The Discount Grocery Cards sumers Have to Say About Information Privacy,” That Don’t Save You Money,” Wall Street Journal, Before the House Commerce Subcommittee on Com- (January 21), [available at http://online.wsj.com/ar- merce, Trade, and Consumer Protection, (May 8).

For further information contact: Chris Jay Hoofnagle University of California, Berkeley 396 Simon Hall Berkeley, CA 94720 Phone: 510.643.0213 Fax: 510.643.4625 E-Mail: [email protected]

252 Marketing and Public Policy Conference Proceedings / 2009 CONSUMER PERCEPTIONS OF ONLINE PRIVACY: BEHAVIORAL TARGETING AND THE VALUE OF PERSONAL INFORMATION

Poonam Kumar, Saginaw Valley State University, University Center Eloise Coupey, Virginia Tech, Blacksburg Mary Peck, Virginia Tech, Blacksburg

SUMMARY reports that the number of Internet searches doubled each month since the summer of 2006, to 14.6 billion searches Behaviorally targeted advertising generates attention and in January 2007. Major companies like Yahoo, Google, controversy among marketers, consumers, and policy Microsoft, AOL and MySpace record at least 336 billion makers. Behavioral targeting is the practice of delivering transmission events in a month. focused content to an online consumer, in which the content is tailored, based on consumer characteristics, Companies compete to get access to users’ behavioral such as location and other demographics, as well as data, because the costs of being shut out are high. Desir- consumer interests, as indicated by previous, tracked, able aggregate-level consumer responses to companies online behaviors, including site visits and web searches. who have the technological and data-rich capabilities to Advertisers tout the benefits of behavioral targeting as the deliver a tailored information environment may include ability to gain economies in advertising, for companies increased site traffic; heightened site (and brand) loyalty; and for consumers; by improving the match between the increased revenue; and augmented market share. For ad content and consumer response metrics, such as aware- strategic functioning, establishing a behavioral targeting ness and click-through. strength may deter potential market entrants; confer a first mover advantage; and result in proprietary technology. The forces that motivate the adoption of behavioral targeting models for firms are very different from those that may The rapid growth in use of behavioral data by online mediate or moderate their acceptance – whether implicitly marketing firms – which eMarketer forecasts will top $4.4 or explicitly – in the consumer sector. Purported benefits billion by the end of 2012 – has raised concerns about of behavioral targeting for consumers include subsidized consumer privacy and online data security (Dobosz, Green, (and even customized) web content, enhanced ad rel- and Sisler 2006; Federal Trade Commission [FTC] 2000, evancy, and browsing efficiencies, each of which may 2007; Milne 2000; Miyazaki 2008; Sheehan 2002). De- influence behavioral responses (e.g., click-through, pur- mand for access to behavioral data increases the value of chase, loyalty) and psychological responses (e.g., confi- third-party advertising networks to companies without dence, satisfaction, trust). internal databases of across-session and across-site behavioral information. The introduction of these compa- The present research examines consumers’ willingness to nies into the cost-benefit equation potentially decreases trade off personal information, in exchange for the ben- control over the acquisition, storage, and dissemination of efits marketers believe behavioral targeting makes avail- consumer information. able. In particular, we examine the role of consumer knowledge types, including subjective and objective Researchers have examined the theoretical bases and knowledge, on perceptions of online privacy, and the substantive impact of Internet technologies on consumer effect of these perceptions on Internet-related behaviors. perceptions of privacy (Culnan and Milne 2001), and their privacy-related attitudes and behaviors (e.g., Norberg and Behavioral Targeting: The Cost-Benefit Back-Story Horne 2007). Interestingly, however, while consumers express concerns about privacy, self-protective behaviors With increasing frequency, companies view behavioral remain low (Milne, Rohm, and Bahl 2004). A holistic targeting as an important element of a strong competitive review of privacy research indicates that: (1) consumers strategy. The cost-benefit logic for companies is straight- express concerns about privacy; and (2) consumers view forward; with the behavioral data of users, companies can privacy as an important concern in online environments; target ads to specific consumers based on their interests but (3) consumers do not protect fully their privacy online, and preferences. As more people spend more time on the given available methods; and furthermore, (4) consumers Internet, companies can collect more data. Comscore expose personal data to companies who may be tracking it.

Marketing and Public Policy Conference Proceedings / 2009 253 Conceptual Background clarative knowledge about aspects of privacy protection. Data was also obtained about subjects’ procedural knowl- We develop a cost/benefit model of consumer participa- edge, as implied by their responses to questions about the tion in online exchanges. Further, we propose that con- frequency and manner in which with which they per- sumer willingness to share personal information, as in formed privacy-protective tasks. allowing behavioral tracking, is influenced by the consumer’s knowledge of the online environment, in- To determine whether consumers are willing to trade off cluding the policies and technologies employed by the privacy for an online benefit, we asked a series of ques- companies involved, and trust in those companies, as well tions constructed to isolate the tradeoff without biasing as the consumer’s own ability – perceived and actual – to the responses. In sum, the sample was highly resistant to protect desired levels of online privacy. a variety of implementations of behavioral targeting, including those ranging from active provision of personal People exhibit at least two different types of knowledge: information to more passive tracking of browsing behav- objective knowledge and subjective knowledge. In the iors. Additional questions tapped the role of situational marketing literature, objective knowledge has been con- factors, such as the availability of online access, on the ceptualized as the accurate information a consumer pos- willingness to expose personal information online. For sesses about a domain, including product knowledge and instance, we asked whether respondents would be willing procedural knowledge, while subjective knowledge has to accept sponsored access – receiving ads on their screens – been characterized as consumers’ perceptions of what in lieu of a fee. As predicted, subjects were willing to make they think they know (Brucks 1985; Spreng, Divine, and tradeoffs, accepting targeted advertising, with the under- Page 2001). When subjective knowledge exceeds objec- lying assumption of behavioral tracking, in exchange for tive knowledge, overconfidence is present (Arkes, Dawes, free access. and Christensen 1986). Overconfidence may negatively affect the quality of decision making (cf., Coupey and Twelve depth interviews were conducted to examine Narayanan 1996). For instance, in online environments, preferences and perceptions regarding online privacy and consumers may expose personal information with inad- behavioral targeting. Participants, recruited by flyer and equate privacy safeguards. word-of-mouth within a campus community, comprised a broad range of ages (20’s through 40’s) and occupations, Method and Results and an even gender split. Several themes emerged: (1) relative comfort with the privacy of personal information We conducted a two-part program of research to examine online, particularly with companies that they elected to perceptions of online privacy, and the impact of objective visit when browsing; (2) higher concern for threat to and subjective knowledge types on perceptions and be- privacy posed by hackers, rather than by companies; (3) a haviors. With an online survey, we assessed aggregate- tiered structure of concern for information, in which the level indicants of objective and subjective knowledge, in security and privacy of financial information is para- conjunction with expressed willingness to trade off online mount; (4) low levels of knowledge about the nature and privacy in exchange for benefits, such as wireless access intent of behavioral targeting; and (5) an implicit, or free email. To clarify insights from the survey, we reputational heuristic for assessing the trustworthiness of conducted in-depth interviews to examine the value placed a company as a function of how much it has to lose, in on different types of personal information exchanged terms of consumer good will/patronage. between companies and consumers in online environments. An online survey of 495 students assessed basic Conclusions demographics and Internet-related activity, as well as knowledge of technologies and methods related to infor- This research postulates a cost-benefit process by which mation collection, privacy protection, and online adver- consumers evaluate willingness to provide personal infor- tising. mation in exchange for online resources. The results suggest that tradeoff quality is influenced by knowledge For subjective knowledge, respondents rated their pri- types. The research can be used to develop proactive, vacy protection knowledge with four five-level Likert pragmatic policy applications for information manage- items. Objective knowledge was assessed with a ten item, ment by consumers and marketers, and policy makers. True-False questionnaire that examined participants’ de-

254 Marketing and Public Policy Conference Proceedings / 2009 For further information contact: Eloise Coupey Virginia Tech 2016 Pamplin College of Business Blacksburg, VA 24060 Phone: 540.231.4724 Fax: 540.231.3076 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 255 POLICY IMPLICATIONS OF RECIPROCITY AND INDEPENDENT JUDGMENT

Jeff Langenderfer, Meredith College, Raleigh Marlys J. Mason, Oklahoma State University, Stillwater Xiang Fang, Oklahoma State University, Stillwater

SUMMARY gratitude is the most significant factor in gift induced compliance (Gouldner 1960; Tsang 2007). Substantial anecdotal and experimental evidence exists which suggests that gift-giving induces reciprocity and Policy Implications compliance in receivers. Yet the impact that such incentives have on the ability of individuals to make unbiased Though it is well-known that the receipt of a gift may judgments is not well understood. This paper is an initial induce the receiver to repay the favor, there are many effort to investigate the impact of reciprocity on subse- instances in which public officials or other professionals quent judgments and its purpose is threefold. First, we are expected to exercise unbiased independent judgment review the evidence that gifts may bias judgment due to even thought they have gotten gifts or other payments feelings of indebtedness, liking, and gratitude. Second, from those about whom judgments must be rendered. we discuss the implications of gift-induced bias for policy Some organizations have rules intended to minimize the development and implementation. Finally, we outline our effects of gifts on such theoretically unbiased judgment, current investigation into testing the nature and extent of by limiting gifts to small amounts or requiring that gift the impact of favors and gifts on independent judgment. recipients refrain from evaluating those who have provided remuneration or other benefits. It is still common, in Background many instances, for gift recipients to be expected to put aside their feelings of indebtedness, liking or gratitude There is substantial anecdotal evidence that gift giving engendered by a gift and remain independent in their generates in the recipient a desire to repay the favor. judgments: Legislators routinely vote on matters that Charitable organizations find that including small gifts in affect campaign contributors. Judges often rule on cases solicitation requests improves responses (Cialdini 1984) in which campaign contributors are parties to the action. and free samples have long been used as an effective Physicians make treatment recommendations to their pa- marketing technique both to induce trial and engender tients having received gifts or substantial payments from indebtedness (Packard 1957). The effect of gift giving on pharmaceutical firms. In some cases, physicians who compliance has also been demonstrated experimentally. serve on FDA review panels may be called upon to Regan (1971) showed that an unsolicited gift increases the evaluate the safety and efficacy of drugs whose makers degree of compliance with a subsequent request and have made payments to those same physicians. Account- several more recent studies have made similar demonstra- ing firms who function as highly paid consultants to tions (Goei et al. 2007; Tsang 2007; Algoe, Haidt, and corporations are asked to audit and review the financial Gable 2008). disclosure practices of those same firms. Can those judgments truly be unbiased? Notwithstanding the substantial evidence that favors induce compliance with subsequent requests, it is not nec- This paper reports on two studies designed to test the essarily clear what factor or factors drive this result. Some nature and extent of the impact of favors and gifts on researchers, including Regan (1971), posit that feelings of independent judgment. A replication of Regan (1971) is indebtedness are the primary drivers of the propensity to conducted to examine to persistence of gift-induced com- repay a gift (Greenburg and Frisch 1972). Other research- pliance. A second study is assessing the extent to which ers suggest that liking is a significant mediator, playing an underlying judgment is affected, and whether an induce- important role in willingness to comply (Burger et al. ment to remain unbiased affects such judgment. 2001). Still other researchers argue that the emotion of

256 Marketing and Public Policy Conference Proceedings / 2009 REFERENCES Gouldner, Alan (1960), “The Norm of Reciprocity: A Preliminary Statement,” American Sociological Re- Algoe, Sarah B., Jonathan Haidt, and Shelley L. Gable view, 25, 161–78. (2008), “Beyond Reciprocity: Gratitude and Rela- Greenberg, Martin S. and David M. Frisch (1972), “Effect tionships in Everyday Life, Emotion, 8 (3), 425–29. of Intentionality on Willingness to Reciprocate a Burger, Jerry M., Shelley Soroka, Katrina Gonzago, Emily Favor,” Journal of Experimental Social Psychology, Murphy, and Emily Somervell (2001), “The Effect of 8, 99–111. Fleeting Attraction on Compliance to Requests,” Packard, Vance (1957), The Hidden Persuaders. New Personality and Social Psychology Bulletin, 27, 1578– York: D. McKay Company. 86. Regan, Dennis T. (1971), Effects of Favor and Liking on Cialdini, Robert B. (1984), Influence: The Psychology of Compliance, Journal of Experimental Social Psy- Persuasion. New York: William Morrow & Com- chology, 67, 599–609. pany. Tsang, Jo-Ann (2007), “Gratitude for Small and Large Goei, Ryan, Anthony Roberto, Gary Meyer, Kellie Carlyle Favors: A Behavioral Test,” Journal of Positive Psy- (2007), “The Effects of Favor and Apology on Com- chology, 2, 157–67. pliance,” Communication Research, 34 (6), 575–95.

For further information contact: Jeff Langenderfer Meredith College 3800 Hillsborough Street Raleigh, NC 27607 Phone: 919.760.8673 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 257 CONSUMER CONFUSION IN THE MARKETPLACE

Paula Fitzgerald Bone, West Virginia University, Morgantown Jeremy Kees, Villanova University Karen Russo France, West Virginia University, Morgantown John Kozup, Villanova University

SUMMARY health claim (i.e., “May Reduce the Risk of Certain Cancers”) remained constant across all conditions. Three Background levels of scientific certainty were indicated by a checked box next to word descriptors of “strong,” “moderate and Consumers live in a world full of information – informa- promising,” or “weak.” A mock press release attributed to tion offered by firms, advocacy organizations, scientists, the American Medical Association (AMA) was used as a government agencies and even from uninformed and/or credible external source and was varied to be supporting highly biased agenda-driven groups. Most of this infor- or not supporting of the health claim. Finally, the two mation is designed to help consumers make decisions for different products were used in the study an unbranded themselves regarding important aspects of consumption. selenium supplement and an unbranded packaged tuna Yet, the information provided to consumers is not straight- fish product. forward and, in fact, often conflicts with other pieces of information. Three hundred and sixty-two adults drawn from an online panel participated in this study (45% male; median age In this paper, we explore how consumers react when bracket was 26–35). Respondents were exposed to the confronted with multiple sources of conflicting informa- mock label and the mock press release simultaneously and tion. We offer a model which integrates extant treatment continuously as respondents answered questions regard- of information ambiguity and conflict and further devel- ing behavioral intentions, search, product quality, product ops the construct of consumer confusion to examine efficacy in preventing cancer, cognitive and emotional confusion’s influence on search, intentions, product qual- confusion, and trust in the industry. ity and trust. We then present results from a nation-wide experimental study which focuses specifically on cogni- Results and Discussion tive and emotional confusion. Our first research question asks whether conflicting infor- Herein, we address very specific questions: (1) Does mation leads to cognitive and/or emotional confusion. conflicting information lead to cognitive and/or emo- Results suggest that emotional confusion for both prod- tional confusion? (2) Does conflicting information im- ucts is highest when the press release claims “weak” pact search, behavioral intentions, quality and efficacy evidence and the disclaimer states that the evidence is perceptions, and firm trust? And (3) does confusion (cog- strong (i.e., the two conflict with each other). Addition- nitive and emotional) mediate the impact of conflicting ally, emotional confusion is lowest when the press release information on search, behavioral intentions, quality and and the disclaimer state that the evidence for the health efficacy perceptions, and firm trust? claim is strong (i.e., the two support each other). Overall, there is evidence that conflicting information affects both Experiment cognitive and emotional confusion. The impact on emotional confusion is stronger (R2 = 0.14) than that on A 3 (level of disclaimer: strong scientific support, moder- cognitive confusion (R2 = 0.04). ate and promising scientific support, or weak scientific support) by 2 (external source: supporting or not support- The second research question examines the impact of ing health claim) by 2 (product category: food or supple- conflicting information on several outcome variables – ment) factorial design was used to explore the above information search, behavioral intentions, product per- research questions. Study participants were shown full- ceptions (quality and efficacy) and firm trust. Wilks’ color mock labels, based on actual product labels in the Lambda indicated main effects for product (F = 5.38, p < marketplace featuring a health claim and disclaimer. The 0.01), the press release (F = 21.38, p < 0.01) and an

258 Marketing and Public Policy Conference Proceedings / 2009 interaction between the product and the press release This study was an initial step toward better understanding which meets a more liberal standard of statistical signifi- consumer response to conflicting information and how cance (F = 1.86, p < 0.09). The interaction between the subsequent confusion affects information processing. The product and the press release significantly affected inten- results indicate that confusion drives product perceptions tions to seek more information (F = 8.15, p < 0.01) as well and search intentions. Additionally, the strength of scien- as intentions to consume the product (F = 5.63, p = 0.02) tific evidence qualifications as currently mandated by such as these intentions were highest for tuna when the FDA leads to confusion on the part of consumers. Whether press release suggested “weak” evidence. When the press that confusion leads to poor decisions or increased vigi- release provided “strong” evidence, intentions to search lance on the part of consumers is the subject of further and intentions to consume were equivalent for the two inquiry. products.

For further information contact: Jeremy Kees Villanova University Bartley 3006 Villanova, PA 19085 Phone: 610.519.6145 Fax: 610.519.5043 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 259 THE MERE CATEGORIZATION EFFECT ON PERCEIVED HEALTHINESS AND CONSUMPTION OF FOOD ITEMS

Caglar Irmak, University of South Carolina, Columbia Beth Vallen, Loyola College, Baltimore

SUMMARY nutritional value for identical food items. For instance, imagine that you are ordering dinner from a restaurant In the absence of objective nutrition information, con- menu that presents the items under different category sumers often rely on cues, such as product name, to titles (e.g., pastas and salads). Our results suggest that you provide a basis for evaluating the healthiness of foods. Yet will perceive a food item that is presented in a relatively making the distinction between healthy and unhealthy unhealthy food category (e.g., pasta) to be less nutritious items based on such limited information might be difficult than an identical food item that is presented in a healthier for some consumers. For example, smoothies bear a food category (e.g., salad) based merely on the product striking resemblance to milkshakes while flavored waters category implied by the name of the item. Moreover, we are often equivalent to soda in terms of calorie and sugar show that these effects are more likely to be demonstrated content (Buia 2003; “Smoothie Operator” 2003). While by dieters due to their expertise in the domain of food- some items do offer nutritional benefits over less healthy related decisions, which leads to heavy reliance on cat- alternatives, the value of others might be illusory. Thus, to egory-based heuristics to form inferences about products make the right food-related decisions, consumers may (Sujan 1985; Sujan and Dekleva 1987). In four studies, we need to seek more specific nutritional information than the demonstrate that product category information influences product name. dieters’ perceptions of the nutritional value of food products, such that items that belong to relatively unhealthy Oftentimes, however, we do not have access to objective categories are perceived to be less nutritious than identical nutritional information (e.g., calories, fat grams, sodium food items that belong to healthier categories. These content, etc.) or all of the information that would be perceptions of nutritional value, in turn, impact actual needed to accurately infer the nutritional value of food consumption behavior. items (e.g., ingredients, side dishes, cooking methods, etc.). As a result, we are likely to rely on signals that In the first two studies, participants were presented with a provide us with a basis for evaluation. For example, dish containing identical ingredients, but with different consumers perceive that low cholesterol foods are also product category assignment communicated by the item low in fat (Andrews, Netemeyer, and Burton 1998). name (i.e., pasta vs. salad or sausage links vs. veggie Moreover, foods that carry low fat claims are assumed to links). They then evaluated the item’s nutritional value have larger serving sizes relative to their full fat counter- and responded to questions assessing dieting behavior. As parts (Wansink and Chandon 2006). Such health halos expected, while non-dieters showed no difference in their can also be generated by brand name; recent research nutrition-related perceptions of the items, dieters per- demonstrates that when a fast-food restaurant claims to ceived items assigned to the unhealthy categories (i.e., serve healthy food (e.g., Subway), consumers are more pasta and sausage links) to be significantly less healthy likely to underestimate the caloric content of main dishes than the items assigned to healthier categories (i.e., salad and to choose higher-calorie side dishes, drinks, or des- and veggie links), despite the fact that the actual items serts compared to when no such claims are made (e.g., were identical. McDonald’s; Chandon and Wansink 2007). In the third study we sought to investigate the proposed Prior research has explored the impact of categorizing categorization process underlying the effects observed in different food items; consumers have been shown to the prior studies. To do so, we manipulated the processing classify foods as “good” (e.g., raisins) or “bad” (e.g., approach of study participants. After providing partici- potato chips) and to assume that calorie equivalent por- pants with the same pasta-salad pair used in Study 1, half tions of the good foods promote less weight gain than the of the participants engaged in an effortful, piecemeal bad foods (Oakes 2005; Oakes and Slotterback 2007). In process by estimating the healthiness of each ingredient of this work, we posit that the category implied by the name the food item before evaluating its overall healthiness. of a food item can similarly impact perceptions of the The remaining participants engaged in a holistic, catego-

260 Marketing and Public Policy Conference Proceedings / 2009 rization process by estimating the healthiness of each demonstrate any of these inferences nor was their con- ingredient after they evaluated the overall healthiness of sumption influenced by product category. the food item. In line with the categorization explanation, the results show that when individuals first estimated the Together, these findings demonstrate that perceptions of healthiness of each ingredient (i.e., piecemeal process), the nutritional value of food items can differ depending on respondents perceived no significant difference between the food category in which they are presented. For dieters, the nutritional value of the two items, irrespective of food items that belong to relatively unhealthy categories dieting behavior. On the other hand, dieters – but not non- are assumed to be less nutritious, while those items dieters – considered the pasta to be less healthy than the deemed nutritious due to their association with healthier salad when they evaluated the nutritional value of the item categories are liked more and are more likely to be before its ingredients (i.e., categorization process). purchased and consumed. Non-dieters are less likely to rely on category-based heuristics. Ironically, this suggests In our final study, we explored the impact of food category that the knowledge that is acquired in order to make assignment on various product inferences as well as informed decisions may accentuate the impact of health consumption quantity. We provided participants of this halos created by food categories. study with a sample of an actual confectionary product identified as either “Candy Chews” or “Fruit Chews” and These findings contribute to the emerging body of litera- measured perceptions of nutritional value, attitude to- ture that examines cues that consumers utilize to infer ward, and purchase likelihood of the item, as well as the unknown information about food items (e.g., Chandon amount of the product consumed. As in prior studies, and Wansink 2007). Moreover, they have implications in when the item was associated with the less healthy cat- terms of the FDA regulations surrounding statements of egory (i.e., “Candy Chews”), dieters perceived the item to identity on food labels. While the FDA states that com- be lower in nutritional value compared to when it was mon or usual food names should be listed on product associated with a healthier category. Moreover, they ex- packaging in a manner that is not misleading (U.S. FDA pressed more positive attitudes toward and showed stron- 2008), specifications for the naming of potentially am- ger intentions to purchase the item when it was associated biguous food categories is limited to very few products with the healthy (vs. unhealthy) category (i.e., “Fruit (e.g., fruit juice vs. cocktail). Our results suggest that Chews”). Importantly, these inferences translated into inferences based on food name and implied category behavior; mediation analysis demonstrated that dieters extend to many different food types and, hence, have the consumed greater quantities of the item associated with potential to greatly impact consumers’ evaluation of food the relatively healthy category because they believed it to products and food-related decisions. References are avail- be more nutritious. Non-dieters, on the other hand, did not able upon request.

For further information contact: Beth Vallen Loyola College 4501 N. Charles Street Baltimore, MD 21224 Phone: 410.617.5466 Fax: 410.617.2117 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 261 ARE PARENTS VULNERABLE, SUSCEPTIBLE, OR SAVVY TO ADVERTISING STRATEGY? AN ANALYSIS OF EMOTION-BASED APPEALS IN FOOD ADS

Deirdre T. Guion, North Carolina Central University, Durham Julie V. Stanton, Pennsylvania State University, Media

SUMMARY consumer may be linked to his/her emotional involvement with a product. Their lack of sophistication – or “mental The battle for the mind of the consumer is increasingly incompetency” – may arise from situations, including difficult to win given the array of products, brand and time pressures, lack of knowledge of a particular topic and media through which to communicate. To break through insecurities about whether products they use will lead to the clutter, advertisers often utilize appeals based on eventual injury. emotion, whether fear, happiness or humor. Using negative emotional cues is often expected to make consumers In this study, the focus is on parents of small children and uncomfortable enough to want the “solution” offered in the strategies used by advertisers to influence parents’ the advertisement. choices of food for their children. Our argument is that the importance of making the right choices has been elevated In their research into advertisers’ own impressions of the by an increasing focus on health and food in the United use of negative appeals, Cotte and Ritchie (2005) identi- States and parents of small children are largely inexperi- fied three consumer types and the advertisers’ theories on enced with making what they perceive as the “right” how they react to negative appeals. “Desensitized” con- decision. These consumers would better fit with Cotte and sumers may only be reached by dramatic messages, and Ritchie’s “tribal” consumers – wanting to be the right kind negative emotional appeals have the potential to both grab of parent and wracked by guilt if they do not make the right attention and raise demand for the solution. “Sophisti- choices – than with their “sophisticated” consumer, at cated” consumers understand advertising strategy and are least with respect to their children’s nutrition. The same a risky group to target with emotional appeals. It follows, consumers may be sophisticated about automobile choices. that less sophisticated consumers would be less risky targets for negative appeals. “Tribal” consumers respond Our study utilizes a content analysis of food advertise- to ad messages based on a need to be part of a larger group, ments in Parenting and Parents magazines, the two most and negative appeals are considered to be particularly highly circulated in their category, from 1999–2008. effective with these consumers. Consumers who can be Emotional cues, linkages to regulated health claims and made to feel guilty that they do not fit with the norm may other strategies are assessed. To complement the content be more easily persuaded into product consumption. analysis, we interviewed parents of young children ranging in age from 8 weeks to 11 years old. While these distinctions are useful mechanisms to advertisers, they mask the greater heterogeneity of consumers, Preliminary analysis of 204 food ads from Parenting particularly since they do not capture the breath, depth and magazine shows that 38 of the ads included an emotional context of the vulnerability that consumers may feel. The appeal, and can be separated into those about infants (18 same consumer may reflect “sophisticated” behavior in ads or 47.3%), toddlers (3 ads) and small children (17 some product categories, for instance, while exhibiting ads). References to the emotional topic of breastfeeding v. more “tribal” tendencies in others. formula were often included as well, elevating the potential that parents perceive this decision as crucial to their The framework developed by Morgan et al. (1995) to baby’s well-being. The majority (86%) of the parents identify vulnerable populations offers a method for cap- interviewed also thought that the ads for baby formula turing such situational vulnerability. Among their groups were the most compelling appeals especially to expectant are consumers whose level of sophistication is in doubt, and new parents. Given the strong tribal sentiments leading to a naivety in product choices. The authors expressed by parents of infants, these cues may be more describe the group using terms such as “gullible” and impactful on them than on the FDA’s “reasonable” suggest that the level of sophistication demonstrated by a consumer.

262 Marketing and Public Policy Conference Proceedings / 2009 Our sample also demonstrates that the strength of the On preliminary evidence, it appears that marketing strat- emotional appeal varies by product (e.g., formula, meal, egy may indeed be taking advantage of a certain type of snack); whether the product is meant for adult, family or tribal consumer’s lack of sophistication with certain as- infant consumption; and other positioning cues (e.g., pects of food choices and, that appropriate policy to health, convenience). Preliminary interview data sug- protect those consumers may be needed. Additionally, the gests that the appeals of ads targeted to parents of toddlers editorial focus of the individual magazines suggests that and small children were the least compelling in terms of the book’s credibility may be considerably higher than potential vulnerability. As our sample expands to include that of more general interest magazines on matters of child Parents magazine and more formal content analysis and welfare. This may result in an unwarranted credibility of interviews are performed, we expect to examine these the magazine’s ad claims on such matters as well and strategies from multiple angles including changes over suggests that a policy designed to protect consumers time and brand-specific patterns. would need to take such context into account.

REFERENCES Morgan, Fred W., Drue K. Schuler, and Jeffrey J. Stoltman (1995), “A Framework for Examining the Legal Cotte, June and Robin Ritchie (2005), “Advertisers’ Theo- Status of Vulnerable Consumers,” Journal of Public ries of Consumers: Why Use Negative Emotions to Policy and Marketing, 14 (2), 267–77. Sell?” Advances in Consumer Research, 32, 24–31.

For further information contact: Deirdre T. Guion North Carolina Central University 1801 Fayetteville Street Durham, NC 27707 Phone: 919.530.7380 Fax: 919.530.6163 E-Mail: [email protected]

Marketing and Public Policy Conference Proceedings / 2009 263 USING THE THEORY OF PLANNED BEHAVIOR AS A FRAMEWORK FOR UNDERSTANDING THE ORGANIC FOOD CONSUMER: AN EXPLORATORY STUDY

Sally A. Williams, Texas State University, San Marcos Karen H. Smith, Texas State University, San Marcos

ABSTRACT qualitative research. By collecting and analyzing consumers’ open-ended responses to questions grounded Using depth interviews, the relationships between con- in theory, we hoped to obtain richer data than can be sumption of organic foods and attitudes, moral norms, provided by quantitative methodologies. past behavior, self-identity, and values and lifestyle were examined within an expanded Theory of Planned Behav- LITERATURE REVIEW AND RESEARCH ior framework. The qualitative approach uncovered sub- QUESTIONS stantial complexity in the core concepts “health,” “quality,” and “price,” with implications for policy makers. The Theory of Planned Behavior (Ajzen 1991) is a subjective expected utility model that has been used to INTRODUCTION predict volitional behavior across a great variety of contexts (see Armitage and Conner 2001), including food choice Between 1995 and 2006, sales of organic food rose by 17 (e.g., Armitage and Conner 1999; Robinson and Smith percent every year (Dimitri and Oberholtzer 2008). World 2002; Sparks and Shepherd 1992). According to this retail value rose from $10.3 billion in 2002 to $21.9 billion model, a consumer’s likelihood to engage in a particular in 2007, while the market size in the U.S. increased from behavior is dependent on the intention to perform the $4.6 billion to $10.2 billion (Global Market Information behavior. The intention construct is meant to encompass Database 2008). Although the recent economic downturn all elements of motivation that would lead a person to do has reduced sales in the last year (Springen 2008), the something. The basic model contends that behavioral industry is poised for significant growth in the future as intention is driven by three factors: the consumer’s attitude more consumers want to eat healthy and are concerned toward the behavior, subjective norms, and perceived about the safety of non-organic food (Whitman 2000). behavioral control. However, organics comprise only 2.8 percent of total food and beverage sales in the U.S. (Martin and Severson The consumer’s attitude toward the behavior is based on 2008). Thus, despite the fact that many consumers express the consumer’s salient beliefs about the behavior. In the concerns about non-organic food, few are buying organic. context of organic food, research has found advantages to To help explain this inconsistency, Hughner, McDonagh, include being healthier and better for the environment and Prothero et al. (2007) have called for qualitative research disadvantages to include being more expensive and less on the values, attitudes, and lifestyles of consumers of available (Hughner et al. 2007). In keeping with the organic foods. Therefore, the purpose of this study is to rational decision-making orientation of the TPB, attitude conduct qualitative depth interviews with consumers of is viewed as a cognitive rather than an affective construct organic food in order to develop a greater understanding (Dean, Arvola, Vassallo et al. 2006). Given that prior of consumer perspectives on this important trend. research has found that the attitude construct consists of separate cognitive and affective components, the TPB This investigation drew upon several prominent may be limited by not including affective measures of psychological theories to provide guidance in developing attitude (Arvola et al. 2008). Inquiry into the emotional the interview instrument: the Theory of Planned Behavior component of organic food purchasing may give a more (TPB), the Prototype Approach to Self-Identity, lifestyle complete picture of consumer attitudes. analysis, and the Elaboration Likelihood Model of Persuasion. These theories have most commonly been R1: What are the cognitive and affective components of applied and validated in survey or experimental settings, attitude toward organic food consumption held by but they also offer a useful conceptual framework for consumers?

264 Marketing and Public Policy Conference Proceedings / 2009 Subjective norms form the basis for the consumer’s per- R5: What prototypes of self-identity are relevant to the ception of approval or disapproval from important indi- purchase of organic food? viduals or groups. Moral norms, which represent a sense of personal obligation to behave in a particular way, can Demographic variables appear to be weakly predictive of influence behavior above and beyond the impact of atti- organic food purchase, and the relationships are not tudes and social norms, increasing the explained variance always clear (Hughner et al. 2007). A consideration of in TPB models (Ajzen 1991). Prior research has found an consumer values and lifestyles may be a richer and more independent impact of moral norms on the purchase of productive approach to understanding what drives or- organic foods (Arvola et al. 2008; Sparks and Shepherd ganic food consumption. Values may be important in 2002). Perceived behavioral control indicates the degree many aspects of consumer behavior, including food (Gold- to which the consumer believes he or she is capable of smith, Freiden, and Henderson 1995). Research has also performing the behavior. found that the same people often have a preference for organic food, belief in alternative medicine, adherence to R2: What are the normative beliefs held by consumers of a vegetarian diet, and an orientation toward environmen- organic food? tally-sound behaviors (Fox and Ward 2008), suggesting that for some consumers, organic food purchase may not R3: What are the behavioral control beliefs held by con- be an isolated choice, but rather part of a “green” lifestyle sumers of organic food? that prioritizes nature, health, and sustainability. R6: How does organic food consumption relate to the Although the basic TPB model has received substantial consumer’s values and lifestyle? empirical support, researchers have included additional variables to improve the predictability of the model (see The Elaboration Likelihood Model (ELM) examines the Conner and Armitage 1998). One obvious omission in the effects of source, message, and audience variables on TPB is the role of past behavior. Some research has found information processing and attitude change. This would that past behavior influences behavioral intention directly apply to advertising, in-store displays, label information, rather than being mediated by other constructs in the and government regulations. In the ELM, the listener’s model; including past behavior improves the explained level of involvement with the product can lead to in-depth variance of the model (Ouellette and Wood 1998). The scrutiny of messages and arguments for high involvement relevance of past behavior may be especially important consumers and the reliance on cues such as source cred- when the behavior in question, such as food purchase, is ibility and expertise for low involvement consumers (Petty, performed frequently (Smith et al. 2008). Cacioppo, and Schumann 1983).

R4: How does past behavior influence current behavior R7: How do involvement and source credibility affect the for consumers of organic food? consumers’ use of advertising and other information about organic food? Self-identity is another variable frequently used to enhance the TPB (e.g., Cook, Kerr, and Moore 2002; Sparks STUDY METHODS and Shepherd 1992; Pierro, Mannetti, and Livi 2003). One dimension of self-identity, personal identity, relates Depth interviews were conducted with a convenience to “self-definitions in terms of unique and idiosyncratic sample of thirty adults who reported that they purchased characteristics” (Pierro et al. 2003, p. 48). An individual’s organic foods. The interview began by asking respon- personal identity can be created, augmented, and sup- dents to define organic food, followed by several ques- ported by the actions he takes, including the purchase, tions related to each of the variables in the expanded TPB, consumption, and display of consumer goods (Kleine, values and lifestyles, responses to advertising and govern- Kleine, and Kernan 1993). Thus, a consumer may make ment communications, and demographics (age, gender, purchase decisions for symbolic reasons that are not ethnicity). Interview questions are available upon request. easily captured by expected utility models such as the The interviews lasted from twenty to forty minutes. The TPB. The Prototype Approach offers one way of constru- sample consisted of ten men and twenty women, ranging ing the influence of personal identity on behavior. Ac- in age from 19 to 64, with a median age of 37. The sample cording to this theory, people develop characteristic im- contained primarily white, non-Hispanic respondents and ages of the kind of person who engages in a behavior and some Hispanic respondents. Most respondents lived in compare their own self-images to these prototypes (Pierro highly populated areas where purchasing organic foods is et al. 2003). relatively convenient (e.g., Austin, TX).

Marketing and Public Policy Conference Proceedings / 2009 265 RESULTS pesticides are to blame for “cancer or other catastrophic things” (Jill, 30) but tends to “feel more confident about Definition of Organic Food [organic food’s] nutrition” (Jeff, 47) and consumes it as a “precautionary step” (Rob, 31). A third, although smaller, The United States Department of Agriculture defines group does not view organic food as healthier. One organic in the following way: “Organic crops are raised respondent, a dietician, views organics as “part of the mix without using most conventional pesticides, petroleum- of good food choice” and perceives that an all-organic diet based fertilizers, or sewage sludge-based fertilizers. Ani- is not healthier than one with a blend of foods (Barbara, mals raised on an organic operation must be fed organic 51). Another consumer mentioned that eating organic feed and given access to the outdoors. They are given no food can cause a contrast effect with other food: “Once antibiotics or growth hormones. The NOP regulations you start, it can sort of make you feel that the non-organic prohibit the use of genetic engineering, ionizing radiation, foods aren’t good for you. . . . I don’t actually have any and sewage sludge in organic production and handling. reason to think it’s healthier” to eat organic food (Tara, As a general rule, all natural (non-synthetic) substances 33). are allowed in organic production and all synthetic substances are prohibited” (USDA 2008, p. 1). Comparing Respondents disagreed as to whether organics are of the official definition to respondents’ definitions indi- superior quality compared to non-organics, with organic cated a high degree of correspondence. Consumers fre- produce being particularly controversial. Some respon- quently mentioned the lack of “pesticides” and “chemi- dents extol the better taste of products such as organic cals.” Some consumers also noted the non-use of growth carrots, broccoli, bananas, and apples, while others find hormones for animals. One consumer described the food organic fruits and vegetables to be inferior in quality to as being grown without “any toxic or unhealthy or their conventionally-grown counterparts. One manmade ingredients” (Susan, 50). Others elaborated that respondent’s husband characterizes organic food as “ugly organic food is not “industrially” or “mass produced.” food” (Barbara, 51) whereas another consumer’s boyfriend buys organic eggs because “brown eggs are pret- Cognitive and Affective Components of Attitude To- tier” (Lisa, 45). A respondent noted that “when certain ward Organic Food Consumption chemical sprays are not an option, suppliers must get food to market while it’s still fresh” (Ryan, 46), but several The vast majority of consumers identified improved health other consumers noted that organic produce does not last as the primary benefit of eating organic food. While some as long as non-organic food once it is brought home. There respondents mentioned health and well-being in a general appears to be no relationship between beliefs about food sense, others specified the reduced consumption of pesti- quality and the stores where people shop for organic cides, hormones, and other potentially harmful chemicals products or what specific organic products they buy. or the increased nutritional value as advantages over conventionally-grown food. Because “things affect kids The higher price of organic food was unanimously men- more than adults” (Brad, 30), several respondents with tioned as a disadvantage of purchasing organic products, children are especially concerned with the food their with several consumers reporting “sticker shock” at see- children eat. One parent noted that “we’ve gone around ing the prices. One consumer commented that “when I the bend adding growth hormones and pesticides . . . and look at the receipt at the cash register, I get a slightly the FDA allows or ignores it” (Gina, 38). Most consumers queasy feeling in my stomach” (Aaron, 36). However, see a difference between organic and non-organic food, some respondents felt that the cost was becoming more though they vary in the degree to which they are con- comparable to conventionally-grown foods over time. vinced of organic food’s benefits, particularly with regard For the most part, the benefits of organic food are seen to to specific health claims. justify the higher price. Both supply side factors (such as the longer time for growing organic crops and the loss of One group believes that there are definite, serious health crops to bugs) and demand side factors (such as superior consequences to eating conventionally-grown food, espe- taste and health benefits) are seen to validate the price cially over a long period of time, and eat organic food to differential. Belief in the suitability of organic food prices avoid these problems. Some diseases attributed to non- amongst those who purchase it would be expected based organic food include: “allergies, obesity, heart disease, on the principle of cognitive consistency (Festinger 1957). you name it!” (Matt, 34), “health issues like cancer, A couple of consumers even appear to use their purchas- allergies, and birth defects” (Diana, 33), and “shorter life ing of organic food as a method for determining their own expectancy” (Ryan, 46). Another group is less certain that attitudes on the matter: “The things that I buy feel worth

266 Marketing and Public Policy Conference Proceedings / 2009 it to me to buy and the ones I don’t buy seem that they little connection between the perceptions of others and aren’t” (Tara, 33) and “[The price] doesn’t seem to pre- respondents’ behavior. vent me from paying it, so I guess it’s not that high” (Jill, 30). However, there is a concern that producers or sellers Purchase of organic food can have a significant ethical are taking advantage of people’s willingness to pay more dimension because of the perceived negative societal for organic products that are healthier or “that people consequences from conventional food production. Envi- believe in and will buy” (Jill, 30). Two organic food ronmental concerns were a factor in the organic food consumers doubt that the price is justified; one wondered purchase decision for over a third of respondents. One whether the price reflects “a snob appeal associated with respondent views the impact on the environment as “im- it” (Jeff, 47) while the other believes “a lot of it is over- mense – fertilizers, non-biodegradable pesticides, worker priced, based on a marketing perception” (Dan, 37). A few exploitation, etc.” (David, 51). Conventional farming respondents commented that eating organic food is a techniques are also seen to lead to “environmental degra- “luxury.” One consumer noted that organic “is also like a dation” (Susan, 50) and “continuing pollution” of the name brand, like . . . ‘oooh, it’s organic’” (Rob, 31). planet (Karen, 37). Although some consumers believe that their organic purchases will contribute to a broad Respondents reported feeling healthier, happier, and more “movement” or “sea change” toward more environmen- relaxed when purchasing, preparing, and eating organic tally-sound practices throughout the food production in- food. Some respondents feel that they are doing some- dustry, others appear to see their impact as more limited thing good for themselves and their families and “contrib- and the purchase as more symbolic; one respondents notes uting in some small way” (David, 51) to a better environ- “I don’t feel like I’m saving the trees by buying an organic ment. Consumers frequently emphasized the feeling of apple, but I feel I’m doing something helpful” (Jill, 30). making a conscious, intelligent, and/or responsible choice that comes from buying organic food. Consumers’ nega- Others are concerned about animal welfare, encompass- tive feelings toward non-organic food can motivate them ing both the health of animals and their living conditions: to purchase organic as well; respondents reported feeling “Animals are fed on food that makes them grow fast and “scared” of pesticides and “creeped out” by unnatural unnaturally and I’m sure that can’t be healthy or good for products. One respondent noted that organic food “doesn’t the animals” (Anna, 49) and “I dislike the image of make me feel anything special, but eating non-organic thousands of chickens cooped up in small cages, the bad food makes me feel ill” (Daphne, 62). smell all around” (Ryan, 46). Some consumers are motivated to purchase organic because they want to “support Normative Beliefs small businesses and farms” (Carla, 49) and the “community of people who raise their products” organically (Joe, There was no correspondence between the views of the 31). participants and their friends and family members. How- ever, there was a slight tendency for friends to have Perceived Behavioral Control and Control Beliefs greater involvement with organic food than family members. One consumer noted that his “mom makes an effort Respondents viewed purchasing organic food as some- to buy something [organic] and I think impress me” when thing within their control, and they all intend to continue he goes home (Brad, 30). Some respondents noted differ- purchasing it. Organic foods are generally seen as avail- ences in the motivations for eating organic food between able in their local areas. Consumers believe that the themselves and others in their lives: “[My boyfriend] eats availability has increased over time, with more supermar- more than organic than me, but for weirder reasons” (Lisa, kets making organic options available and a greater choice 45) and “My husband thinks of it as health . . . I’m more of organic products overall. However, some consumers a ‘what about the chickens?’ kind of person” (Karen, 37). wish that organic food was more available in restaurants: “I wish we had more options in going out, but sadly it Respondents also varied a great deal in their beliefs about doesn’t prevent us from doing it” (Rob, 31) and “I don’t how others evaluate their consumption of organic food. eat at restaurants; I’m always conscious that it’s not One respondent described the responses as ranging from organic food” (Daphne, 62). “at least she’s trying to do the right thing” to “idiot, why is she paying 98 cents for a can of beans when she can get Past Behavior them for 48 cents?” (Tara, 33). Participants reported being described as “picky,” “hippie and a little indulgent,” Consumers appear to use three different strategies for “over-the-top,” “kind of extreme,” “bizarre,” and “tree- selecting the products they will purchase. One strategy is hugger crazy people.” Nevertheless, there appears to be to select specific product categories for which buying the

Marketing and Public Policy Conference Proceedings / 2009 267 organic version seems particularly important due to health convinced of, the benefits of organic food or who may not or environmental concerns: “Round and yellow things be able to afford it. However, several respondents did hold because I’ve heard it’s important for those” (Lisa, 45) and negative images of those who do not eat organic food: “Definitely berries, those are the worst for pesticides” non-organic purchasers “probably expose themselves to a (Gina, 38). A second strategy involves buying specific shorter life, more diseases, and more health care costs” brands for different product categories. The third strategy (Daphne, 62) and may even be “self-destructive” (Diana, is to look at the price discrepancy between the organic and 33). One respondent noted that he tries to remind himself conventionally-grown versions to determine what to pur- that they are “not just a bunch of slobs, who don’t care chase: “With the price difference for vegetables, I don’t about themselves or the environment” (Rob, 31). Another always buy organic” (Daphne, 62) and “If organic apples consumer characterized them in general as being “not my are much more than conventionally-grown, I will do kind of people” (Karen, 37). Purchasing organics is a way without apples . . . I think there are plenty of alternatives that these consumers can distance themselves from an to buying a $3 apple” (Barbara, 51). Respondents differ in identity as an uneducated, lower class, careless person. how long they have been organic food customers, ranging from one year to more than twenty years. These consum- Values and Lifestyle ers have incorporated organic food purchases into their regular schedule with most buying at least one organic Eating organic food does appear to be part of an overall food item per week. This suggests that the consumption lifestyle choice for most consumers, with an emphasis on behavior has become habitual for them to some extent, the values of personal and environmental health. For especially for those who regularly purchase particular example, only five consumers did not mention regular products or brands of organic foods to control costs. exercise as part of their lifestyle, and most respondents follow other dietary restrictions, including avoiding junk Self-Identity food, fats, and sugar; more than half of respondents follow a vegetarian diet to some extent. Many respondents re- Consumers most commonly hold an image of organic cycle and limit their driving in order to “be conscientious food consumers as smart and conscientious: “Concerned, of our footprint on the earth” (Karen, 37). Some respon- mindful folks who seek to be healthy and lessen man’s dents are highly involved with animal rights causes. impact on the planet” (David, 51). For these consumers, Almost one-third make an effort to eat locally grown food, personal purchases of organic foods may reinforce posi- and more would purchase it if it were more readily tive aspects of their self-identity because the action is available. associated with intelligence, thoughtfulness, and concern for important values. A second popular prototype is the Consumers’ Use of Advertising and Other Informa- organic food consumer as hippie: “You mean those nuts tion About Organic Food and berries?” (Jeff, 47) and “Hippie type people” (Lisa, 45). One respondent noted that there has been a change in Many consumers who purchase organic foods themselves the image of organic food consumers: “It used to be . . . are nonetheless unaware of any advertising for organic extreme or eccentric, but now, it’s just ordinary people” products. While some of them believe that increased (Sandra, 61). However, a minority of respondents hold advertising would be beneficial in changing people’s somewhat negative stereotypes of other organic food eating habits, other think that advertising organic foods consumers, believing them to be overly idealistic, moti- can be “misleading,” would be a “waste of money” (Ryan, vated by doing what is “trendy,” or “misguided in their 46), or would undermine the perception that organic foods perception of the value” of organic foods (Dan, 37). are inherently superior products that do not require marketing spin to be preferred. A couple of respondents About half of the respondents view consumers who do not have negative attitudes toward current advertising practices buy organic food neutrally; these organic food buyers see that involve scare-mongering about conventionally-grown it as a matter of “people [having] a choice to make, and it’s food: “I don’t like the way organic food advertising makes their choice” (Joe, 31), think that other consumers may people afraid of regular food” (Tara, 33) and “I think it’s believe that “something else is more important” (Carla, a little extreme . . . it makes you think that things are worse 49), and have elected to not “hold it against people for than they are” (Amy, 19). Overall, consumers do not going with the cheaper alternative” (Brad, 30). Another believe they place much importance on organic food group of organic food consumers see non-organic pur- advertising when buying food. A few respondents chasers as being “commonsensical, working class normal mentioned sometimes seeking out particular items or people” (Tara, 33) who may not know about, or be brands they had seen advertised, but in general, despite the

268 Marketing and Public Policy Conference Proceedings / 2009 high product involvement for these consumers, message- difference between consumers who do or do not buy response involvement for engaging with advertising organic may be how knowledgeable they are about what outside the purchase environment is low. constitutes “organic.”

Another source of information is government regulations. The results were broadly supportive of the applicability of Consumers expressed three different views of the actions the TPB to organic food purchases. The purchase inten- the government should take regarding organic and non- tions of those sampled are very positive toward continu- organic food. A small group believes that it would be ing to buy and consume organic food. Organic food wrong for the government to take any action that plays consumers have developed cognitive beliefs and affective favorites between organic and conventionally-grown foods responses to organic food that are generally consistent or that would “dictate who can eat what” (Vicki, 20). with their purchase behavior, following the principle of Another group favors the government protecting the vi- cognitive consistency. This suggests that non-consumers ability of small organic producers and encouraging addi- of organics may hold negative cognitive beliefs, negative tional organic production through subsidies and other affective responses, or both toward organics. Organic forms of support. This group is looking for a “truer consumers also believe their behavioral control to be economics” (Carla, 49) than is demonstrated by current high; this is partly due to the fact that the respondents in agricultural policies. The largest group of consumers this sample tend to live in areas where organic food is wants the government to continue or increase regulations readily available. regarding certification of organic food, including strict definitions for organic food, inspections, and accurate Although social norms seem to have no correlation to labeling of food in stores. organic food consumption, moral norms do appear to strongly influence behavior. Care for the environment is Respondents were evenly split between those who mostly an important value that favors organic consumption since trust that foods are truly organic when they are labeled as fertilizers and other chemicals used in conventional farm- such and those who believe that food is frequently sold ing are perceived as bad for the land and wildlife. Con- under the “organic” label that has not been grown to legal cerns about the small farmer and the conditions in which standards. These skeptical consumers tend to use the animals are raised also contribute to the consumption of credibility of the source as a way of judging the odds of organic foods. Organic food consumption was often con- getting truly organic food. They prefer to shop at stores sidered part of the respondents’ self-identity. Positive that they know well, have researched, or believe have prototypes included “intelligent,” “healthy,” and “mind- performed an adequate vetting process. They will gener- ful,” along with “hippie.” Some respondents also consid- ally have confidence in the product if they “trust the ered non-consumers to be less informed or less concerned merchant, like Whole Foods . . . If you get an organic about their health. Organic food consumption appears to cookie from Walgreens, I would be very suspicious” be part of a lifestyle choice that includes exercise, eating (Rob, 31). Reports that “some food stores have their healthy, and concern for animal rights and the environ- produce washed off and changed into containers” that are ment. labeled organic led one respondent to use established farmers’ markets for her organic shopping (Barbara, 51). Though the key themes of health, quality, and price were Consumers also look at the reputation of the brand, with common to respondents’ beliefs, these concepts were some respondents trusting “bigger brand names” (Susan, shown to be complex, multidimensional constructs that 50) while others are wary of “a big corporation” (Courtney, differed across the sample. One of the most commonly 36), particularly if it also markets non-organic foods. held beliefs among organic food consumers is that these foods are “healthier” than conventionally-grown prod- DISCUSSION AND IMPLICATIONS FOR ucts. However, some consumers are convinced that only PUBLIC POLICY organic products are truly safe to eat. Others eat organics because of the possibility that they provide more nutrients These regular consumers of organic food were quite or help avoid chemicals with unknown health effects. consistent overall with their definition of organic food, Some see organics as simply one part of a healthy diet. describing it primarily in terms of what it does not contain Respondent views on food quality were even more wide- (pesticides, chemicals, and additives). This indicates that ranging. “Quality” was described with reference to taste, it is possible to gain useful perspective on the organic food appearance, and freshness, and respondents rating organ- category from this group. It further suggests that one ics as high along one of these dimensions frequently

Marketing and Public Policy Conference Proceedings / 2009 269 mentioned the lack of one of the others (e.g., tasty but natively, word of mouth communication, testimonials, ugly). Further complicating matters is that the organic and informative PSAs from experts could increase knowl- label was sometimes seen as a signal of overall quality and edge about organics for a greater number of consumers, not something pursued for its own sake. Another belief, which could translate into stronger, more positive atti- held by consumers and non-consumers alike, is that or- tudes among non-consumers while avoiding a negative ganic foods cost considerably more than non-organics effect on current consumers. (Hughner et al. 2007). But while organic consumers tend to view this as justified, there remains concern for some Most respondents advocated stricter government inspec- that the higher cost reflects price discrimination by sellers tions and labeling because of concerns about the accuracy and is not sufficiently tied to the underlying value of the of “organic” labels. However, these doubts do allow products. Respondents also noted the differences in atti- resellers like Whole Foods and producers like Earth tudes between themselves and other organic consumers in Bound Farms to engender loyalty from shoppers who their own household. The views of non-purchasers may come to rely on the brand name rather than a general be even more diverse. government label as a mark of assurance. It is unclear how the organic label interacts with other nutrition information This variety in attitude content has implications for rein consumer decision-making. To the extent that a con- searchers, marketers, and policy makers. The use of sumer relies on simplifying beliefs like “food labeled quantitative methods that do not cover the broad range of organic is healthier and better,” some consumers may beliefs and affective associations covered by terms such make poorer decisions in some situations. For instance, as “health” or “quality” will be poorly suited to uncover- they may select an organic snack cracker with low nutrient ing the ways in which consumers relate to food and make value over a more nutritious non-organic product. Fear of purchase decisions. Further research could also develop pesticides may lead some consumers to eat fewer fruit and theoretical models of the different segments of consum- vegetables than is recommended for their health if they ers, allowing marketers to more effectively position their cannot find or afford organic produce. products. Depending on the details of policy decisions, different aspects of health (e.g., safety vs. nutrient con- In summary, we have discovered that attitudes, moral tent) could be influenced by government actions. norms, values and lifestyle, self-identity, and past behavior are important factors explaining the current and future Organic consumers stated that they are unaware of much behavior of organic consumers, while social norms ap- advertising about organic food and, in fact, think more pear less significant. Furthermore, this in-depth qualita- organic advertising may be counter-productive. One rea- tive approach has uncovered some of the complexity in son is that informed consumers who are concerned about the concepts commonly associated with organic food – their health will purchase them regardless of the quantity such as “health,” “quality,” and “price” – that may not be of advertising messages. Another reason is that advertis- evident in purely quantitative studies (e.g., “prevents ing would diminish the perception of quality held by many cancer” vs. “part of the mix of good food choice”; “ugly organic consumers. Finally, scare tactics encouraging food” vs. “pretty eggs”). It also provides a broader view consumers to avoid non-organic foods are seldom effec- of the variety of ways in which the concepts are linked in tive and might backfire by creating a negative perception the minds of consumers. of the trustworthiness of messages about organics. Alter-

REFERENCES Psychology, 9 (4), 261–72. ______and ______(2001), “Efficacy of Ajzen, Icek (1991), “The Theory of Planned Behavior,” the Theory of Planned Behavior: A Meta-Analytic Organizational Behavior and Decision Processes, Review,” British Journal of Social Psychology, 40 50 (2), 179–211. (4), 471–99. Armitage, Christopher J. and Mark Conner (1999), “Pre- Arvola, Anne, Marco Vassallo, Moira Dean, Piritta dictive Validity of the Theory of Planned Behavior: Lampila, Anna Saba, Liisa Lahteenmaki, and Rich- The Role of Questionnaire Format and Social Desir- ard Shepherd (2008), “Predicting Intentions to Pur- ability,” Journal of Community & Applied Social chase Organic Food: The Role of Affective and

270 Marketing and Public Policy Conference Proceedings / 2009 Moral Attitudes in the Theory of Planned Behavior,” business/18organic.html]. Appetite, 50 (2/3), 443–54. Ouellette, Judith A. and Wendy Wood (1998), “Habit and Conner, Mark and Christopher J. Armitage (1998), “Ex- Intention in Everyday Life: The Multiple Processes tending the Theory of Planned Behavior: A Review by Which Past Behavior Predicts Future Behavior,” and Avenues for Further Research,” Journal of Ap- Psychological Bulletin, 124 (1), 54–74. plied Social Psychology, 28 (15), 1429–64. Petty, Richard E., John T. Cacioppo, and David Schumann Cook, Andrew J., Geoff N. Kerr, and Kevin Moore (1983), “Central and Peripheral Routes to Advertis- (2002), “Attitudes and Intentions Toward Purchas- ing Effectiveness: The Moderating Role of Involve- ing GM Food,” Journal of Economic Psychology, 23 ment,” Journal of Consumer Research, 10 (2), 135– (5), 557–72. 46. Dean, Moira, Anne Arvola, Marco Vassallo, Liisa Pierro, Antonio, Lucia Mannetti, and Stefano Livi (2003), Lahteenmaki, Monique M. Raats, Anna Saba, and “Self-Identity and the Theory of Planned Behavior in Richard Shepherd (2006), “Comparison of Elicita- the Prediction of Health Behavior and Leisure Activ- tion Methods for Moral and Affective Beliefs in the ity,” Self and Identity, 2 (1), 47–60. Theory of Planned Behavior,” Appetite, 47 (2), 244– Robinson, Ramona and Chery Smith (2002), “Psychoso- 52. cial and Demographic Variables Associated with Dimitri, Carolyn and Lydia Oberholtzer (2008), “U.S. Consumer Intention to Purchase Sustainably Pro- Organic Handlers Mostly Small, Focus on Fruit and duced Foods as Defined by the Midwest Food Alli- Vegetables,” Amber Waves, Washington: 6 (4), (Sep- ance,” Journal of Nutrition Education and Behavior, tember) 8. 34 (6), 316–25. Festinger, Leon (1957), A Theory of Cognitive Disso- Smith, Joanne R., Deborah J. Terry, Antony S.R. Manstead, nance. Stanford, CA: Stanford University Press. Winnifred R. Louis, Diana Kotterman, and Fox, Nick and Katie Ward (2008), “Health, Ethics, and Jacqueline Wolfs (2008), “The Attitude-Behavior Environment: A Qualitative Study of Vegetarian Relationship in Consumer Conduct: The Role of Motivations,” Appetite, 50 (2/3), 422–29. Norms, Past Behavior, and Self-Identity,” Journal of Global Market Information Database (2008), Euromonitor Social Psychology, 148 (3), 311–33. International, (accessed October 15, 2008), [avail- Sparks, Paul and Richard Shepherd (1992), “Self-Identity able at http://www.portal.euromonitor.com]. and the Theory of Planned Behavior: Assessing the Goldsmith, Ronald E., Jon Freiden, and Kenneth V. Role of Identification with ‘Green Consumerism,’” Henderson (1995), “The Impact of Social Values on Social Psychology Quarterly, 55 (4), 388–99. Food-Related Attitudes,” Journal of Product & Brand ______and ______(2002), “The Role of Management, 4 (4), 6–14. Moral Judgments Within Expectancy-Value-Based Hughner, Renee Shaw, Pierre McDonagh, Andrea Attitude-Behavior Models,” Ethics & Behavior, 12 Prothero, Clifford J. Schultz III, and Julie Stanton (4), 299–321. (2007), “Who Are Organic Food Consumers? A Springen, Karen (2008), “Best Organics for the Buck,” Compilation and Review of Why People Purchase Newsweek, New York: 152 (15), (October 13). Organic Food,” Journal of Consumer Behavior, 6 (2/ United States Department of Agriculture (2008), “Na- 3), 94–110. tional Organic Program,” (accessed October 15, Kleine III, Robert E., Susan Schultz Kleine, and Jerome B. 2008), [available at: http://www.ams.usda.gov/ Kernan (1993), “Mundane Consumption and the AMSv1.0/getfile?dDocName=STELDEV3004443]. Self: A Social-Identity Perspective,” Journal of Con- Whitman, Deborah B. (2000), “Genetically Modified sumer Psychology, 2 (3), 209–35. Foods: Harmful or Helpful?” ProQuest Discovery Martin, Andrew and Kim Severson (2008), “Sticker Shock Guides, (accessed October 15, 2008), [available at: in the Organic Aisles,” New York Times, http://www.csa.com/discoveryguides/gmfood/ (nytimes.com), April 18, (accessed October 15, 2008) overview.php]. [available at http://www.nytimes.com/2008/04/18/

Marketing and Public Policy Conference Proceedings / 2009 271 For further information contact: Karen H. Smith Texas State University McCoy Hall 422 601 University Drive San Marcos, TX 78666 Phone: 512.245.3192 Fax: 512.245.7475 E-Mail: [email protected]

272 Marketing and Public Policy Conference Proceedings / 2009 PERCEPTIONS OF SERVICE QUALITY WITH CREDENCE SERVICES

Ann M. Mirabito, Baylor University, Waco Leonard L. Berry, Texas A&M University, College Station Jelena Spanjol, University of Illinois at Chicago

SUMMARY cleanliness and attractiveness. We found that the technical quality dimension and the interpersonal quality di- Little is known about how people evaluate credence mension were highly correlated, pointing to the difficulty attributes, that is, those attributes which the consumer consumers of credence services have in disentangling the often cannot fully evaluate even after purchasing and two dimensions. In addition, we found that each of the consuming the product. And yet consumers struggle to quality dimensions had a unique relationship with overall evaluate quality in several important product categories quality, satisfaction and value. dominated by credence attributes such as medical services, legal services, and pharmaceuticals, among others. Buyers of credence services have a difficult challenge: The research explores the processes by which people form how to use the signals from observable service compo- quality evaluations of services high in credence attributes nents to infer the quality of unobservable attributes. We and the consequences of those evaluations. Drawing on identified two paths to evaluation with the role of trust the service quality, dual-process social information pro- distinguishing between the two approaches. In the first cessing, expert-novice and risk literatures, we develop a approach, trust is a consequence of quality evaluations. conceptual model to illustrate how skill and motivation Patients integrate evidence of the physician’s capabilities, moderate the ways people seek and integrate observable practices, and prior outcomes to reach evaluations of information to infer unobservable quality. The influence technical quality. Patients similarly reach evaluations of of quality evaluations on outcome, satisfaction, value, and interpersonal and organizational quality. Then trust flows loyalty is mapped. The model is tested in the context of a from the quality evaluations, consistent with other re- classic credence service, health care services, with two search which finds that trust is a consequence of quality large datasets using structural equation modeling. (Garbarino and Johnson 1999).

Study 1 draws on an existing patient satisfaction database Under the rival approach, trust is used as a heuristic for (6,280 records) to measure the sources and consequences evaluating unobservable technical quality. Specifically, of quality evaluations. Study 2 validates Study 1 findings perceptions of interpersonal and organizational quality and extends those findings to show the moderating roles influence trust formation; trust, in turn, influences techni- of product expertise and perceived risk on quality evalu- cal quality perceptions. ation processes. The second study is tested with 1,379 consumers (patients) drawn from an online consumer Finally, we propose that perceived risk may influence the panel. way credence quality evaluations are made. We suggest that higher levels of risk may increase the motivation for Sources and Consequences of Quality thoughtful evaluations of technical quality and, at the same time, decrease the ability and increase reliance on Following the service quality literature (Parasuraman, cognitive heuristics. Berry et al. 1991; Rust and Oliver 1994; Brady and Cronin 2001), the dimensions of service quality of a physician are Implications conceptualized to include: the physician’s technical quality as evidenced by medical skills, knowledge, adherence The healthcare system in the U.S. is challenged by to best practice and prior outcomes (Donabedian 1980); unaffordable high costs, lack of access, and uneven qual- the interpersonal quality of the physician as evidenced by ity. Some healthcare policymakers advocate a market- communication and caring; and the organizational qual- based solution in which patient-consumers will reward ity as evidenced by the competence and friendliness of the healthcare providers who offer the best combination of auxiliary staff, access convenience, and physical setting quality, prices, and services with higher fees. For the

Marketing and Public Policy Conference Proceedings / 2009 273 market-based approach to work, patients must have the insight into how patients assess medical providers and the ability to act like consumers. The fine-grained theoretical consequences of those evaluations. and empirical model developed in this research offers

REFERENCES Roles of Satisfaction, Trust, and Commitment in Customer Relationships,” Journal of Marketing, 63 Brady, M.K. and J.J. Cronin (2001), “Some New Thoughts (April), 70–87. on Conceptualizing Perceived Service Quality: A Parasuraman, A., L.L. Berry et al. (1991). “Refinement Hierarchical Approach,” Journal of Marketing, 65 and Reassessment of the SERVQUAL Scale,” Jour- (July), 34–49. nal of Retailing, 67 (Winter), 420–51. Donabedian, A. (1980), The Definition of Quality and Rust, R.T. and R.L. Oliver (1994), “Service Quality: Approaches to Its Assessment. Ann Arbor, MI: Health Insights and Managerial Implications from the Fron- Administration Press. tier,” Service Quality: New Directions in Theory and Garbarino, E. and M.S. Johnson (1999), “The Different Practice, 1–20.

For further information contact: Ann M. Mirabito Hankamer School of Business Baylor University One Bear Place #98007 Waco, TX 76798 Phone: 254.775.4316 Fax: 254.710.1068 E-Mail: [email protected]

274 Marketing and Public Policy Conference Proceedings / 2009 REDUCING RESISTANCE TO PLANNING FOR END-OF-LIFE: A SOCIAL MARKETING APPROACH

Elyria Kemp, Texas State University, San Marcos Steven W. Kopp, University of Arkansas, Fayetteville

SUMMARY ticipants were 309 adults from a geographically dispersed household research mail panel in the southern part of the Because of the significance and magnitude of making United States. end-of-life decisions, a number of consumerist groups have made efforts to encourage consumers to plan ahead The experiment included two different advertisements. and make decisions about death care and other end-of-life Both ad message stimuli included information about the alternatives. This research examined how communication benefits of preplanning for a funeral; however, in the messages encouraging planning might be made more Alpha message an added incentive indicated that an effective by using an approach/avoidance framework and individual’s “loved ones would appreciate it” if they incorporating consumers’ attitudes about death. preplanned. The goal of the Omega message was to encourage preplanning, but in incremental steps. The Goals, offers, and opinions are complex stimuli that message stressed that an individual could start planning engage motives. Some of these motives are approach for end-of life by “simply writing down their final wishes motives that push opinions and attitudes toward the goal and requests and sharing them with family and friends.” and other motives enable avoidance. Alpha persuasive strategies activate the approach forces, thereby increasing Attitude about death, or fear of death, was an individual motivation to move toward the goal. Omega persuasive difference measure that was not manipulated. Fear of strategies promote change by minimizing the avoidance death was measured using the Death Attitude Profile – forces, reducing the motivation or resistance to move Revised (Wong, Reker, and Gesser 1994). Subjects rated away from the goal. Alpha strategies attempt to persuade their attitude toward the ad and their response to the ad. by increasing approach forces and the attractiveness of an They were also asked if they would be interested in alternative. Offerings are made more attractive by adding receiving information about funeral planning. incentives or inducements for compliance, creating more convincing reasons, and finding more credible sources Results (Knowles and Linn 2004). In contrast, Omega strategies reduce the reluctance or dislike of an alternative. Ac- Findings show that there was a main effect for communi- knowledging consumers’ resistance to planning for end- cation message on attitude toward the ad. Individuals in of-life might help to mitigate the “avoidance forces” that the Omega message condition expressed more favorable prevent them from doing so. Because this resistance may attitudes toward the ad than individuals in the Alpha stem from an individual’s unfavorable attitude about message condition, M = 5.2 and M = 4.65, p < .01. death, the implementation of Omega strategies in mes- Additionally, there was a main effect for message on sages that encourage consumers to make incremental response to the ad. Subjects in the Omega message condi- progress toward the preplanning process might be effec- tion exhibited more favorable responses to the ad than tive. The effectiveness of using both Alpha and Omega those in the Alpha message condition, M = 4.7 and M = strategies in communication messages for encouraging 4.1, p < .01. Results from planned contrasts also indicated planning behavior was tested empirically. that individuals in the high fear of death and Omega message conditions had a greater response toward the ad Methodology than those in the high fear of death and Alpha message conditions, M = 4.7 and M = 4.2, p < .05. Although To evaluate the effectiveness of using Alpha and Omega individuals in the high fear of death and Omega message persuasive communication strategies for end-of-life plan- conditions, expressed more favorable attitudes toward the ning, a 2 (Communication Message: Alpha/Omega) X 2 ad than those in the high fear of death and Alpha condi- (Attitude about Death: Low Fear of Death/High Fear of tions, M = 5.1 and M = 4.6, these differences did not reach Death) between-subjects experiment was performed. Par- statistical significance, p = .06.

Marketing and Public Policy Conference Proceedings / 2009 275 Individuals in the Omega message condition were more The issue of end-of-life planning and death care consump- interested in receiving information (58%) than those in tion involves complex behaviors and processes. In this the Alpha message condition (42%), Kendall’s tau b -.15, study, results suggested that the Omega message was p < .05. Additionally, individuals in the high fear of death more effective in addressing the resistance individuals and Omega message conditions (69%) were more inter- might have toward engaging in planning for end-of-life. It ested in receiving information than those in the high fear was also found that females had more favorable responses of death and Alpha message conditions (31%), (Kendall’s than men for the Omega ad. This is especially significant tau b -.24, p < .05). Further analysis indicated that there since women tend to outlive men, and are more prone to was a significant Message x Gender interaction. Females spousal bereavement and funeral planning (DeSpelder in the Omega message condition had more favorable and Strickland 2008). Hence, social marketing messages responses toward the ad than males in the Omega message might be made more effective by addressing the resistance condition, M = 4.18 and M = 4.9, p < .05. that individuals have about planning for end of life.

For further information contact: Elyria Kemp McCoy College of Business Administration Texas State University 601 University Drive San Marcos, TX 78666 Phone: 512.245.7882 Fax: 512.245.7475 E-Mail: [email protected]

276 Marketing and Public Policy Conference Proceedings / 2009 WHAT’ S BEHIND MY GOOD DEEDS? SELF-ENHANCEMENT, SELF-VIEW, AND DONATION

Joon Yong Seo, University of Utah, Salt Lake City Debra L. Scammon, University of Utah, Salt Lake City

SUMMARY information in different ways in their judgments and decisions (Torelli 2006). We hypothesize that informa- Most people evaluate themselves as more honest, coop- tion about others’ donations moderates the relationship erative, and considerate than others are (e.g., Goethals between self-construal and self-enhancement (H5). et al. 1991). This finding illustrates self-enhancement which refers to unrealistically positive self-evaluations. Study 1 The current research examines self-enhancement in the altruistic domain. Our interest is in understanding how A 2 (self-construal manipulation: between-subjects) X 2 self-enhancement might affect individuals’ motivation (self vs. other evaluation: within-subject) mixed design for involvement in pro-social behaviors. Factors influenc- was used. Participants first completed self-construal ma- ing self-enhancement in the altruistic domain are likely to nipulations followed by a manipulation check. They then influence various helping behaviors, and understanding evaluated themselves and, separately evaluated other stu- the antecedents of helping behavior should have practical dents at the same university, on charitable traits. The implications for many pro-social behaviors. We begin this results showed that participants in the interdependence exploration by focusing on a personal characteristic (self- condition, compared to participants in the independence construal) and a situational factor (availability of social condition, rated themselves in more positive terms than information) that relate to self-enhancement on charitable they rated others. This finding supports H1. traits and donation decisions. Study 2 We demonstrate that one important antecedent of self- enhancement in the altruistic domain is one’s self-construal Participants rated themselves on charitable traits relative which refers to the degree to which individuals see them- to other students. They also rated the extent to which each selves as separated from (independent) or connected with trait was important to their self-identity. They then com- others (interdependent) (Markus and Kitayama 1991). pleted the Self-Construal Scale (Singelis 1994). Finally, Research shows that self-enhancement is more pervasive they were given three dollars in cash and given a chance among independents than interdependents (e.g., Heine to donate to their preferred charity. Interdependents con- et al. 1999). Although this is the prevailing view in the sidered charitable attributes more personally important literature, other research suggests that both independents than did independents, and this personal importance me- and interdependents self-enhance on personally impor- diated the self-enhancement tendency (H2). Interdepen- tant attributes (e.g., Sedikides et al. 2003). Markus and dents gave more to a charity than did independents (H3). Kitayama (1991) suggest that interdependents strive for Self-construal influenced donation amount via self-en- connections and harmony with others through consider- hancement, such that strong interdependent self-view → ing and meeting the expectations of significant others. strong self-enhancement → higher donation amount (H4). This suggests that individuals with dominantly interdependent self-views may value charitable traits more than Study 3 others do. On the basis of these findings, we hypothesize that interdependents self-enhance on charitable traits more A 2 (self-construal measurement) X 2 (social information: than do independents (H1), and personal importance of $2.50 vs. $7.50) between-subjects design was used (DV: charitable traits underlies this effect (H2). We then extend self-enhancement). Participants were given a message the implications of these relationships to charitable giv- that solicited donations to local charities. They then read ing. We hypothesize that interdependents, compared to information about how much other students at the same independents, are likely to give more to charities (H3). We university donated to the charities. Half of the participants further hypothesize that self-construal influences dona- were told that other students donated $2.50 on average to tion via self-enhancement (H4). Previous research sug- the local charities, while the other half were told that other gests that independents and interdependents use social students donated $7.50 on average. Next, participants

Marketing and Public Policy Conference Proceedings / 2009 277 rated themselves on charitable traits relative to other of interdependents. Moreover, we demonstrated that per- students. Finally, they completed the Self-Construal Scale. sonal importance of charitable traits underlies this effect. The results showed that the effect of self-construal on self- Finally, our research is one of the first to demonstrate the enhancement was qualified by the differential social in- interaction between self-construal and social information, formation. In the $2.50 condition, self-enhancement which subsequently impacts self-enhancement in the al- emerged among interdependents, but not among indepen- truistic domain. dents. This pattern was not observed in the $7.50 condition, supporting H5. Practical Implications. When given the opportunity to donate, high interdependents donate more. Charities may Discussion attempt to target individuals with strong interdependent self-views and/or design communications that induce Theoretical Implications. Our study adds to the self- interdependent self-views. In terms of intentions to do- enhancement literature by investigating this tendency nate, those with a high self-enhancement tendency are specifically on charitable traits and from the perspective more likely to donate. Non-profit practitioners would of self-construal. In contrast to the prevailing view in the benefit from understanding their target audiences’ self- literature that interdependents do not self-enhance, our enhancement tendency. findings attest to substantial self-enhancement on the part

For further information contact: Joon Yong Seo University of Utah 1645 E. Campus Center Dr. Salt Lake City, UT 84108 Phone: 801.581.8651 E-Mail: [email protected]

278 Marketing and Public Policy Conference Proceedings / 2009 SPECIAL SESSION

DEVELOPMENTAL PERSPECTIVES ON CHILDREN’S RESPONSE TO MARKETING STIMULI

Session Chairs: Wendy Attaya Boland, American University, Washington, D.C. Paul M. Connell, Stony Brook University

Discussant: Sonya Grier, American University, Washington, D.C.

SESSION OVERVIEW understanding of sales improves with age. However, regardless of this understanding, children favor conceptu- The primary objective of this session is to bring together ally easier sale promotions even when economically sub- three papers that examine effects of marketing to children optimal decisions result, providing useful insights into the from a developmental perspective. That is, the studies in deficiencies that many children have when it comes to this session investigate how children react to marketing sales. This research has implications for parents, educa- stimuli differently as they age and gain cognitive sophis- tors, and policy makers. While many parents and teachers tication. The first paper explores age differences in view consumer education as a desirable learning out- children’s understandings of the symbolic meaning of come, school curricula do not consistently include or consumption, and the second paper examines develop- stress the development of consumer skills. To date, there mental aspects of children’s understandings of sales pro- have been only limited attempts by policymakers to ad- motions. Finally, the third paper explores how advertising dress the communication of sales to children, none of experienced during early childhood can have long-lasting which directly address temporary price reductions. While effects that persist into adulthood due to a lack of persua- regulation in retail settings is more difficult to enforce, sion knowledge at time of encoding. primarily because of the difficulty in determining what constitutes a promotion directed to children, policymakers The first paper (by Chaplin and Lowrey) focuses our might consider establishing guidelines that require dis- attention on how children form stereotypes of other people counts to be computed and presented to young consumers based upon what they consume, also known as consump- in ways that reduce confusion. Ultimately, the findings of tion constellations. To date, research on consumption this work provide direction in protecting children from constellations has primarily focused on adults. The au- potentially confusing or deceptive marketing practices, thors seek to provide deeper understanding for the origins and in building better foundations for children’s con- of consumption constellations and how they develop. sumer education. They find that stereotypes become increasingly rigid and strong as children reach adolescence, but that these stereo- The third paper (by Connell, Brucks, and Nielsen) at- types relax and become more flexible by late adolescence. tempts to explain long-term effects of advertising to This research has important policy implications because it children on judgment that result in biases persisting into sheds light on children’s understanding of marketers’ adulthood, and what can increase effectiveness of public cross-promotional tactics, and how these tactics affect policy initiatives aimed at undoing some of those effects. children’s development of consumer stereotypes that have The authors explore this research question in the context the potential to carry into adulthood. of products that are high in sugar content in order to explore potential policy implications related to the public The second paper (by Boland and Erickson) demonstrates health issue of obesity. Results suggest that emotion felt that children’s purchase decisions are influenced by the toward early childhood advertising objects, but not late presence of a sale and that their understanding is suscep- childhood advertising objects, leads to biased judgments tible to the manner in which the sale is communicated. The in favor of the product, due to age differences in persua- authors find that children’s conceptual and computational sion knowledge when the advertising objects were ini-

Marketing and Public Policy Conference Proceedings / 2009 279 Special Session: Developmental Perspectives on Children’s Response to Marketing Stimuli

tially encoded into memory. In addition, priming health advertising objects) because their attitudes are resilient to motivation or actively engaging cognitive defenses against such interventions. advertising are effective only when individuals experience lower levels of affect toward the advertising objects. Following the presentation of the three papers, Sonya Therefore, well-intended policy interventions are likely to Grier synthesizes the theoretical concepts to compare the not be effective for consumers that are arguably the most papers, propose questions for discussion, and moderate important ones to reach (i.e., those with high affect toward audience input.

280 Marketing and Public Policy Conference Proceedings / 2009 Special Session: Developmental Perspectives on Children’s Response to Marketing Stimuli

YOU DRIVE A PRIUS? I BET I KNOW WHAT BRAND OF CEREAL YOU EAT! CHIDLREN’S UNDERSTANDING OF CONSUMPTION CONSTELLATIONS

Lan Nguyen Chaplin, University of Arizona, Tucson Tina Lowrey, University of Texas at San Antonio

SUMMARY form consumption constellations for different social roles without the aid of experimental stimuli? (2) Are there People make assumptions about others based on the social-cognitive developmental milestones that would products and brands they own. Thus, if one sees a woman help to explain age differences in the structure (e.g., size) in designer fashions, driving an SUV, and talking on a cell and content (e.g., using products and brands vs. other phone, one might make inferences about her personality items) of consumption constellations? and (3) Are there and other products and brands she owns. Adults form such developmental changes that occur after the onset of ado- stereotypes (referred to as consumption constellations) lescence? The purpose of our research is to provide more quite often (Lowrey et al. 2001). However, little is known insight into the development of children’s consumption about when such consumption-based stereotypes begin to constellations by answering these questions. In doing so, develop, what they look like in early stages, and how they we provide the first theoretical explanation for why we develop over time. might observe age differences in children’s consumption constellations. Over the past two decades, the degree to which marketers have scaled up efforts to reach children is remarkable Three different methods were used to examine the devel- (Schor 2004). Research on children’s consumption con- opment of consumption constellations in children ages 5– stellations should shed light on how diverse marketing 16, finding an increasing linear age trend in the number of cues are received by children and how their levels of brand products and brands children use to form constellations, consciousness and materialism may be affected. It should as well as the degree to which these elements display also help parents, educators, and other concerned con- symbolic complementarity. However, by early adoles- stituents understand how marketers’ increasingly popular cence, stereotypes become stronger, with constellations cross-promotional tactics affect children’s knowledge of becoming smaller and less flexible. Although seventh social roles, which can lead to stereotypes and feelings of graders use more products and brands to form constella- prejudice that may carry into adulthood. tions than younger children, they do so in place of other ways to define roles, such as personality traits, therefore To date, research into the nature and underlying processes forming smaller constellations. By late adolescence, indi- of consumption constellations has focused on adults, not viduals develop larger, more flexible constellations. children. Unanswered questions include: (1) Can children

REFERENCES Verification of Consumption Constellations: Impli- cations for Advertising Strategy,” Journal of Adver- Lowrey, Tina M., Basil G. Englis, Sharon Shavitt, and tising, 30 (1), 29–39. Michael R. Solomon (2001), “Response Latency Schor, Juliet B. (2004), Born to Buy. New York: Scribner.

Marketing and Public Policy Conference Proceedings / 2009 281 Special Session: Developmental Perspectives on Children’s Response to Marketing Stimuli

For further information contact: Lan Chaplin Eller College of Management University of Arizona P.O. Box 210108 Tucson , AZ 85721 Phone: 520.626.3701 Fax: 520.621.7483 E-Mail: [email protected]

Tina M. Lowrey The University of Texas at San Antonio One UTSA Circle San Antonio, TX 78249–0638 Phone: 210.458.5384 Fax: 210.458.6335 E-Mail: [email protected]

282 Marketing and Public Policy Conference Proceedings / 2009 Special Session: Developmental Perspectives on Children’s Response to Marketing Stimuli

DO SALES PROMOTIONS PERSUADE CHILDREN TO BUY?

Wendy Attaya Boland, American University, Washington, D.C. Lance-Michael Erickson, University of Arizona, Tucson

SUMMARY Four experiments were conducted to test whether or not: (1) children understand that sales reduce the price of an A growing body of evidence suggests that children, at a item (experiments 1–3); (2) children’s purchase decisions very young age, understand the sequence of events in- are influenced by the presence of a sale (experiments 1– volved in a shopping trip (e.g., Berti and Bombi 1998; 4), and (3) children favor conceptually easier sales promo- Peracchio1992). Children frequently accompany their tions even when they are not economically sub-optimal parents to retail stores and amass a good deal of knowl- (experiment (4). Since we expect to see differences across edge about aspects of shopping such as store layouts, age groups, our sample includes both second and fifth product offerings, and exchanging money for goods grade children. (McNeal 1992). However, one particular aspect of the shopping experience that is not well understood is how Our results provide support that children are able to sale cues impact children’s purchase decisions. Retailers conceptualize and react to sale cues that are portrayed communicate sales to customers in a variety of ways, both semantically and numerically. By simulating a pur- including purely semantic cues such as “clearance” or in chase situation, our results also indicate a significant ways that indicate the amount of the price reduction such increase in the choice share of sale items, providing as “take 30% off.” The aim of this paper is to broaden our evidence that sale cues play a role in children’s purchase understanding of how children make purchase decisions decisions. Lastly, we find that framing a promotion in when sale cues are present. Explicitly, we propose that terms of a quantity discount such as a “Buy One, Get One children recognize sale cues as a reduction in price and Free” sale is preferred to a “60% off” sale because of the that this conceptualization is used as an input to their perceived complexity associated with conceptualizing a decision-making. However, we also propose that the sale in terms of percentages, leading to an economically manner in which a sale is communicated affects their inferior result. We conclude with a discussion of the understanding and may lead some children to make eco- implications of our findings for parents, educators and nomically sub-optimal decisions. policy makers, including some suggestions for consumer education.

REFERENCES Lexington Books. Peracchio, Laura A. (1992), “How Do Young Children Berti, Anna and Anna Bombi (1988), The Child’s Con- Learn to be Consumers? A Script-Processing Ap- struction of Economics. New York: Cambridge Uni- proach,” Journal of Consumer Research, 18 (March), versity Press. 425–40. McNeal, James U. (1992). Kids as Customers. New York:

Marketing and Public Policy Conference Proceedings / 2009 283 Special Session: Developmental Perspectives on Children’s Response to Marketing Stimuli

For further information contact: Wendy A. Boland Kogod School of Business American University 4400 Massachusetts Ave, NW Washington, D.C. 20016 Phone: 202.885.1976 Fax: 202.885.2691 E-Mail: [email protected]

Lance-Michael Erickson Eller College of Management University of Arizona P.O. Box 210108 Tucson , AZ 85721 Phone: 520.626.4168 Fax: 520.621.7483 E-Mail: [email protected]

284 Marketing and Public Policy Conference Proceedings / 2009 Special Session: Developmental Perspectives on Children’s Response to Marketing Stimuli

LONG-TERM EFFECTS OF ADVERTISING TO CHILDREN ON JUDGMENT IN ADULTHOOD

Paul M. Connell, Stony Brook University Merrie Brucks, University of Arizona, Tucson Jesper Nielsen, University of Arizona, Tucson

SUMMARY (i.e., when motivation and ability to think are high). Based on this position, we challenge the change of meaning In this research, we argue that brand relationships that are assumption in the persuasion knowledge model. That is, formed in early childhood have unique properties that individuals may not be aware of their previous biases (i.e., make them especially prone to create attitudinal biases favorable attitudes toward a brand that were developed that favor them. This is because they were first experi- before the development of persuasion knowledge), and enced when the individual had limited ability to effec- subsequently do not correct bias. Similarly, individuals tively defend against advertisements. That is, children do may not be sufficiently motivated to correct bias. not understand advertisements in the same way that adults do. Rather, they incrementally gain knowledge about Given that such attitudinal bias in judgments may occur advertising as they age (John 1999). As children age and from these brand relationships, we investigate possible develop the ability to take the perspective of others, they public policy interventions to facilitate bias correction in consistently recognize the advertiser’s intent to persuade three experiments. Results from three experiments sug- them by age seven or eight (Ward, Wackman, and Wartella gest that: (1) emotion felt toward early childhood adver- 1977). tising objects, but not late childhood advertising objects, leads to biased judgments in the healthfulness of products In their persuasion knowledge model, Friestad and Wright with high sugar content, presumably due to age differ- (1994) argue that as individuals age and gain persuasive ences in persuasion knowledge when the advertising knowledge, they spontaneously correct for previous bias objects were encoded into memory, and (2) priming through what the authors call a “change of meaning.” health motivation or actively engaging cognitive defenses However, this argument was conceptual, discussed rather against advertising are effective only when individuals briefly, and was not supported with empirical evidence. experience lower levels of affect toward the advertising Petty and Brinol (2008) assert that bias correction is best objects. facilitated when individuals elaborate more on messages

REFERENCES of Children: A Retrospective Look at Twenty-Five Years of Research,” Journal of Consumer Research, Friestad, Marion and Peter Wright (1994), “The Persua- 26 (December), 183–213. sion Knowledge Model: How People Cope with Petty, Richard and Pablo Brinol (2008), “Persuasion: Persuasive Attempts,” Journal of Consumer Research, From Single to Metacognitive Processes,” Perspec- 21 (June), 1–31. tives on Psychological Science, 3 (2), 137–47. John, Deborah Roedder (1999), “Consumer Socialization

Marketing and Public Policy Conference Proceedings / 2009 285 Special Session: Developmental Perspectives on Children’s Response to Marketing Stimuli

For further information contact: Paul M. Connell College of Business Stony Brook University Stony Brook, NY 11794 Phone: 631.632.7139 Fax: 631.632.8181 E-Mail: [email protected]

Merrie Brucks Eller College of Management University of Arizona P.O. Box 210108 Tucson, AZ 85721 Phone: 520.621.3368 Fax: 520.621.7483 E-Mail: [email protected]

Jesper H. Nielsen Eller College of Management University of Arizona P.O. Box 210108 Tucson, AZ 85721 Phone: 520.621.4902 Fax: 520.621.7483 E-Mail: [email protected]

286 Marketing and Public Policy Conference Proceedings / 2009 AUTHORS

Adkins, Natalie Ross 8 Daniloski, Kim 98 Adkins, Randall E. 8 Davis, Brennan 165, 166 Ahuja, Bob D. 143 Day, Julie 194 Aikin, Kathryn 44, 46 Dermody, Janine 81 Amos, Clinton 200 Desrochers, Debra M. 59, 60 Andreasen, Alan 164 DeVincenzo, Marie Hafey 178 Andrews, J. Craig 62, 164 Devlin, James 94 Apirattanapimolchai, Apichai 146 Diamond, William D. 150 Austin, Lucinda 130 Droms, Courtney 202 Auty, Susan 21 Eisenstein, Eric M. 204 Baker, Stacey Menzel 55, 58, 192 Erickson, Lance-Michael 283 Bann, Carla 45 Evett, Sophia R. 221 Beales III, J. Howard 1 Fang, Xiang 256 Becker, Karen 139 Fein, Sara B. 40 Bell, Catherine 3 Finn, Adam 102 Berkman, Nancy D. 45 Fitzgerald, Tania 45 Bernstein, Joan Z. 1 Fitzsimons, Gavan J. 75 Berry, Leonard L. 273 France, Karen Russo 258 Block, Lauren G. 75 Frank, Joshua M. 217 Boeke, Abbe 132 Fuller, Dan A. 206 Boland, Wendy Attaya 279, 283 Garma, Romana 176 Bone, Michael 102 Gates, Rahima Jan 45 Bone, Paula Fitzgerald 258 Gendall, Philip 38, 239 Bone, Sterling A. 219, 223 Gorin, Daniel 3 Borders, Leila 55, 56 Gould, Stephen J. 136 Brinberg, David 98 Grant, Susan 48 Briones, Rowena 139 Grau, Stacy Landreth 176, 200 Brobeck, Stephen 48 Grier, Sonya 279 Brucks, Merrie 285 Grover, Aditi 174 Bui, My 118 Grube, Joel W. 36 Burton, Scot 62, 118 Grumbach, Kevin 194 Butler, Sam 194 Guion, Deirdre T. 55, 56, 262 Calfee, Jack 77 Hakstian, Anne-Marie G. 221 Campbell, Margaret C. 73 Hang, Haiming 21 Carlson, Les 59 Hanmer-Lloyd, Stuart 81 Chaplin, Lan Nguyen 281 Hastak, Manoj 1, 52, 53 Chaudhary, Sarita Ray 92 Hedderley, Duncan 239 Chavez, Pablo 48 Henderson, Geraldine R. 221 Choiniere, Conrad J. 40, 87, 88, 89 Hettche, Matt 10 Christensen, Glenn L. 223 Hibbard, Judith 45 Coble, Garrett 33 Hibbert, Sally A. 152 Cohen, Joel B. 164 Hilger, James 52 Connell, Paul M. 279, 285 Hill, Ronald Paul 55, 58, 164 Corus, Canan 79, 82 Hoch, Stephen 204 Costello, Ayse Olcay 110 Hoek, Janet 38, 196, 239 Costello, Thomas G. 110 Hogarth, Jeanne 3, 50, 90 Coupey, Eloise 253 Hogg, Margaret K. 152 Craciun, Georgiana 42 Hoofnagle, Chris Jay 245 Crosby, Elizabeth 120 Howlett, Elizabeth 118 Cude, Brenda J. 159 Hoy, Mariea Grubbs 243

Marketing and Public Policy Conference Proceedings / 2009 287 Huang, Yi 40 Miller, Tom 77 Huntley-Fenner, Gavin 155 Milne, George R. 48, 243 Irmak, Caglar 260 Mirabito, Ann M. 273 Jiménez, Fernando 33 Mish, Jenny 100 Jones, Lukas 104 Mitakakis, Chrissy 136 Jones, Sandra C. 229 Mitra, Anu 89, 227 Kahle, Lynn R. 104, 106 Mittelstaedt, John D. 64, 192 Kamins, Michael 174 Miu, Carol 108 Karniouchina, Ekaterina V. 206 Miyazaki, Anthony D. 161 Kaufman-Scarborough, Carol 96 Mohr, Gina S. 73 Keefe, Brian 139 Moore, Elizabeth S. 60 Kees, Jeremy 172, 258 Moore, William L. 206 Kemp, Elyria 124, 275 Mukherjee, Sayantani 134 Kendrick, Kathlene 77 Munro, Miles 155 King, Jennifer 245 Murphy, Patrick E. 1, 164 Kinnear, Thomas C. 164 Netemeyer, Richard G. 62 Klein, Thomas A. 148 Neuhauser, Linda 44, 47 Knowles, Isaac 53 Newman, Christopher 163 Kopp, Steven W. 118, 124, 163, 275 Niculescu, Ruxandra 96 Kozup, John 7, 84, 258 Nielsen, Jesper 285 Kulemeka, Owen 114 Nowlis, Stephen M. 168 Kumar, Poonam 253 O’Donoghue, Amie 44, 46 Labiner-Wolfe, Judith 40, 86 Ohene-Frempong, Janet 45 Laczniak, Gene 192 O’Lexy, Mark 96 Langenderfer, Jeff 256 Otnes, Cele C. 120 Lee, Evonne 161 Owen, Scott 104, 106 Lee, Joonwhan David 198 Ozanne, Julie L. 79, 192 Lee, Jung-Sook 126 Pappalardo, Janis 52 Lee, Seul 159 Patel, Sejal 130 Letzler, Robert 50, 53 Payne, Collin 167 Levy, Alan S. 84, 87, 88, 165 Pechmann, Cornelia 166 Limbu, Yam B. 92 Peck, Mary 253 Lin, Chung-Tung Jordan 241 Peeler, C. Lee 1 Lokken, Kristine 132 Pence, Karen M. 35 Louviere, Jordan 38 Perry, Vanessa Gail 198 Lowrey, Tina 281 Piacentini, Maria G. 152 Lubin, Paul C. 225 Pierce, Meghan 98 Lubran, Meryl 130 Pilcher, Kenneth 132 Lwin, May O. 161 Polonsky, Michael Jay 176 Lynch, Jr., John G. 35 Pounders, Kate 138 MacInnis, Deborah J. 65 Raish, Carol 55, 57 MacLeod, William C. 1 Raymond, Mary Anne 64 Mandel, Naomi 168 Rich, Jessica L. 48 Martin, Ingrid M. 55, 57, 134, 174 Riker, Elise 223 Martin, Wade E. 55, 57 Rodgers, Edwina 194 Mason, Marlys J. 33, 173, 256 Rosa, José Antonio 169 Maubach, Ninya 239 Russell, Cristel Antonia 36 Mazis, Michael B. 164 Russell, Dale W. 36 McAlvanah, Patrick 53 Saatcioglu, Bige 83 McCormack, Lauren A. 45 Santos, Nicholas 192 McKay-Nesbitt, Jane 141 Scammon, Debra L. 55, 56, 77, 164, 194, 206, 277 McLean, Rachael 196 Scarpaci, Joseph 10 Merry, Ellen A. 5 Schertzer, Clint S. 143 Michels, Tara A. 143 Schiffino, Ryan 96

288 Marketing and Public Policy Conference Proceedings / 2009 Schillinger, Dean 45 Vallen, Beth 75, 260 Schultheiss, Patricia 77 Venger, Olesya 112 Scott, Kristin 180 Verlegh, Peeter W.J. 73 Scott, Maura L. 168 Vitartas, Peter 146 Scullion, Richard 81 Vlietstra, Andrew D. 90 Seo, Joon Yong 277 Walker, Kristen L. 116 Shimshack, Jay P. 54 Walker, Mary M. 143 Shin, Dongwoo 42 Wan, Heather Zhong 65 Skubisz, Christine 130 Wansink, Brian 167 Smith, Karen H. 182, 264 Ward, Michael 54 Smith, Malcolm C. 141 Wiener, Joshua L. 173 Spanjol, Jelena 169, 273 Wijaya, Mindawati 161 Squiers, Linda 44, 45 Wilcox, Keith 75 Squire, Claudia 45 Wilkie, William L. 35 Stanaland, Andrea J.S. 161 Williams, Jerome D. 219, 221, 223 Stanton, Julie V. 262 Williams, Sally A. 264 Stauffer, Julien 98 Williamson, JoAnna 157 Subrahmanyan, Saroja 128 Wolburg, Joyce M. 175 Sullivan, Helen 46 Worthy, Sheri L. 132 Swasy, Jack 44, 46, 227 Zank, Gail M. 182 Tam, Leona 165, 169 Zhang, Jason Q. 42 Tanner, Jr., John F. 64 Zigas, Barry 35 Turner, Monique Mitchell 130 Zwicki, Todd J. 35

Marketing and Public Policy Conference Proceedings / 2009 289