Volume 23 The Changing Role of Policy in Consumer Well-Being · 2013 Marketing & Public Policy...

2013Marketing & Public Policy Conference

The Changing Role of Policy in Consumer Well-Being

Volume 23

May 30 – June 1, 2013 Washington D.C.

EDITORS: Linda Salisbury and Kathleen Seiders

“The Changing Role of Policy in Consumer Well-Being”

May 30–June 1, 2013Loews Madison Hotel

Washington, D.C.

Marketing and Public Policy

Conference Proceedings 2013

Editors:Linda SalisburyKathleen Seiders

Volume 23

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

ii

Copyright © 2013, American Marketing Association

Printed in the United States of America

Production Editor: Marilyn StoneComposition by Marie Steinhoff, CAP, Southeast Missouri State UniversityCover design by Kristina Walton

ISBN: 0-87757-353-0

All rights reserved. No part of the material protected by this copyright notice maybe reproduced or utilized in any form or by any means, including photocopying andrecording, or by any information storage and retrieval system without the writtenpermission of the American Marketing Association.

iii

2013 MARKETING AND PUBLIC POLICYCONFERENCE AWARDS

Best Paper in Conference“How Effective Are Alternative Plain Pack Designs?”

Philip Gendall, University of Otago and Massey UniversityJanet Hoek, University of Otago

Damien Mather, University of OtagoRichard Edwards, University of Otago

Brenda Derby Memorial Award“Be a Good Cookie: Cause-Related Marketing’s Influence on Health Halo Formation”

Elizabeth Minton, University of Oregon

The Brenda Derby Memorial Award is presented nationally to an outstanding doctoral student who demonstratesexcellence as an emerging policy researcher. The award honors a strong advocate of the public policy and marketingsubdiscipline and is supported through the generosity of the Center for Marketing and Public Policy Research atVillanova University.

iv

PREFACE

The theme of this year’s Marketing and Public Policy Conference (MPPC) is “The Changing Role of Policy inConsumer Well-Being.” Relating marketing and public policy to consumer well-being has been an enduring focus ofscholars and researchers, and currently, consumer well-being is at the heart of crucial global issues and initiatives. Inthe United States, for example, recent health care and financial regulation debates are highly relevant to consumer well-being and have key economic implications. Our goals for the 2013 conference are to feature critical issues at the nexusof consumer well-being and public policy and to provide scholars, policy makers, and other experts with interesting andapplicable insights.

This is MPPC’s 24th year—an inspiring statement of the marketing and public policy research community’scommitment. We defined our conference theme broadly to include multiple dimensions of well-being. Health care policyand consumer wellness, financial stability and decision making, environmental sustainability and regulation, andsocietal safety and protection all represent focal areas where well-being and policy intersect. Exciting research advanceshave been made in these areas, and compelling opportunities remain.

As the conference program and proceedings indicate, MPPC 2013 encompasses a range of topics spanning areasthat affect consumer well-being, including influencing prosocial behavior, environmental effects on food choices,tobacco marketing and smoking behavior, communicating risk and benefit information, pro-environmental marketing,recycling and repurposing, policy relating to consumer debt, Federal Trade Commission policy on advertising claims,diversity and policy research, financial decision making, maladaptive consumption, and front-of-package nutritionlabeling.

The conference has a diverse representation of scholars, authors, and panelists from academe, the public sector, andthe private sector. Increasingly, MPPC has gained a global orientation, with participants from institutions in Australia,Brazil, Canada, China, Colombia, Denmark, France, Germany, Hong Kong, Japan, New Zealand, Nigeria, Portugal,Turkey, and the United Kingdom. From the public sector, the program includes participants from the ConsumerFinancial Protection Bureau, the Federal Trade Commission, Centers for Disease Control and Prevention, and the U.S.Food and Drug Administration. Other organizations represented include the Ad Council, Center for Financial ServicesInnovation, GfK SE Health, Health Promotion Agency, Hudson Cook LLP, Prevention Research Center, Qintuit,Research Triangle Institute International, Tala Pasifika, and Whakauae Research Services.

We are indebted to Ron Hill, John Kozup, and the Villanova Center for Public Policy and Marketing Research fortheir sponsorship—including program planning and implementation—of the 2013 Marketing & Public Policy Work-shop and Doctoral Seminar. This preconference event showcased leading scholars in the field and facilitated theirinteraction with doctoral students and junior faculty interested in the public policy and marketing area. The programcovered methodological issues, research domains, publishing in the field, federal agency–based research, anddiscussions with journal editors. Villanova generously covered all costs for the attendees. This kind of support anddedication helps promote strong research standards and secure the future of our field.

This year’s conference and workshop would not have been possible without the support of many organizations andindividuals. We thank the planning group for valuable early ideas. We thank the American Marketing Association(AMA) for its support in MPPC scheduling, planning, logistics, and overall management of the conference. Weespecially thank Matt Weingarden, AMA Program Manager, for his wonderful work on the conference, his profession-alism and collegiality, and the consistently excellent service he provided to us throughout every stage of the process overthe past year.

Most important, we are very grateful to the conference reviewers for their hard work and responsiveness in helpingus process the 144 competitive paper and special session submissions we received. The MPPC reviewers truly are at theheart of the conference.

We acknowledge and thank the Marketing Science Institute, a loyal friend of MPPC, for its financial support. Wealso thank the AMA Academic Council and the Marketing and Society special interest group for ongoing contributions.Finally, we are most appreciative of our home institution, Boston College, Carroll School of Management, for significantfinancial support of the conference as well as for encouraging our service to the profession in this capacity.

Linda Salisbury and Kathleen SeidersBoston College

v

2013 MARKETING AND PUBLIC POLICYCONFERENCE REVIEWERS

Claudia Rosa AcevedoUniversity of São Paulo

Kathryn AikinU.S. Food and Drug Administration

Alan AndreasenGeorgetown University

Anshu Saxena AroraSavannah State University

Ioannis AssiourasESC Toulouse Business School

Courtney Nations BakerUniversity of Wyoming

Connie BatemanUniversity of North Dakota

Samantha BietschAmerican Military University

Melissa G. BublitzUniversity of Wisconsin Oshkosh

Les CarlsonUniversity of Nebraska–Lincoln

Jesse R. CatlinWashington State University Tri-Cities

Angela ChangCentral Michigan University

Elizabeth CrosbyUniversity of Wisconsin La Crosse

Catharine CurranUniversity of Massachusetts Dartmouth

Prakash DasUniversity of Calgary

Brennan DavisBaylor University

Cassandra DavisUniversity of Arkansas

Yoshiko DeMottaFairleigh Dickinson University

Debbie DesrochersUniversity of Westminster

Zhirong DuanTsinghua University

Akon Elizabeth EkpoWestern Michigan University

Pam Scholder EllenGeorgia State University

Jodie L. FergusonVirginia Commonwealth University

Margaret Paula FitzgeraldWest Virginia University

Andrea Godfrey FlynnUniversity of San Diego

George FrankeUniversity of Alabama

Henrik HagtvedtBoston College

Janet HoekUniversity of Otago

Michael R. HymanNew Mexico State University

Pauline M. IppolitoU.S. Federal Trade Commission

Djavlonbek KadirovEastern Institute of Technology

Andrew KaikatiUniversity of Georgia

vi

Carol Kaufman-ScarboroughRutgers University–Camden

Jeremy KeesVillanova University

Rohit KhannaIn Vivo

Leslie KoppenhaferUniversity of Oregon

Christine KowalczykEast Carolina University

Ohjin KwonConcordia University

Kenneth K. KwongHang Seng Management College

Monica LaBargeQueen’s University

Jeff LangenderferMeredith College

Yam B. LimbuMontclair State University

Kelty LoganUniversity of Colorado at Boulder

Michael LuchsCollege of William and Mary

Ingrid Maria MartinCalifornia State University Long Beach

Jane McKay-NesbittBryant University

James MeadUniversity of Kentucky

Jessica MikeskaUniversity of Nebraska–Lincoln

Elizabeth Gelfand MillerUniversity of Massachusetts Amherst

Rohan Miller University of Sydney

Elizabeth MintonUniversity of Oregon

Chrissy MitakakisIona College

Dick MizerskiUniversity of Western Australia

Fred MorganWayne State University

Jill MostellerPortland State University

Iman NaderiUniversity of North Texas

Gergana NenkovBoston College

Mauro Jose OliveiraCentro Universitário da FEI

Nitha PalakshappaMassey University

Ross PettyBabson College

Marcus PhippsUniversity of Melbourne

Dante M. PirouzWestern University

Rob RichersonUniversity of Kentucky

Spencer M. RossUniversity of Massachusetts Amherst

Chryssoula RouvakiAthens University of Economics and Business

Cristel RussellAmerican University

Linda Court SalisburyBoston College

Kristin ScottMinnesota State University, Mankato

vii

Maura L. ScottUniversity of Kentucky

Kathleen SeidersBoston College

Robin SosterUniversity of Arkansas

Wenbin SunRockhurst University

Remi TrudelBoston University

Beth VallenFordham University

Eric Van SteenburgUniversity of North Texas

Richard VannUniversity of Wyoming

Kristen WalkerCalifornia State University Northridge

Michael Francis WalshWest Virginia University

Carla WalterCalifornia Lutheran University

S. Todd WeaverPoint University

Josh WienerOklahoma State University

Jerome Denean WilliamsRutgers University

viii

TABLE OF CONTENTS

CONFERENCE AWARDS iii

PREFACE iv

MANUSCRIPT REVIEWERS v

TABLE OF CONTENTS viii

INFLUENCING CONSUMER GOALS AND PROSOCIAL BEHAVIOR

Charitable Giving to Controllable Misfortunes: The Role of Deliberation and VictimIdentifiability

Yoshiko DeMotta, Sankar Sen, Stephen Gould 1

Consuming Functional Innovations: Are Utilitarian Behaviors Enhanced orUndermined?

Aaron M. Garvey, Lisa E. Bolton 3

Misforecasting Positive Affective Responses to Prosocial BehaviorStefanie Rosen Robinson, Caglar Irmak, William O. Bearden 5

Predicting Preventive Health Behavior: The Effects of Self-Prophecy and Self-EfficacyCarmen-Maria Albrecht, Hans H. Bauer, Kai Bergner, Tamara Gogia,David E. Sprott 7

CONSUMER RESPONSES TO HOW PRODUCT ATTRIBUTES ARECOMMUNICATED

That’s Not How I Remember It: Willfully Ignorant Memory for Ethical Product AttributeInformation

Rebecca Walker Naylor, Julie R. Irwin, Kristine R. Ehrich 9

When Might Consumers Choose Superior Sustainability Despite a Trade-Off with OtherProduct Attributes?

Michael G. Luchs, Minu Kumar 11

Be a Good Cookie: Cause-Related Marketing’s Influence on Health Halo FormationElizabeth Minton 13

Translated Attributes as a Choice Architecture Tool: Trick and TreatAdrian R. Camilleri, Christoph Ungemach, Richard P. Larrick, Elke U. Weber,Eric J. Johnson 15

POLICY AND ENVIRONMENTAL EFFECTS ON THE HEALTHFULNESS OFFOOD CHOICES

The Unintended Consequences of Calorie Information on Menus: The Use of Attractionand Compromise Effect

Ryall Carroll, Beth Vallen 17

ix

Environmental Cues and Food Consumption Among Dieters and Non-DietersBeth Vallen, Chrissy Mitakakis 19

The Impact of Food Groups, Dietary Variety and Quality, and Demographics on DailyCaloric Intake

Adwait Khare, J. Jeffrey Inman, Chung-Tung Lin, WenYen Juan 21

Conjoint Analysis of the Consumer Trade-Off Decisions Associated with aBeverage Tax

Jess Mikeska, Les Carlson, Dwayne Ball 23

PROMOTING HEALTH AND QUALITY OF LIFE FOR SPECIFIC CONSUMERSEGMENTS

Restorative Senior Center Servicescapes and Senior HealthMark S. Rosenbaum, Jillian Sweeney 25

Transportation Studies Reveal Barriers for Persons with Disabilities: Using theConcept of Continuous Sequence

Carol Kaufman-Scarborough 27

The Tortoise and the Hare Effect: Aged “Marathoners” Best Youngster “Sprinters”When Learning Experientially

Ashley Goerke, Eric M. Eisenstein, Ayalla Ruvio 29

THE EVOLVING DYNAMICS OF SOCIAL MEDIA AND INTERACTION: ISSUESAND OPPORTUNITIES

Adolescent Bullying Bias and Negative Consumption: The Moderating Roles of SocialActivity, Teacher Support, and Long-Term Goals

Brennan Davis 31

Private Information in a Social World: A Policy Perspective of Consumer Privacy onSocial Networking Sites

Alexa K. Fox, Marla B. Royne 32

The Role of Online Talking in Achieving Health Behavior ChangeBo Liang, Debra L. Scammon 34

Alcohol Culture: The Effects of Social Media in Relation to Alcohol ConsumptionGwyneth Howell, Rohan Miller, Dijana Apostolovski 36

CURRENT INVESTIGATIONS OF TOBACCO MARKETING AND SMOKINGBEHAVIOR

Identifying the Latent Effects of Selecting Samples Using Extreme Responses: LESSERAlvin Lee, Dick Mizerski, Jinchao (Alex) Yang, Shasha Wang 38

The Effect of Political Ideology on Reactions to Warning LabelsMitchel R. Murdock, Caglar Irmak, James F. Thrasher 40

How Effective Are Alternative Plain Pack Designs?Philip Gendall, Janet Hoek, Damien Mather, Richard Edwards 42

x

COMMUNICATING RISK AND BENEFIT INFORMATION TO CONSUMERS

Do Dissonant Ad Visuals Cause Consumers to Discount Prescription Drug Side Effects?David J. Hagenbuch 44

The Presentation of Quantitative Benefit Information in Direct-to-ConsumerTelevision and Print Advertisements for Prescription Drugs

Helen W. Sullivan, Amie C. O’Donoghue, Kathryn J. Aikin, Dhuly Chowdhury,Rebecca R. Moultrie, Douglas J. Rupert 46

No Good Deed Goes Unpunished: Citing Sources in Print Advertising ClaimsLeslie Koppenhafer, Catherine Armstrong Soule, Robert Madrigal 48

The Effect of Health Risk Presentation Format on Consumer Perceptions and ChoiceJesse R. Catlin, Elizabeth Gelfand Miller, Ann M. Mirabito 58

THE IMPACT OF PRO-ENVIRONMENTAL MARKETING ON CONSUMERPREFERENCES

Eco-Labeling: Approach or Avoidance?M. Paula Fitzgerald, Joshua D. Dorsey 60

Scale and Metric Design as Choice Architecture ToolsAdrian R. Camilleri, Richard P. Larrick 62

The Role of Centrality of Green Attributes in Green Product EvaluationAndrew D. Gershoff, Judy K. Frels 64

An Exploratory Study of Behavior Placements in Television SitcomsChristine M. Kowalczyk, Jennifer Martinez 66

RECYCLING AND REPURPOSING DECISIONS

An Empirical Study of the Determinants of Green Shopping Behavior in BrazilHelenita R. da Silva Tamashiro, José Augusto Giesbrecht da Silveira,Claudia Rosa Acevedo 68

Bringing It Back to Life: Repurposing Products to Make Them New AgainS. Todd Weaver, Kristin Scott 70

Crossing the Recycling Bin Boundary: The Effect of Product Distortion and Size inConsumer Disposal Decisions

Remi Trudel, Jennifer Argo 72

When Environmental Messages Should Be Assertive: The Moderating Effect of EffortInvestment on Consumers’ Recycling Intentions

Tae Hyun Baek, Sukki Yoon, Seeun Kim 74

CONSUMER DEBT AND PUBLIC POLICY

Ethics Audit Reveals “Little Dutch Boy” Policy Leaks: Credit Card Act (2009) FailsCollege Students

Connie R. Bateman, Mary K. Askim-Lovseth 76

xi

Motivating Consumers to Get Out of Debt FasterKeri L. Kettle, Remi Trudel, Gerald Häubl 78

Susceptibility to the Credit Card Effect: A New Concept and Measurement ScaleSandra Awanis, Charles C. Cui 80

APPROACHES TO DISCOURAGING RISKY BEHAVIORS

Temptation and Prevention Provided by the Gambling Industry: Main and InteractiveEffects on Gamblers

Linda Lemarié, Jean-Charles Chebat 82

The Relationship Between Social Marketing Policies on Public Health andEpidemiological Situations: An Exploratory Research Based on TelevisionAdvertisements Preventing HIV/AIDS in France, Germany, Portugal, and Italy

Beatriz Casais, João F. Proenca 85

Positive Marketing to Promote Healthy Foods and Advance Food Well-BeingMelissa G. Bublitz, Laura A. Peracchio 87

Choice, Control, and the Paradox of Smoking During PregnancyJanet Hoek, Ninya Maubach, Heather Gifford, Stephanie Erick,Richard Edwards, Rhiannon Newcombe 89

MARKETING PRACTICE AND INDUSTRY SELF-REGULATION

Is the Price Right? Policy Implications Associated with Scanner Data InaccuracyDavid M. Hardesty, Anthony Miyazaki, Ronald Charles Goodstein,Dhruv Grewal, Praveen Kopalle 91

Product Content Ratings in the Entertainment IndustryJoseph Matthes, Les Carlson 92

The Political CEO: An Event Study Comparing Consumer Attributions of CEO BehaviorJessica Mikeska, Elise Johansen Harvey 94

AUTHOR INDEX 96

Marketing and Public Policy Conference Proceedings / 2013 1

CHARITABLE GIVING TO CONTROLLABLE MISFORTUNES: THEROLE OF DELIBERATION AND VICTIM IDENTIFIABILITY

Yoshiko DeMotta, Fairleigh Dickinson University, MadisonSankar Sen, Baruch College, CUNY, New York

Stephen Gould, Baruch College, CUNY, New York

SUMMARY

Causal controllability is the degree to which an observerof another person’s misfortune perceives that misfortuneto be the fault or responsibility of the person in need ofhelp. When a person encounters another in need of help,that person spontaneously judges the causal controllabil-ity of the situation (Weiner 1985). When the need iscaused by an uncontrollable circumstance (e.g., geneticillness), sympathy is elicited, and the perceiver is likelyto help the victim; in contrast, when the need is caused bya controllable circumstance (e.g., drunkenness), angerand blame are evoked, and the perceiver is likely towithhold help giving (Weiner 1980). Based on this theo-ry, a nonprofit organization’s donation request is likely tobe more effective if it communicates uncontrollability ofthe need; that is, the victim’s plight is not caused byirresponsibility on the part of the victim. However, socialcauses that nonprofit organizations attempt to aid are notalways uncontrollable. For example, causes having amental-behavioral origin such as drug abuse and obesityare generally considered to be controllable, and accord-ingly it is difficult to solicit donations for those causesbecause they tend to elicit potential helpers’ anger andblame toward the victims (Weiner, Perry, and Magnus-son 1988). How can we decrease blame toward the victimof the controllable misfortune and increase people’swillingness to help the victim? Drawing on the literatureof the identifiable victim effect (Small and Loewenstein2003) and the feeling- versus deliberation-based mode ofprocessing (Hsee and Rottenstreich 2004), we proposethat a charitable request that excludes the victim’s per-sonal information and evokes a potential donor’s deliber-ation will increase donations when a cause is construed ascontrollable.

The identifiable victim effect refers to people’s greaterhelpfulness towards a personalized, single victim (i.e.,the victim’s information such as a photo and name isavailable for perceivers), compared to an abstract, uni-dentified victim (Kogut and Ritov 2005; Small and Loe-wenstein 2003). Research has shown that a perceiver’saffective reaction, whether it is sympathy or anger, is

stronger toward an identified victim than toward anunidentified victim (Small and Loewenstein 2005). Thus,in order to lessen a perceiver’s negative reaction towardthe victim of a controllable misfortune, exclusion of thevictim’s personal information will result in not only a lessnegative reaction, but also in an increased deliberation ofthe controllable misfortune. If a potential donor consid-ers a misfortune within the context of a broader scope(Hsee and Rottenstreich 2004), that donor is likely toperceive the misfortune as stemming from a societalfailure rather than from an individual failure (Piff et al.2010), and feel more sympathy toward the victim, lead-ing to more help giving. Our three studies support thisprediction.

This research makes several theoretical and practicalcontributions. First, this is the first study that examinesthe identifiable victim effect from the perspective of thedifferent levels of causal controllability of a misfortune.By doing so, we show that the conventional belief in theidentifiable victim effect is only true when people per-ceive the misfortune to be uncontrollable. Second, thisresearch identifies a condition in which deliberationincreases philanthropic impulses, which is new to theliterature of charitable giving. Third, this paper contrib-utes on a practical level to nonprofit organizations thatsupport controllable misfortunes by suggesting commu-nication methods that reduce the negative evaluationsfrom the perception of high controllability without losingthe opportunity to communicate preventability of themisfortune.

H1: People are more likely to donate to help an uniden-tified victim of a controllable misfortune than anidentified victim of the same misfortune when peo-ple’s deliberation is high. Victim identifiability willnot affect people’s donation decisions when theirdeliberation is low.

H2: Sympathy will mediate the effect of victim identifi-ability and deliberation on donations to a controlla-ble misfortune.

2 Marketing and Public Policy Conference Proceedings / 2013

Three experiments examined and supported our hypoth-eses. Study 1 tested hypothesis 1 and found that anunidentified victim increases donations to a controllablemisfortune when potential donors’ degree of deliberationhigh (vs. low). In addition, the same study replicated theprevious finding of Small et al. (2007) and found that anidentified victim increases donations to an uncontrolla-ble misfortune when potential donors’ degree of deliber-ation is low (vs. high). Study 2 examined the mediatingrole of sympathy in the effect of identifiabilty and delib-eration on donations for a controllable misfortune (hy-pothesis 2). Study 3 used a different method of manipu-lating the degree of deliberation and replicated the effectsfound in studies 1 and 2.

Presenting personalizing information of a victim of amisfortune and stimulating people to sympathize with thevictim often increases their desire to help that victim. Tothis extent, persuasion techniques that directly increasesympathy have a positive effect. As our findings indicate,however, additional considerations arise when aid issolicited in the context of a charitable appeal that aims tohelp a victim of a misfortune that is perceived as control-lable. Our studies demonstrate that the characteristics ofan appeal that encourage people to feel sympathetic withthe identified victim can be detrimental to its effective-ness. Instead, excluding the victim’s personal informa-tion and stimulating people’s deliberation can effectivelyincrease donations to help victims of controllable misfor-tunes. References are available upon request.

For further information contact:Yoshiko DeMotta

Fairleigh Dickinson University285 Madison AvenueMadison, NJ 07940

Phone: 973.443.8295Fax: 973.443.8870

E-Mail: [email protected]


CONSUMING FUNCTIONAL INNOVATIONS: ARE UTILITARIANBEHAVIORS ENHANCED OR UNDERMINED?

Aaron M. Garvey, The University of Kentucky, LexingtonLisa E. Bolton, The Pennsylvania State University, University Park

SUMMARY

A frequent practice in marketing is the introduction ofproducts with new functionality; that is, a product ismodified with attributes that afford consumers with addi-tional opportunities for action (Dourish 2001; Ziamouand Ratneshwar 2003). Such innovations in functionalitymake a novel set of benefits available that support one ormore functional outcomes, even if the base product orfunctional components are not themselves new to themarket (Alexander, Lynch, and Wang 2008; Hoeffler2003). Examples include functional foods (e.g., omega-3 ice cream) and green products (e.g., natural detergents)that provide benefits over their traditional counterparts,and which represent substantial portions of their respec-tive consumer markets (Price Waterhouse Coopers 2009).

The present research examines how functional innovationsmay unexpectedly and ironically undermine the veryoutcomes their novel functionality is intended to support.Although prior research has deeply studied factors whichinfluence preference for (Hirschman and Holbrook 1982;Naylor, Droms, and Haws 2009) and categorization of(Moreau, Markman, and Lehmann 2001; Moreau,Lehmann, and Markman 2001) functional innovations,research is sparse examining the downstreamconsequences of consuming these products. Extantresearch has demonstrated that evaluative processing ofnovel innovations results in a heightened focus uponincremental benefits versus traditional products(Alexander, Lynch, and Wang 2008; Meyers-Levy andTybout 1989). We propose that this “functional focus”may influence subsequent consumption patterns. Priorresearch has demonstrated that explicit self-perceptionsof utilitarian behavior have the tendency to subsequentlyincrease hedonic behaviors through a variety ofmechanisms, including reducing perceived risk (Bolton,Cohen, and Bloom 2006), guilt reduction, (Wansink andChandon 2006), executive resource depletion (Vohs andHeatherton 2000), and utilitarian goal-release (Laran,Janiszewski, and Cunha 2008).

Our research indicates that consumption of functionalinnovations subsequently increases hedonic-seekingbehaviors which undermine utilitarian outcomes. Ourresults are consistent with a process of passive goal-guidance (Laran and Janiszewski 2009), in which con-sumption of a functional innovation releases relevantutilitarian goals, followed by involuntary emergence ofconflicting hedonic desires (Carlson, Meloy, and Miller2013) that subsequently guide perception and behavior.For example, consumption of functional foods (eco-products) releases health (environmental) goals, therebyinstigating a hedonic state that increases engagement inless healthy (environmentally responsible) consump-tion. The result is a so-called hedonic “boomerang”that undermines overall utilitarian outcomes and thefunctional benefits of the innovation. Furthermore, wepropose a method to interrupt this process and therebymitigate the “boomerang” by priming a utilitarian goalbeforehand—which shields (Chartrand et al. 2008) againstthe release of the functional goal.

A series of real behavior and scenario studies supports theproposed model and intervention, while ruling out alter-native explanations. Study 1 examined the impact ofconsuming a functional innovation upon subsequent he-donic-utilitarian behavior. An environmental innovationin the technology domain (smartphones) was employed.Following purchase of a functionally innovative (“withEco-Safe battery technology”) or non-innovative smart-phone, participants indicated choice likelihoods for mul-tiple phone application pairs containing one hedonicallyoriented and one utilitarian oriented option. As predicted,participants purchasing the functional innovation indi-cated increased intentions to purchase hedonic apps.

Studies 2A-B were conducted to (i) provide direct pro-cess evidence for a heightened hedonic and suppressedutilitarian state following consumption of a functionalinnovation distinct from conscious choice, and (ii) dem-onstrate that a similar “hedonic boomerang” effect doesnot occur for traditional, non-innovative products that


provide comparable utilitarian benefits, thereby (iii) rul-ing out conscious “justification” process explanationssuch as guilt reduction (Wansink and Chandon 2006), orconscious licensing due to inferences of goal progress(Fishbach and Dhar 2005). Study 2A directly measuredinvoluntary activation of a heightened hedonic state dueto functionality without soliciting conscious choice. Instudy 2A, participants first imagined purchasing either anenvironmentally responsible (washable cotton towel) orirresponsible (disposable paper towels) product that waseither functio nal (“with Eco-Nurture additives”) or non-functional. To measure activation of a hedonic state,participants were next exposed to a 15x18 letter matrixcontaining an equal number of utilitarian (e.g., “re-sponsible”) and hedonic (e.g., “exciting”) words. Aspredicted, functionality significantly increased (de-creased) the number of hedonic (utilitarian) wordsfound in the puzzle, regardless of the base producttype. These results are consistent with activation of ahedonic state that guides attention outside of con-scious awareness. Furthermore, these results precludeexplanations for behavioral shifts due to conscious guilt-reduction or “balancing” mechanisms, which proposethat conscious justification of prior behavior is used asa “license” to engage in hedonic behaviors. Study 2Bemployed the same design and initial stimulus, butmeasured behavioral intentions through choice likeli-hood ratings of environmentally responsible and irre-sponsible options. Consistent with the process find-ings of 2A, functional innovations (regardless of basetype) subsequently decreased preference for environ-mentally responsible options.

Study 3 explored in a real-behavior context whetherconsumption of a functional innovation undermines util-itarian behavior. As part of a taste test, participants firstconsumed a real pre-packaged cookie labeled as func-tional (“with antioxidants”) or non-functional, then spentfifteen minutes viewing movie trailers and eating actualM&Ms. Consistent with our predictions, individuals con-suming a functional innovation subsequently ate moregrams of the candy. Study 4 replicated this effect with(non-) functional innovation versions of a “healthy” baseproduct, yogurt. Finally, Study 5 provided further evi-dence for the proposed process by demonstrating thatprior activation of a utilitarian (vs. hedonic) goal miti-gates the effect of functionality upon subsequent con-sumption. Thus prior utilitarian goal activation inocu-lates consumers and introduces a method for “undoing”the functional boomerang.

In summary, this research is, to our knowledge, the firsteffort to understand the downstream consequences ofconsuming functional innovations. Consumption of afunctional innovation intended to support utilitarianoutcomes is demonstrated to subsequently underminethese outcomes. This unintended consequence occursbecause the functional innovation instigates a hedonicstate resulting from the release of functionally-relatedgoals. Beyond theoretical contributions to the innovationand motivation literatures, this work has substantiveimplications for managers and consumer welfare, partic-ularly given the focus of multiple industries (e.g., mobiletechnology, health food) upon product hybridization as asource of growth. References are available upon request.

For further information contact:Aaron M. Garvey

Marketing455U Gatton College of Business and Economics

The University of KentuckyLexington, KY, 40506Phone: 859.257.2869



MISFORECASTING POSITIVE AFFECTIVE RESPONSES TOPROSOCIAL BEHAVIOR

Stefanie Rosen Robinson, North Carolina State University, RaleighCaglar Irmak, University of South Carolina, Columbia

William O. Bearden, University of South Carolina, Columbia

SUMMARY

Consumer giving represents 75% of all charitable contri-butions (Giving USA 2010), and while these numbersindicate that consumer giving is a major industry, con-sumer giving has steadily declined over the last few years(Giving USA 2010; Department of Labor 2011). Also,when consumers do give, approximately half are moti-vated to only give once and not give again (NonprofitTimes 2003). In order to continue to serve those inneed, charities must effectively request help fromtarget donors. One way in which charities can beeffective in their requests is to understand consumers’feelings before and after they engage in prosocialbehavior. More importantly, charities need to under-stand whether any differences arise between consum-ers’ predicted and experienced emotions and how thisimpacts future help toward the charity.

There is an implied assumption that consumers are awareof the rewards, or positive affect, they will feel whenhelping others (Sprecher et al. 2007). However, researchhas not addressed the extent to which consumers areaccurate in their forecasts or the factors that influencetheir emotional estimations. This is important to under-stand, because as shown in this research, inaccurateforecasts impact consumers’ likelihood to engage inrepetitive charitable giving. Overall, charities can bemore effective in their requests if they understand con-sumers’ feelings surrounding the prosocial behavior.

Affective misforecasting is the difference between fore-casted and experienced affect (Patrick, MacInnis, andPark 2007). While prior research has demonstrated thedifferent mechanisms that can lead to affective misfore-casting (e.g., Dunn et al. 2004; Gilbert et al. 1998), it hasnot explained why misforecasting occurs with regard tohelping others. We argue that the focus of costs versusbenefits differs when people think about helping behav-ior versus when they engage in such behavior. There aretwo main components to prosocial behavior, the benefitto others and the costs incurred to the self. There arevarious ways in which a future event can be envisioned,either focusing on the process (the steps necessary for the

event to occur) or on the outcome (the end result that willbe achieved) (Taylor et al. 1998). When consumers thinkabout helping others, consumers focus on the outcome oftheir help (i.e., “others”); whereas, when consumersengage in the helping behavior, focus shifts to the process(i.e., costs to “self”) involved in helping others. We showthat this difference causes consumers to overestimatetheir positive affective responses to prosocial behavior(Studies 1 and 2).

Since consumers overestimate their positive affect whenthey forecast helping (i.e., experienced affect is less thananticipated affect), they are likely to be disappointedwhen their experienced affect does not match their antic-ipated affect. When consumers do not forecast theiraffective responses, they do not have a comparison pointto use against their experienced affect and therefore donot have an opportunity to be let down. Thus, consumers’experienced affect is expected to be higher when con-sumers do not forecast their emotions (vs. forecast theiremotions). If consumers have a positive experience help-ing others, they may be inclined to help again in the future(Jenner 1982). Therefore, not forecasting (vs. forecast-ing) one’s positive affect will increase consumers’ will-ingness to engage in future helping behavior (Study 3).

The next question we ask is whether consumers are everaccurate in predicting their positive affect? While ineveryday helping behaviors, donors are not required toput in much effort or costs, there are instances whenconsumers do engage in more costly helping behaviors,such as building homes for Habitat for Humanity, givingblood, and donating organs. All helping requires somecosts on the part of the giver (Bartlett et al. 2002), forexample, one’s time, money, and/or effort (Dovidio et al.1991). With costly helping behavior, the help itself ismore intense (Piliavin et al. 1981) and therefore theprocess of helping others, i.e., the cost involved to theself, is apparent both when thinking about and engagingin prosocial behavior. Therefore, consumers are lesslikely to misforecast their emotions when the cost ofhelping is high. We show that high cost helping providesa boundary condition to the misforecasting effect associ-ated with prosocial behavior (studies 4 and 5).


Across five studies, results provide evidence that con-sumers misforecast their positive affective responses inregard to prosocial behavior. With general everydayhelping behavior, consumers overestimate their positiveaffective responses. This overestimation is due to con-sumers’ perception of greater benefit to others when theythink about versus engage in the behavior. Furthermore,overestimating positive affect decreases consumers’ like-lihood to help the charity again in the future. A boundarycondition for this effect shows that, with costly helping,misforecasting diminishes.

While Mother Teresa said, “Nothings makes you happierthan when you reach out in mercy to someone who isbadly hurt” (Myers 1992, p. 194), evidence is providedhere to show that sometimes the anticipation of helpingothers will make you happier (than the act of helpingitself). Understanding whether misforecasting of posi-tive affect occurs in regard to prosocial behavior, why itis happening, and how to fix misforecasting errors canprovide benefits to consumers and prosocial causes, andin the end, society as a whole. References are availableupon request.

For further information contact:Stefanie Robinson

North Carolina State University2801 Founders DriveRaleigh, NC 27695

Phone: 919.515.6953Fax: 919.515.2120



PREDICTING PREVENTIVE HEALTH BEHAVIOR: THE EFFECTSOF SELF-PROPHECY AND SELF-EFFICACY

Carmen-Maria Albrecht, University of Mannheim, GermanyHans H. Bauer, University of Mannheim, Germany

Kai Bergner, GfK SE, Health, GermanyTamara Gogia, University of Mannheim, Germany

David E. Sprott, Washington State University, Pullman

SUMMARY

One main objective of marketing research in the field ofhealth care is to develop mechanisms to promote behav-iors and lifestyles of people that maximize their longterm health. In this regard, the main task is to increasepeople’s adherence to different prevention- and treat-ment-related regimens (Kazarian and Evans 2001).Different approaches have been developed to help toinfluence health behaviors such as the theory of plannedbehavior (Ajzen 1991; Fishbein and Ajzen 1975), thehealth belief model (Becker 1974; Janz and Becker1974), and the protection motivation theory (Madduxand Rogers 1983; Rogers 1983), to name only a few.However, when health care professionals try to imple-ment these frameworks in practice, they face the chal-lenge that these frameworks are rather complex and noteasily applicable. In addition, these frameworks often failto provide practitioners with specific procedures on howto alter health behaviors (Sprott et al. 2006).

Self-prophecy, however, can be regarded as a simple wayto influence health-related behaviors and can easily betransferred to practice (Sprott et al. 2006). The self-prophecy effect suggests that asking people to predictwhether they will perform a socially normative behaviorleads to the increased likelihood of them performing thatbehavior.

While self-prophecy research demonstrated its theoreti-cal explanation to be based on the theory of cognitivedissonance (Spangenberg et al. 2003) and found somemoderators of the self-prediction effect, it still requiresidentification of further conditions under which self-prediction is most effective (Sprott et al. 2006).

Since self-efficacy has generally been shown to influ-ence health behaviors (e.g., Jayanti and Burns 1998;Keller 2006), self-efficacy is such a potential factor.Except for one study (Longstreet et al. 2011) which testedthe self-prediction effect on computer-related task per-formance, the question of interaction between self-proph-ecy and self-efficacy has not yet been empirically studiedin prior research.

The current research investigates self-prophecy in thehealth behavior context to accomplish three objectives:(1) to provide evidence that self-prophecy has a signifi-cant effect on performance of a preventive health behav-ior, (2) to show that self-efficacy positively affects pre-ventive health behavior, and (3) to test whether self-efficacy moderates the self-prophecy effect on the focalbehavior.

The results of an experiment (n = 267) in the context ofsports behavior indicate that self-prophecy has an effecton preventive health behavior, thus providing furtherevidence that making a self-prediction changes people’sbehavior. Moreover, we are able to show that this effectis moderated by a person’s level of self-efficacy. Inparticular, people low in self-efficacy are more likely torespond to the effect of self-prediction than are peoplehigh in self-efficacy. This outcome suggests that the self-prophecy question can be particularly helpful for peoplewith lower self-efficacy beliefs in bringing their level ofexercising, for example, to the same level as people withhigher self-efficacy beliefs who actually do not need aself-prediction question. References are available uponrequest.


For further information contact:Carmen-Maria AlbrechtUniversity of Mannheim

L 9, 1Mannheim, 6813

GermanyPhone: +49.621.181.1573

Fax: +49.621.181.1571E-Mail: [email protected]


THAT’S NOT HOW I REMEMBER IT: WILLFULLYIGNORANT MEMORY FOR ETHICAL PRODUCT

ATTRIBUTE INFORMATION

Rebecca Walker Naylor, The Ohio State University, ColumbusJulie R. Irwin, The University of Texas at Austin

Kristine R. Ehrich, University of San Diego

SUMMARY

Consumers often encounter products in the marketplacethat have implications for larger ethical issues, such asproducts that have negative environmental consequencesor products that are manufactured in a way that is unfairto workers. When they encounter this type of informa-tion, consumers may be motivated to use coping strate-gies to avoid the emotional costs of incorporating knowl-edge about poor performance on these types of ethicalattributes into their decisions (e.g., Eckhardt, Belk, andDevinney 2010; Luce 1998). An especially direct way toavoid ethical issues in purchase situations is simply to notgather ethically-relevant product information. As Ehrichand Irwin (2005) have shown, consumers sometimesactively avoid searching for information about ethicalattributes. This “willful ignorance” is especially commonamong consumers who care about the underlying ethicalissue. This avoidance is particularly notable because theconsumers who avoid the information will actually usethis information in their decision making if it is given tothem.

Given that even consumers who avoid ethical attributeinformation often use it when it is available, a marketresearcher or brand manager might conclude from thisprior research that the solution to willful ignorance (i.e.,the key to getting consumers to incorporate ethical infor-mation into their purchase decisions) is simply to provideethical attribute information to the consumer. However,many psychological processes need to occur betweenattribute exposure and purchase and post-purchase eval-uations. Like many judgments and decisions, consumerdecisions can be made “online,” with all of the relevantinformation present in the decision environment whilethe decision is being made, or they can be memory-based,with some or all of the relevant information retrievedfrom memory at the time the decision is made (Hastie andPark 1986).

The present research explores the idea that simply pro-viding ethical attribute information to consumers is not

enough to ensure that this information is used becauseany role that memory plays in subsequent processing ofthe information could affect the mental representationof the product. Even if consumers are presented withethical attribute information, they may still not incor-porate information about poor (vs. good) performanceon an ethical attribute into their decisions. Instead, theyare motivated to systematically forget this information asa coping strategy to avoid the emotional costs of thisinformation. Thus, their memories for the informationwill be inaccurate, in a predictably biased direction.

Drawing on the literature on the malleability of memory(e.g., Braun 1999; Braun, Ellis, and Loftus 2002;Cowley 2008; Cowley and Janus 2004; Loftus 1979)and motivated reasoning (e.g., Kunda 1990; Santioso,Kunda, and Fong 1990), we predict (and demonstrateacross three studies) that consumers exhibit a particulartype of willfully ignorant memory that is unique to ethicalattributes. This willfully ignorant memory leads consum-ers to fail to recall poor performance on an ethicalattribute (e.g., that a desk is made from non-sustainablerainforest wood) at a higher rate than they do goodperformance on the same attribute (e.g., that a desk ismade from sustainable tree farm wood). The malleabilityof memory, we suggest, therefore protects consumersfrom having to process unsettling ethical attribute infor-mation because they simply forget that the product isunethical. Thus, in the service of protecting themselves,consumers develop inaccurate memories about the ethi-cality of the products they encounter.

We call this motivated forgetting willfully ignorant mem-ory and hypothesize that consumers engage in it as acoping strategy. If consumers can disregard poor perfor-mance on the ethical attribute, either by forgetting itentirely (i.e., not being able to recall any informationabout that attribute) or by remembering it incorrectly(i.e., incorrectly recalling performance on the ethicalattribute as good when it is actually poor), then a sense ofhelplessness about solving the ethical issue in questioncan be avoided. We predict and demonstrate that this typeof memory distortion does not take place for good perfor-


mance on an ethical attribute because good performancedoes not lead to as strong a sense of helplessness about theunderlying ethical issue. Our proposed effect is in directconflict with the negativity bias (Baumeister et al. 2001;Rozin and Royzman 2001). According to this bias, peo-ple tend to pay more attention to negative than to positiveinformation. If differences in memory were driven solelyby the negativity bias, we would expect to find thatconsumers would remember poor performance on anethical attribute better than good performance. Instead,we find that consumers’ motivation to forget poor ethicalattribute performance trumps any effects of the negativ-ity bias for these attributes.

Consistent with Ehrich and Irwin’s (2005) work onwillful ignorance in search, we also find that the willfullyignorant memory effect is particularly strong for con-sumers for whom the ethical issue is especially important.It is these consumers who are most likely to feel helpless

about solving the ethical issue in question when theyencounter products with poor performance on the ethicalattribute in the marketplace.

Finally, we propose that motivated forgetting is a copingmechanism (Luce 1998) used by consumers to avoid thefeelings of helplessness that thinking about solving large,intractable ethical problems can generate. Thus, if aconsumer is already feeling helpless about ethical issueswhen he or she encounters information that a productperforms poorly on an ethical attribute, then there is noneed to engage in motivated forgetting to stop the activa-tion of all of the negative memory nodes linked to theattribute value. By demonstrating that consumers do notmisremember poor values on the ethical attribute whenthey are primed to feel helpless, we therefore provideevidence that the pattern of forgetting we demonstrate inour studies is driven by a desire to cope/avoid feelinghelpless. References are available upon request.

For further information contact:Rebecca Walker Naylor

The Ohio State University2100 Neil Ave.

Columbus, OH 43210Phone: 614.247.6433

Fax: 614.292.0879E-Mail: [email protected]


WHEN MIGHT CONSUMERS CHOOSE SUPERIORSUSTAINABILITY DESPITE A TRADE-OFF WITH

OTHER PRODUCT ATTRIBUTES?

Michael G. Luchs, College of William and Mary, WilliamsburgMinu Kumar, San Francisco State University

SUMMARY

Building on prior research that suggests consumers ofteninfer a trade-off between sustainability and other productattributes (Luchs et al. 2010), the objective of the currentresearch is to contribute to our understanding of whenand why consumers will choose a product that is moresustainable, despite an explicit trade-off with other val-ued attributes. Specifically, we investigate the differentemotional and behavioral consequences of scenarios inwhich consumers face a trade-off between either productsustainability and utilitarian value, or between productsustainability and hedonic value.

Prior research has demonstrated that different types ofproduct attribute trade-offs can have very different emo-tional and behavioral consequences (Chitturi, Raghu-nathan, and Mahajan 2007, 2008). Choices involvingutilitarian attributes have been shown to be a conse-quence of prevention oriented emotions, such as confi-dence (Chitturi, Raghunathan, and Mahajan 2007; Hig-gins 2001; Luchs, Brower, and Chitturi 2012). On theother hand, choices involving hedonic attributes havebeen shown to be a consequence of promotion orientedemotions, such as excitement (Chitturi, Raghunathan,and Mahajan 2007; Higgins 2001). With respect to trade-offs with sustainability, we suggest that the orientation ofthe goals fulfilled (or unfulfilled) depends on what spe-cifically is being traded off with sustainability, as do theconsequent emotions and choice. Specifically, we pro-pose that it is especially difficult to justify the choice ofhedonic value (vs. utilitarian value) over sustainability;doing so is especially likely to increase guilt (and reducepride) relative to the morally superior sustainable option.In addition, consistent with the precedence principle(Berry 1994), it is also plausible that consumers are morelikely to choose utilitarian value (vs. hedonic value) oversustainability due to the greater confidence (and lowerdistress) experienced when choosing a relatively moreutilitarian product. Thus, we predicted that consumersare more likely to choose a product with superior utilitari-an value over a more sustainable product than tochoose a product with superior hedonic value over amore sustainable product, and we expected this effect

to be mediated by greater pride (less guilt) and by greaterconfidence (less distress). Further, we expected thiseffect of trade-off type on choice likelihood to be moder-ated by consumers’ attitude toward sustainability suchthat consumers are more likely to choose sustainability inthe context of a trade-off with hedonic value (vs. utilitar-ian value) as their attitude toward sustainability becomesmore positive.

In Study 1, participants were presented with a choicebetween two kitchen blenders that were described asdiffering along two dimensions. In the first condition,one option was superior (inferior) with respect to itsproduct performance (sustainability) whereas the otherwas superior (inferior) with respect to its sustainability(product performance). In the second condition, partici-pants were presented instead with a trade-off betweensustainability and aesthetic design. Within the productperformance trade-off condition, more participants choseproduct performance over sustainability (68:7), χ2 =32.81 p < .0001. However, there was no difference inchoice likelihood when the choice was between supe-rior aesthetics and superior sustainability (42:32), χ2 =1.34, p > .10. This effect of Trade-off type on Choicewas significant, χ2 = 18.84, p < .0001, confirming thatchoice likelihood – given a trade-off with sustainabil-ity – depends on the type of trade-off: specifically, asustainable product is more likely to be chosen in thecontext of a trade-off with hedonics than in a trade-offwith utilitarian value.

For our mediation analyses, we created a series of threeemotion dimension pair measures (confidence-distress,excitement-disappointment and pride-guilt) after con-firming that all emotion pairs were related as expected(all pairs inversely related, p < .0001). A bootstrappedparallel mediation analysis, per Hayes (2012), suggestedthat only greater confidence (less distress) mediated theeffect of Trade-off Type on Choice at the 95% confidencelevel. However, a bootstrapped serial mediation analysis,per Hayes (2012) suggested that greater pride (less guilt)may have indirectly played a significant role on Choicethrough its effect on confidence (distress) and excitement(disappointment). In addition, this effect of Trade-off


Type on Choice was moderated by consumers’ Attitudetoward Sustainability (AtS), χ2 = 5.41, p = .02, such thatchoice of the sustainable option increased amongstparticipants with relatively higher AtS scores, but onlywithin the superior aesthetic design trade-off condition.

The design of Study 2 was similar to the design used inStudy 1 except that participants were also able to “optout” of the choice. In addition, Study 2 employed twodifferent product category contexts: calculators and MP3players. The pattern of results was the same as the patternfound in Study 1. Within the performance trade-offcondition, more participants chose product performanceover sustainability (108:8), χ2 = 50.45, p < .0001. Onceagain, there was no difference in choice likelihood whenthe choice was between aesthetic design and sustainabil-ity (60:49), χ2 = 1.11, p > .10. Similarly, while a boot-strapped parallel mediation analysis, per Hayes (2012),suggested that only greater confidence (less distress)mediated the effect of Trade-off Type on Choice, a followup analysis suggested that pride (guilt) had an indirecteffect on Choice by attenuating – or blunting – theerstwhile greater confidence (lower distress) felt towardthe higher performing (yet less sustainable) product.

Our results extend prior demonstrations of the principleof precedence (Berry 1994), which suggests that utilitar-ian value is often chosen over hedonic value (Chitturi,Raghunathan, and Mahajan 2007), by showing that it ismore likely for utilitarian value (vs. hedonic value) to bechosen over sustainability. Further, we show that a pos-sible disadvantage faced by a more sustainable productcan be attenuated in the context of a trade-off withhedonic value as consumers’ attitudes toward sustain-ability become increasingly positive. Finally, we foundthat the effect of some of the emotions in our context –such as pride (guilt) – may be indirect. This finding,conceptually consistent with the Appraisal TendencyFramework (Han, Lerner, and Keltner 2007), is novelgiven the demonstration of a blunting effect with differ-ent emotions as well as demonstration of the bluntingeffect within a given choice context (i.e., as opposed to aseries of tasks). Beyond the theoretical and publicpolicy implications of our research, we believe thatour findings also present very promising opportunitiesfor future research on sustainable consumption which isneeded given the inherent complexity and conflictingemotions experienced within this unique context (Phippset al. 2012). References are available upon request.

For further information contact:Michael G. Luchs

College of William and MaryP.O. Box 8795

Williamsburg, VAPhone: 757.221.2906



BE A GOOD COOKIE: CAUSE-RELATED MARKETING’SINFLUENCE ON HEALTH HALO FORMATION

Elizabeth Minton, University of Oregon

SUMMARY

This research investigates how health-oriented causerelated marketing initiatives can cause a consumer toperceive a packaged good as healthier than it actually is(i.e., formation of a health halo). In a series of threestudies, findings reveal that products with a health causeon a package (e.g., “a portion of proceeds are donated tothis cause”) lead consumers to form a health halo sur-rounding the package. Addition of a disclaimer signifi-cantly reduces health halo effects but not purchase inten-tions. These effects are moderated by health interest, adskepticism, and BMI. Public policy implications arediscussed.

Conceptual Overview

A growing body of research has investigated the develop-ment of health halos and related policy implications toencourage consumer healthy decisions. Wansink andChandon (2006) describe a health halo as when a con-sumer overgeneralizes one healthy trait of a product (e.g.,low fat) to represent the totality of the product (e.g., alsolow sugar, high fiber, low sodium, etc.). This recentresearch on health halos has spawned off the Food andDrug Administration’s (FDA) Nutrition Labeling andEducation Act of 1990 to determine the influence ofpackage labeling on food purchase decisions (Silver-glade 1996).

One area where further research is needed lies in the fieldof corporate social responsibility (CSR) initiatives. WhenKeebler recently introduced cookie packaging with theAmerican Red Cross logo, it is possible that this logocreated a health halo causing the cookies to appearhealthier than they actually are. The American Red Crosslogo will be placed on approximately 175 million cookiepackages each year while the campaign runs (CauseMarketing Forum, 2012). Keebler is receiving positiveresults from the cause-related marketing initiative withincreased sales and improved brand perception. How-ever, what effect does this cause-related marketinginitiative have on perceptions of healthiness of the cook-

ies? Also, how could more knowledge in this area help toencourage consumer healthy decision making?

Although much research has investigated attributes ofcause-related marketing, no research to date has exam-ined the relationship between health halo creation andcause-related marketing. Previous research has shownthat cause-related marketing initiatives can result in con-sumer misperceptions. For example, Bower and Grau(2009) found that when a corporation and nonprofitpartner in a cause-related marketing initiative, consum-ers can misperceive that the nonprofit is giving a “seal ofapproval” for the corporation. The greater the fit betweenthe corporation and the nonprofit, the more likely theconsumer is to perceive the nonprofit approves of thecorporation. Thus, companies involved in cause-relatedmarketing initiatives should be cautious to ensure thatconsumer perceptions of the initiative do not result innegative consequences for either the corporation or non-profit. Just as consumers misperceive seals of approval,consumers could also misperceive a nonprofit endorsinga product as healthy and something that should be con-sumed in greater frequency.

Therefore, the purpose of this research is to answer twogeneral research questions: (1) how do cause-relatedmarketing initiatives on food packages result in creationof health halos, and (2) what role does public policy havein reducing health halo formation? To answer thesequestions, a variety of literature bases are broughttogether including priming, health, CSR, and cause-related marketing literatures to best understand underwhat circumstances health halos are created within cause-related marketing initiatives.

Methods

Three studies were conducted to investigate this problemusing undergraduate business students. Participants wereshown a mock package (Study 1: cookies, Studies 2 & 3:crackers) and asked to evaluate the overall healthiness,purchase intentions, and taste perceptions along withseveral moderating variables. Each study utilized threeconditions: (1) package with no cause, (2) package with


health-oriented cause, and (3) either a package with anon-health oriented cause (Study 1) or a package with ahealth cause and disclaimer (Studies 2 & 3).

Results and Discussion

Study 1 (N=109) reveals that products with a health causeon a package (e.g., a portion of proceeds are donated tothe American Heart Association) lead consumers to per-ceive the product is significantly healthier than a packagewith no health cause or a non-health cause. Qualitativeresponses suggest that consumers perceive the cause asendorsing the packaged good. Also, health interest mod-erates the relationship between health package designand health perceptions with high health interest consum-ers perceiving a package as less healthy (i.e., health haloreduction) than low health interest consumers.

Study 2 (N=101) replicates the findings from Study 1 thata health cause on a package (in this study, the WorldHealth Organization) significantly increases health per-ceptions resulting in health halo formation. Also, con-sumers classified as overweight (BMI > 25) were signif-icantly more likely to see a health halo around packageswith a health cause in comparison to normal weightconsumers. Results from Study 2 also provide publicpolicy implications by showing that adding a disclaimer

to a package with a health cause (“This is not an endorse-ment by the World Health Organization”), significantlyreduces health halo effects but does not reduce purchaseintentions.

Study 3 (N=99) again replicates the findings from Stud-ies 1 and 2 by showing that a health cause on a packagesignificantly increases health perceptions (i.e., a healthhalo is formed), and addition of a disclaimer reduceshealth halo effects. Also, consumers high in ad skepti-cism had significantly lower health perceptions with apackage containing a health cause and disclaimer than apackage with no cause.

Conclusions

In summary, this research shows that health-orientedcause-related marketing initiatives significantly increasehealth perceptions and result in the formation of a healthhalo. Disclaimers are one way for public policy to getinvolved and reduce formation of health halos. With over35 percent of the US classified as obese and an obesityepidemic facing both the nation and the world, thisresearch has the potential to positively influence con-sumer choice and aid in reducing consumer obesity fora better world ahead. References are available uponrequest.

For further information contact:Elizabeth Minton

University of Oregon1208 University of OregonEugene, OR 97403–1208

Phone: 541.346.9090Fax: 541.346.3341



TRANSLATED ATTRIBUTES AS A CHOICE ARCHITECTURETOOL: TRICK AND TREAT

Adrian R. Camilleri, Duke University, DurhamChristoph Ungemach, University of Columbia, New York

Richard P. Larrick, Duke University, DurhamElke U. Weber, University of Columbia, New York

Eric J. Johnson, University of Columbia, New York

SUMMARY

Choices are never presented in a vacuum: rather, alterna-tive choice options are always presented within a context.Importantly, in many cases the context – the “choicearchitecture” – reflects the meta-decisions made by thedesigner of the context – the “choice architect” (Johnsonet al. 2012; Thaler and Sunstein 2008). A key principleassociated with the choice architecture metaphor is thatthere is no neutral choice context and therefore thoseresponsible for framing decisions will influence choicesin all cases. Motivated by this belief, scientists andpolicy-makers have become increasingly interested inlearning how to best arrange the choice architecture inorder to help mitigate the threat of various social dilem-mas, most recently the threat of climate change (Newelland Pitman 2010; Weber and Stern 2011).

Energy labels, and in many cases the standards thataccompany the labels, are often considered to be the bestavailable tools for governments to manage energy-efficiency policies and climate-change-mitigationprograms (Gillingham, Newell, and Palmer 2009; Sternet al. 1987; Wiel and McMahon 2003). Given that theconsumption of fossil fuels is a major contributor ofgreenhouse gas emissions, one natural target of choicearchitecture intervention are vehicle choices via fueleconomy label design.

One important feature of the fuel economy label is thechoice of metric or metrics to include. A common featureof many consumer labels, including fuel economy labels,has been the presentation of highly related metrics. Forexample, the US fuel economy label, redesigned by theU.S. Environmental Protection Agency (EPA) in 2012,includes three metrics that are all different expressions offuel economy: gallons per 100 miles, estimated annualfuel cost, and greenhouse gas rating expressed on ascale of one to ten. Interestingly, these three metricsare perfectly correlated – they are different transla-tions of fuel economy. We investigated how the presen-

tation of multiple translated attributes could change con-sumer’s preferences.

The consumers’ choice literature has converged on theconclusion that choices are inherently constructive, thatconsumers often do not have well-defined existing pref-erences, and that consumers construct their preferencesusing a variety of strategies contingent on task demands(Payne, Bettman, and Johnson 1988). One non-compen-satory strategy that boundedly rational consumers mayadopt is a simple counting heuristic in which the deci-sion-maker simply counts up the number of attributes onwhich one alternative outperforms the others and selectsthe option with the highest score. Counting heuristicsfavor the option that is better on more attributes. Wetherefore hypothesized that the presentation of multipletranslations of a global dimension (e.g., fuel economy)would shift consumers’ preferences toward the optionthat was better on the majority of the presented attributes(i.e., the translated attributes effect).

Another key feature of translated attributes is the fact thatthey may highlight different characteristics of a globaldimension. Adding a specific translation might thereforeincrease the accessibility of the dimension they describein memory as well as its perceived importance (Shah andOppenheimer 2009). Similarly, translated attributes canremind individuals of additional goals they may haveoverlooked but nonetheless care about. We hypothe-sized that the translated attributes effect would bemoderated by the translated attribute’s congruence withthe consumer’s goals: specifically, those with pro-envi-ronmental values to more often select the efficient optionbut only when one translation was environmental (i.e.,greenhouse gas rating).

We observed that the presentation of multiple translatedattributes did indeed shift preferences. Specifically, wefound that the weight placed on a global dimension (e.g.,fuel efficiency) tended to increase when it was translated


into multiple attributes (e.g., annual fuel cost, GPM,GHG, rating). Put simply, more fuel-efficient vehicleswere selected when more translated fuel efficiency at-tributes were presented, whereas cheaper vehicles wereselected when more translated price attributes were pre-sented. We also found that participants’ likelihood ofselecting the fuel-efficient option was affected by theirpro-environmental attitudes but only when the translatedfuel-efficiency attributes included the greenhouse gasrating.

These results are best captured by two concurrent choicemechanisms: counting and goal activation. One the onehand, the basic translated attribute effect can be ex-plained by the application of a counting strategy accord-ing to which people simply choose the option that ispreferred on most of the available metrics (the “trick”).

On the other hand, the moderating role of personalenvironmental values is in line with a goal activationaccount, according to which different translations high-light different aspects of fuel-efficiency and choices canbe further influenced when there is a match between thetranslation and individual goals (the “treat”). Thus, incomparison with the counting heuristic, goal activationhas much higher specificity.

In summary, we have proposed the novel concept oftranslated attributes – related attributes derived from aglobal dimension by simple mathematical transforma-tion – as another pillar of the choice architecture frame-work. This tool will be useful for policy-makers andmarketing managers alike as they strive to effectivelycommunicate with consumers to guide informed con-sumer choices. References are available upon request.

For further information contact:Adrian R. Camilleri

The Fuqua School of BusinessDuke University100 Fuqua Drive

Durham, USA 27708E-Mail: [email protected]


THE UNINTENDED CONSEQUENCES OF CALORIEINFORMATION ON MENUS: THE USE OF

ATTRACTION AND COMPROMISE EFFECT

Ryall Carroll, St. John’s University, New YorkBeth Vallen, Fordham University New York

SUMMARY

For those of us who live in cities where calorie informa-tion has become mandated, it is not a shock that calorieinformation on menus is being used by consumers tomake their consumption choices. That was, in essence,the whole point of mandate in the first place. But thequestion that remains is: are there any unintended conse-quences of this new calorie information? No longer areyou just choosing an item off the menu, but instead youare choosing an item of a menu, in the context of and incomparison to other items and their calorie information.Is this a good thing or bad thing? In most cases this isprobably a good thing. Prior research has shown that theproviding calorie information on menus can lead tochanges in consumer’s choice (e.g., Howlett et al. 2009).As a result people are hopefully making healthier or atleast more knowledgeable decisions. The goal of thisresearch, however, is to present examples when unin-tended negative consequences occur as a result of thisinformation. This research contributes to the literature byexploring whether the attraction effect and compromiseeffect, two of the most robust findings in consumerbehavior, are likely to occur when consumers are present-ed with calorie information using this unique source ofnutritional information.

When viewed together, our three studies show that a fooditems’ calorie information is not used in isolation, but israther used comparatively against other food items’ cal-orie information. This use of information can result inless optimal choices for the consumer and can even alterthe perception of the food. In study 1 we replicated pastcompromise effect research (Simonson 1989; Sharpeet al. 2008) with the use of calorie information as anattribute. Participants’ chose food items within the con-

text of other options and in both cases the target item sawan over ten percent increase in share choice when the itemwas presented in a choice set as the compromised choice.

Additionally, in studies 1 and 2 we showed the strengthof the attraction effect and how the addition of a decoyitem can result in an over a ten percent increase inselection of the targeted food item. In four out of the fiveattraction effect conditions, the targeted item was thehigher calorie option, meaning that in each attractioneffect condition, over ten percent of participants wouldhave chosen a lower calorie item had a third decoy itemnot been offered.

Lastly, in studies 1 and 3 we show the strength of the“higher calories equal taste” heuristic. In study 1, weshow the power of the heuristic by showing its strength inthe dessert category over the robust compromise effect.The compromise effect did not work in the ice creamcondition, despite having worked in the frozen dinnercondition. In study 3, we show the heuristic can eveneffect people’s perception of taste, quality, and satisfac-tion. Participants tasted the exact same cookie and whentold that one cookie had a higher calorie count, theperception was altered despite there being no differencein the cookie whatsoever.

In this research, we have shown the unintended conse-quence of having calorie information on a menu. Pastresearch has looked at how nutritional panel information(e.g., Keller 1997) is processed and how this can be usedin presentation to the consumer to be most effective. Inthe same vein, future research may need to focus on howinformation is processed when reading a menu in order topresent the information in the most beneficial and effec-tive manner. References are available upon request.


For further information contact:Ryall Carroll

St. John’s University8000 Utopia Parkway, Bent Hall

Queens, NY 11439Phone: 718.990.7348



ENVIRONMENTAL CUES AND FOOD CONSUMPTION AMONGDIETERS AND NON-DIETERS

Beth Vallen, Fordham University, BronxChrissy Mitakakis, Iona College, New Rochelle

SUMMARY

Dieters often start with the best of intentions, planning totruly commit to their diet goal “on Monday,” or perhaps“once the holidays are over.” Unable (or perhaps unwill-ing) to face the food temptations they face on weekendsand special occasions, individuals are more likely tooverindulge once they have engaged in even just a smalltemptation (i.e., the “what the hell” effect; Polivy andHerman 1985) and plan to reengage in their diet goalon a new day in which their prior diet inhibitingactivities do not threaten their goal. While extantresearch has explored special occasion consumptionwith a focus on the typical American holiday season (i.e.,Thanksgiving through New Year’s Day), this researchseeks to explore more generally how dieters and non-dieters differ in their attempts to bolster dietary defensesin the face of external cues related to what we term eatingoccasions—events that serve as a cue to consume tempt-ing food options.

To date, research examining the effects of weak versusstrong food-related cues on consumption has focusedspecifically on tempting food stimuli. We extend thisbody of literature by examining the effects of cues asso-ciated with eating occasions on dieting defenses. Giventhat special occasions and holidays are used as markersfor indulgence (Kivetz and Zheng 2006; Mukhopadhyayand Johar 2009), we posit that they also serve as strongfood-related cues for individuals. This is supported byanecdotal evidence from diet-focused publications thatbombard readers with advice regarding ways in which toavoid indulgence as the holidays approach; for instance,Shape Magazine touts its “No Weight Gain HolidayGuide” (2009) while Men’s Health tells its readers “howto indulge, without showing a bulge” (“Season’s Cheat-ings” 2009). For dieters, who are attempting to controltheir consumption, we argue that such cues will bolsterindividuals’ dieting defenses relative to when these indi-viduals are faced with weaker cues that are not related toan eating occasion (i.e., an ordinary day). We do notpredict differences in bolstering dieting defenses for non-dieters as, in the absence of dietary goals, these individ-uals are not likely to perceive eating occasions as threats.

We explore these predictions in three empirical studiesthat assess consumption behavior during times of eatingoccasions traditionally associated with indulging in food.

In our first study, participants were given a scenario inwhich they were asked to imagine that they were out todinner celebrating either their own birthday or theirfriend’s birthday. All participants were then asked tochoose the specific dessert they would order at the dinner,ranging from extremely unindulgent (i.e., a healthy fruitsalad), too extremely indulgent (chocolate cake á lamode). After the dessert choice task, we assessed dietingbehavior with the following item: (“I am currently diet-ing.”). Supporting our predictions, those scoring rela-tively higher on the dieting measure (i.e., dieters) select-ed the cake á la mode to a greater extent when theoccasion was their friend’s birthday relative to their ownbirthday; those low on the dieting scale (i.e., non-dieters)did not differ in selecting the cake á la mode based on thedifferent birthday occasions. In addition, we also findthat when participants were told it was their own birth-day, dieters were less likely to select the cake á la modethan non-dieters, but when it was their friend’s birthday,they were just as likely to select the cake á la mode asthose who were considered non-dieters.

In the next study, we examined these effects using actualconsumption. We varied the eating occasion cue byrunning the study on two different days using differentsections of the same course: one on Halloween day andthe other on a random, non-holiday weekday severalweeks afterward. Participants were given a bag filledwith 50 M&M candies in a Ziploc bag that they couldenjoy the candy as they filled out a questionnaire assess-ing their dieting behavior and BMI. At the end of the classperiod, participants were asked to count how many M&Mcandies remained in their bags. We find support for ourhypotheses, such that individuals scoring relativelyhigher on the dieting measure and BMI (i.e., dieters whoare overweight) consumed more M&M candies on anordinary day relative to Halloween; those low on thedieting scale (i.e., non-dieters) with a similar BMI did notdiffer in the amount of M&M candies consumed on anordinary day relative to Halloween. Furthermore, results


also demonstrate a directional effect on Halloween, suchthat dieters with a higher BMI ate less candy than thosewith a lower BMI.

Finally, in our third study, we explored whether dietersbolster their defenses more than non-dieters in the pres-ence of eating occasion cues. Similar to Study 2, we ranthe study a few days before Thanksgiving and alsoseveral weeks afterward. Individuals answered four itemsaimed at assessing the extent to which participants tend tobolster their dietary restraint, which we combined inorder to obtain an index measure (Bolstering), as well asa measure of their dieting behavior. Supporting ourpredictions, those scoring relatively higher on the dietingscale (i.e., dieters) bolstered dieting goals to a greaterextent when Thanksgiving was salient relative to anordinary day; those low on the dieting scale (i.e., non-dieters) did not differ in bolstering based on the presenceof weak versus strong food-related cues. Additionally,

we find that around Thanksgiving, dieters bolstered goalsmore than non-dieters. Importantly, we do not see adifference in the bolstering of dieting goals when it wasan ordinary day.

In conclusion, the findings from this research contributeto the existing literature on environmental cues and foodconsumption. Much research has documented the impor-tant role of external cues in the form of tempting fooditems, and has found that dieters guard against tempta-tions that are not in line with their long-term dietinggoals. Similarly, the current research shows that dietersreact defensively in response to holiday cues that repre-sent eating occasions. On the other hand, we also showthat in the absence of holiday cues, self-control breaksdown and dieters do not have the defenses to guardagainst indulgence. The results presented in the currentresearch can have numerous implications for individualswith regard to holiday cues representing eating occa-sions. References are available upon request.

For further information contact:Chrissy Mitakakis

Iona College715 North Avenue

New Rochelle, NY 10801Phone: 914.633.2422



THE IMPACT OF FOOD GROUPS, DIETARY VARIETY ANDQUALITY, AND DEMOGRAPHICS ON DAILY

CALORIC INTAKE

Adwait Khare, University of Texas at ArlingtonJ. Jeffrey Inman, University of Pittsburgh

Chung-Tung Lin, United States Federal Drug AdministrationWenYen Juan, United States Federal Drug Administration

SUMMARY

We examine the influence of multiple food consumptionrelated behavioral measures and demographics on indi-viduals’ daily caloric intake. Given the high frequency offood consumption behavior and its implications for per-sonal, household, social, and national outcomes, suchbehavior has received significant research attention. Wepropose that variables influencing daily caloric intakecan be grouped into three hierarchical levels – day,individual, and household – with days nested withinindividuals in turn nested within households. Based on afood diary panel dataset provided by the NPD Group, aleading market research company, we test the effects ofa range of variables from the three levels on individuals’daily caloric intake.

We use data from NPD’s annual Eating Trends Surveyfrom 2010 and 2011. During the survey, the head of eachparticipating household keeps a daily diary to report thefoods eaten by all household members. The survey isconducted for two consecutive weeks. The sampling isdone in a rolling fashion such that every week somehouseholds start their two-week participation. Our finalsample consists of 8,415 individuals from 3,752 house-holds. For every individual, for every participating day,information for up to twelve at-home versus away-from-home meals is reported. For every meal, all items con-sumed are reported. NPD classifies the multitude ofreported food names in two ways, one finer and onebroader; there are 4,245 finer NLINK codes and these aregrouped further in to the 184 broader codes. We furthergrouped the NLINK2 codes into eight food group codesof which, five are the food groups as defined in the 2010Dietary Guidelines for Americans (DGA) – milk andmilk products, fruits, grains, protein foods, and vegeta-bles – and the remaining three groups are snacks, bever-ages, and fats/oils. Thus, for each item reported, we havea NLINK, NLINK2, and DGA (plus 3) code. The lattertwo are crucial for the behavioral measures we created.

For each item, information about serving size, speciallabel (e.g., light), and item type (e.g., main dish) is alsoreported. Using the NLINK codes and other informationabout individuals (e.g., age, gender), NPD computeseach individual’s caloric and nutrient intake based on anaverage calculated serving size for each food item ingrams. Such intake information is calculated by matchingeach food item with food items listed in the USDA Foodand Nutrient Database for Dietary Studies (v. 4.1, 2010).

Dependent Variable. We aggregated the caloric, nutri-ent, and grams consumed for each individual per reportedday. The daily caloric totals (one for each day anindividual participated in the survey) calculated aftersuch aggregation serve as the dependent variable in ouranalysis.

Day-Level Independent Variables. We calculated perindividual per day the grams of food consumed in the fiveDGA groups and the three additional groups we created.The coefficients on these food group variables indicate alarge range, for example, a gram of snack items provides2.66 calories but a gram of fruit items provides only 0.46calories. Another day-level variable we included is adummy variable for weekend (Fridays, Saturdays, andSundays) versus week days and it shows that weekendintake is higher than week day intake by about 49 calories.

Individual-Level Independent Variables. We analyze theimpact of various measures of dietary variety and qualityand also of other behavioral and demographic measures.Our first measure of variety, NumUniqVariety, is a countof the total number of unique NLINK2s reported by anindividual per day. The higher this count of unique items,the more an individual is consuming overall and so weexpect NumUniqVariety to have a positive effect on dailycaloric intake (supported). Our second measure of vari-ety, PerUniqVariety, is calculated as the percentage ofthe count of the unique NLINK2s from among the totalnumber reported. The higher this percentage, the more


the individual spreads his or her intake across distinctfood types. Since a higher spread means a higher chancethat foods of varying levels of healthiness are beingconsumed, we expect PerUniqVariety to have a negativeeffect on daily caloric intake (supported). Our thirdmeasure of variety, HerfindahlVariety, assesses the ex-tent of the spread in intake (using the formula of theHerfindahl Index) over the five DGA food groups. Ahigher spread for this index means a higher chance thatfoods of varying levels of healthiness are being con-sumed, and so we expect HerfindahlVariety to have anegative effect (supported).

Our first measure of dietary quality, DrewnowskiQuality(Drewnowski et al. 1997) is a 5-point scale based onindividuals’ conformity with key U.S. dietary recom-mendations for fat, saturated fat, carbohydrate, choles-terol, and sodium. Since a higher score means a healthierdiet, we expect DrewnowskiQuality to have a negativeeffect on daily caloric intake (supported). Our secondmeasure of quality, GoodBadDiffQuality, is based on thepopular perception that fruits and vegetables are gen-erally healthier compared to snacks and beverages (ex-cluding alcoholic beverages) and was calculated as thedifference in the percentage of calories from fruits andvegetables versus snacks and beverages. Since a higherdifference would mean a relatively healthier diet, weexpect GoodBadDiffQuality to have a negative effect ondaily caloric intake (supported). Our third measure ofquality, AtHomeVsAwayQuality, is based on the popu-lar perception that foods consumed at home are healthierthan foods consumed away from home and was calculat-ed as the difference in the percentage of calories from at-home versus away-from-home intake. If the perception istrue, a higher difference would mean a relatively health-ier diet and so we expect AtHomeVsAwayQuality tohave a negative effect (supported).

We also modeled the number of diets (surprising nulleffect), number of weight-related medical conditions(negative effect), number of exercises (positive effect),

the number of special labels on the items consumed(surprising positive effect), and individuals’ age (posi-tive effect), gender (men consume 375 more calories perday), and BMI (positive effect) information as well.

Household-level Independent Variables. Except for gen-der, the household counterparts of all individual-levelvariables were calculated as the average of all individual-level variables for all individuals in each household. Amajority of the household-level variables calculated inthis fashion had an effect of the same sign as theircorresponding variable at the individual level. We mod-eled household head’s race (individuals from householdswith White head persons consume the least), Hispanicstatus (no effect), household size (no effect), and income(no effect) also.

Our results indicate measures of dietary variety andquality have a dual effect, via patterns in individuals’own consumption behaviors and via patterns in house-hold members’ consumption behaviors. Such facilitativeeffects of family on consumption have been discussed byothers, but a systematic, “hierarchical” examination likeours has not been possible so far due to data limitations.Our results for dieting and use of special labels suggestthat consuming healthy foods may ironically increase theconsumption of less healthy foods. The dual, individual-and household-level, effects we observe for the variousmeasures indicate that for their caloric intake outcomesconsumers must better understand not only their ownconsumption behaviors but also those of household mem-bers. Thus, it would be useful to craft variety and qualityguidelines at both the individual and household levels.More broadly, consumers must be better educated aboutthe various, multi-level influences on their consumptionso that they are better able to manage it. Our resultssuggest that both dietary variety and quality can becultivated in many ways and that their benefits can beadditive. Consumers must be mindful of the possibilitythat dieting and consuming special label foods can stim-ulate rather than contain consumption. References areavailable upon request.

For further information contact:Adwait Khare

University of Texas at Arlington701 West Street

Arlington, TX 76019Phone: 817.272.0967

Fax: 817.272.2854E-Mail: [email protected].


CONJOINT ANALYSIS OF THE CONSUMER TRADE-OFFDECISIONS ASSOCIATED WITH A BEVERAGE TAX

Jess Mikeska, University of Nebraska – LincolnLes Carlson, University of Nebraska – LincolnDwayne Ball, University of Nebraska – Lincoln

SUMMARY

In response to dangerous and increasing obesity rates(CDCP 2010), public policy leaders and researchers areworking to resolve this trend by understanding the rela-tionship between over consumption of calorically-densefoods and marketing activities. Empirical understandingof this relationship has never been more concerningbecause calorically-dense food is relatively inexpensive(Montopoli 2009)and often the target of recent snacktaxes (Fletcher, Frisvold, and Thefft 2010). At the sametime, Congressional Budgetary Office meetings indicatethat a national beverage tax could be in the offing (Adamy2009; Fried 2010).

While news reports (Adamy 2009; Fried 2010) provideestimates for potential revenue from a sweetened bever-age tax, less prevalent are indications of the possibilitythis tax might have for reducing soda consumption andobesity. It is important to think of an optimal tax in termsof switching behaviors, i.e., how a tax might influenceconsumers to actually change their consumption behav-ior, because soda is linked to over consumption of solidsnack foods as well as significant weight gain (e.g.,Mattes 1996; Raben et al. 2002). In addition, researchalso often finds a tax that raises the price of soda productsleads to increased consumption of either beverages withmore calories or beverages with fewer nutrients than soda(Fletcher, Frisvold, and Thefft 2010). Finally, draftingtax codes that include regular soda but exclude diet sodaruns the risk of producing unintended consequences suchas increased consumption of other foods (e.g., Blundelland Hill 1986; Fowler et al. 2008; Lavin, French, andRead 1977; Mattes 1990).

Prior legislative attempts at controlling marketing activ-ities thought to impact obesity, such as package claims orproduct innovations, have yet to prove efficacious inencouraging nutritious habits (e.g., Moorman 1998;Moorman, Du, and Mela 2005). In addition, obesity-linked soda products are lower in price relative to health-ier alternatives such as fresh meats, vegetables, and

fruits. Consequently, there is a need to investigate furtherhow consumers make tradeoffs regarding beverage at-tributes. The following research questions are posed:

RQ1: Of the tax rates proposed so far, which willdecrease consumption of unhealthy beverages?

RQ2: What beverage types should be taxed to increaseconsumption of healthy beverages?

RQ3: What additional beverage product attributesbeyond price and type impacts consumptionenough such that they should be considered anelement of tax legislation?

RQ4: Which proposed beverage tax rates will impactconsumer BMI levels?

As with other studies of consumer attitudes, collegeundergraduate students represent an appropriate samplefor the research questions posed (e.g., Petty and Cacioppo1996; Petty and Krosnick 1995). Moreover, thosebetween the ages of 12 and 19 years consume soda morefrequently than any other age group in the US (Wang,Bleich, and Gortmaker 2008), and the proportion ofcollege students (about one third) reported to be obese issimilar to that of adults (USDA 2005; CDCP 2010). Assuch, the responses of 82 undergraduate students wereanalyzed in order to address the research questions posed.Sawtooth’s ACA software was used for data collectionbecause its adaptive nature allows accurate part-worthestimation even though the respondent does not maketrade-off decisions for multiple product scenarios. Per-haps of most importance, conjoint analysis is an efficienttool for tapping into consumer product preferences with-out facilitating potential bias arising from asking for suchpreferences directly. The four design attributes, devel-oped through the results of multiple pre-tests, included(1) five levels of price per 6-pack of 20-ounce bottles ($0,$4.00, $5.44, $6.88, $8.32), (2) seven types of beverages(tap water, pure fruit juice, soft drinks, sports drinks,bottled water, coffee, tea), (3) two levels of caffeine


(caffeinated and not caffeinated), and (4) three levels ofsweeteners (sugar, artificial sweeteners, no sweeteners).Respondents rated beverage preference on a seven-pointscale in 24 trade-off scenarios, described by threeattributes at a time, and randomly displayed.

The findings suggest a sweetened beverage tax couldsignificantly decrease soda consumption if the taxexcluded artificially sweetened beverages, but only if thetax made the price of sugar-sweetened beverages consid-erably more expensive. For example, a 1 cent increase perounce tax on soda did not appear to be a particularlyaversive consideration for our respondents. It was onlywhen this tax moved to the 3 cent per ounce increase levelthat the price of a six pack of soda became something toavoid.

More importantly, optimal legislation should not assumeconsumers behave through bounded rationality, but shouldinstead attempt to predict product substitution decisions.For example, strong preference for beverage type over-rides an increase in price in some scenarios. Only whensoda price increases from $4 to $6.88 does consumptionof various sweetened beverages begin to decrease. Sec-ond, findings reveal no relationship between attributepreferences for various beverages and BMI, despite highrates of soda consumption among college age students(Wang, Bleich, and Gortmaker 2008). And a third sam-pled are obese. While the simulated market does reveal

that some consumers are likely to reduce consumption ofcalorically-dense beverages when the prices of theseproducts become expensive, it does show that consum-ers cease consumption of these beverages or that theyincrease consumption of healthier alternatives. Thus,while preference for sweeteners does not appear to berelated to BMI, the larger concern may be switchingobese, or potentially obese, consumers away from highcalorie beverages and toward water. Third, no tax scenar-io appears likely to encourage switches toward healthydrinks, such as tap water, pure fruit juice, bottled water,and unsweetened tea. It appears consumers will look forsweet substitutes given a beverage tax, including artifi-cial substitutes.

If the ultimate goal of beverage tax legislation is to reduceconsumption of soda, our results may complicate thepresumed efficacy of such policy. For example, it is onlyat the higher levels of tax increases where consumers mayactually switch away from soda, though, switching fromsoda to more “healthy” alternatives appears to unlikelyacross soda tax possibilities. Taxing sugared alternativesexclusively could result in consumers switching to dietsoda but, as noted earlier, consumption of diet soda is notwithout its own set of unhealthy consequences. If, on theother hand, the ultimate goal of beverage tax legislationis to generate revenue for resources and education aimedat curbing obesity, our results may be in line with policyobjectives. References are available upon request.

For further information contact:Jess Mikeska

University of Nebraska – Lincoln310 CBA

Lincoln, NE 68588Phone: 402.472.5610



RESTORATIVE SENIOR CENTER SERVICESCAPESAND SENIOR HEALTH

Mark S. Rosenbaum, Northern Illinois UniversityJillian Sweeney, University of Western Australia, Australia

SUMMARY

Elderly people are susceptible to non-communicablediseases, such as cancer, diabetes, and chronic respira-tory disease (Bloom et al. 2011). In addition, the elderlytypically experience negative life events that diminishor curtail their social networks and the support theyonce received (Rosenbaum 2006). Consequently, olderadults are prone to experiencing social and emotionalloneliness, stress, and depression (Mayo Clinic 2012;Rosenbaum et al. 2007). Fatigue, or exhaustion, is oneof the most common complaints among the elderly;estimates reveal that up to 50 percent of the elderlysuffer from mild fatigue (Wick and LaFleur 2007). Asolution is required to help individuals and societiesaffordably and effectively help older-aged adults andthe elderly remedy symptoms associated with loneli-ness, most notably, fatigue. This research finds thatsenior day centers represent a solution that often trans-forms the well-being of their elder patrons by fashion-ing a therapeutic, or a restorative servicescape (Rosen-baum 2009). Thus, senior day centers, which are non-residential facilities operating 10–12 hours per day andoffering affordably priced or complimentary meals;social, educational, and practical assistance; and enter-tainment classes/activities (Calsyn and Winter 2000),emerge as solutions to help the elderly deal with fatigueand improve their well-being.

This research empirically shows that senior day centersplay therapeutic roles in their patrons’ lives. To date,researchers have explored the extent to which theelderly often patronize for-profit, commercial ser-vice establishments, such diners (Rosenbaum 2006),fitness clubs (Rosenbaum 2009), shopping malls (Kimet al. 2005), and beauty shops (Garzaniti et al. 2011), toescape loneliness and maintain commercial friendships(Price and Arnould 1999). However, this research addsto this paradigm by showing the practical and evocativeroles of senior centers in enhancing their patrons’ well-being. Service establishments housed in the public

realm may play significant roles in their patrons’ lives byserving as therapeutic landscapes that transform their well-being by encouraging human restoration.

We obtained the data through a self-administered ques-tionnaire that was randomly distributed to 85 patrons of asenior center located in a major Australian city. No person-al information was asked of the respondent, and one of thestudy’s authors was present during data collection to pro-vide assistance. Of the respondents, 62 (73 percent) werewomen and 23 (27 percent) were men. In addition, 32percent were age 60–69 years, 26 percent were age 70–79years, 20 percent were age 80–89 years, 11 percent wereage 50–59 years, 6 percent were age 90 years or older, and5 percent were age 49 years or below.

The data reveal that some senior center patrons respondto three restorative dimensions found within a servicedesign—being-away, fascination, and compatibility – in amanner that promotes health benefits, including lowerlevels of fatigue, improved quality of life, and betteremotional and physical health. Thus, researchers whoquestion whether there is a place for commercial establish-ments in public health (Frumkin 2003) now have empiricaland descriptive evidence that links public places to humanhealth.

This research expands Bitner’s (1992) seminal work onservicescapes and shows the transformative role of seniorcenters in elder care by acting as therapeutic environments.Along these lines, we encourage service researchers toconsider employing Attention Restoration Theory (Kaplan1995; Kaplan and Kaplan 1989; Rosenbaum 2009) toinvestigate whether younger-aged consumers also seek outand patronize retail establishments (e.g., an Apple store) toachieve restoration. Perhaps, servicescapes that are nottraditionally perceived as restorative (e.g., penitentiaries,youth detention centers) often fail because their patronssuffer from directed attention fatigue. This work revealsthat even these settings could be modified in an innovativemanner that promotes human restoration and healing.


From a practical perspective, the National Council onAging (2012) reported that senior centers are one of themost widely used services among older adults, with morethan one million patronizing senior centers every day.The literature is replete with articles that reveal that olderadults who participate in senior center programs learn tomanage and delay the onset of chronic disease andexperience measurable improvements in their physical,social, spiritual, emotional, mental, and economic well-being. However, to date, little is known about the impactof senior center design on patrons’ attitudes and behav-ior. The current article fills this void by offering manag-ers a practical understanding of how some patrons under-go restorative, transformational changes in senior cen-ters. However, we encourage future researchers toemploy qualitative, descriptive research to betterunderstand why and how some senior center patronsperceive a restorative servicescape.

Although this research is one of the first in marketing toexplore the role of senior centers in consumer health, it isnot without limitations. Most notably, the study employsa smaller-sized, Australian-based sample. Thus, it mightbe speculated whether the research findings generalizebeyond Western, industrialized societies. However, be-cause we offer support for the essential tenets of ART, weare confident of the study’s practical implications tosenior center management and theoretical opportunitiesto service researchers. We also organize several substan-tive literature streams under the burgeoning transforma-tive service research paradigm. Given the global agingtrend and the problems that seniors often experience asthey age, we believe that senior centers are innovative,relatively inexpensive, non-medical places that have thepotential to improve individual and societal well-being.References are available upon request.

For further information contact:Mark S. Rosenbaum

College of Business, Department of MarketingNorthern Illinois University

DeKalb, IL 60115Phone: 815.753.7931



TRANSPORTATION STUDIES REVEAL BARRIERS FOR PERSONSWITH DISABILITIES: USING THE CONCEPT OF

CONTINUOUS SEQUENCE

Carol Kaufman-Scarborough, Rutgers University School of Business, Camden

SUMMARY

A major concern for persons with disabilities istransportation access to the marketplace. Throughout theyears since the ADA was adopted, in-store access hasimproved as ramps, automated doors, and widened aisleshave become customary parts of retail atmospheres.However, traveling to and from a “trip to the store” mayprovide additional unique challenges. While recent stud-ies have examined issues that are in the store environ-ment, none have explicitly considered that a major factorin marketplace access is transportation, its ease of use,and the accommodations that it provides to persons withdisabilities.

This paper will do the following: (1) discuss the find-ings of four national studies of persons with disabili-ties related to transportation; (2) outline the specificADA jurisdictions of public and private transportationsystems; (3) suggest a theoretical framework that inte-grates the role of transportation in marketplace accessi-bility; and (4) utilize the Bureau of Transportation’s 2002National Transportation Availability and Use Survey todraw public policy and marketing implications. Thisspecific database was developed in order to examinewhether the nation’s transportation system is currentlymeeting the needs of persons with disabilities throughoutthe United States. Data collection included the identifica-tion of various modes of transportation used by personswith and without disabilities, including automobiles, airtravel, and public transit.

In 1974, Mehrabian and Russell introduced environmen-tal psychology to urban planners. They proposed thatpeople are likely to approach or avoid various environ-ments based on their reactions to the characteristics ofthose environments. Among other questions, they asked:“How does the physical environment . . . influence thefrequency and the quality of social interactions?” (p. 4).Three specific emotional reactions were the likely conse-quences: these included pleasure-displeasure, arousal(stimulated-relaxed), and dominance-submissiveness. As

a result of their reactions, consumers are thought todetermine whether they wanted to stay or not, whetherthey wanted to avoid or explore the environment, and soforth.

In reality, environmental factors such as limited masstransit “create” a disabling environment for consumerswho are unable to drive, while similar consumers withample mass transit can function independently and maynot feel disabled at all (Kaufman-Scarborough and Baker2005). While most researchers have studied the experi-ences of persons with disabilities actually within specificshopping environments, the “extended environment” lead-ing from their homes should be considered as well. Theimportance of transportation in the sequence of stagesneeded to enter a shopping environment is an extendedpart of the concept of “continuous sequence” within astore environment itself. In a more traditional sense, acontinuous sequence can trace the path of a shopper fromstore entrance, through displays, to restrooms, back tomerchandise, and finally to checkout, also emphasizingthe transitions, such as changes in flooring, that theindividual must traverse.

Persons with disabilities may depend on: (1) specificprivate transportation to travel to store access, such as alocal taxi service; (2) a retailer who provides transporta-tion to and from the store facility, such as a hotel shuttleat an airport; (3) public transportation, such as trains andbuses, to get to shopping areas; and (4) finally, their ownabilities to walk, wheel, or drive themselves to destina-tions where purchases can be made. Such considerationsare important to public policymakers since they cansignificantly affect the likelihood of a person with dis-abilities being able to establish feelings of independence,rather than dependency.

The Bureau of Transportation Statistics developed andexecuted the 2002 Transportation Availability and UseStudy to provide a comprehensive database for analysisby government bodies and public policy researchers. Thedata represent a unique opportunity to examine the mass


transportation perceptions of persons with disabilities incomparison to those without disabilities. These dataenable us to examine the travel behaviors, availability oftravel accommodations, the travel methods used, andtheir perceptions of that travel arrayed by disability types.The study collected interviews from 5,019 persons, in-cluding 2,321 persons with disabilities. Several measuresof disabilities are included in the data set, providing a richopportunity for analysis.

In order to add some detail to the “story” of the travelexperiences of persons with disabilities, the manuscriptfocuses on the responses related to airline travel in orderto narrow the frame for analysis. In contrast to thestereotypical reports that argue that persons with disabil-ities do not travel regularly, over half of those whoreported disabilities, a total of 1362 respondents, took along-distance trip during the past year. When taking the

long trip, 45% traveled by car, either as a driver orpassenger. The next most frequent mode of travel wascommercial airline, reported by 239 respondents (10.6%).Moreover, in 2002, 428 (19.1%) of the respondents tookat least one round trip on a commercial airline, with theirtotal number of trips ranging from 1 to 35 trips.

As anticipated, responses included problems with curbcuts and stairs, broken elevators and escalators, obsta-cles, passing space, problems with moving sidewalks andtrams, the lack of available wheelchairs, and inadequateparking. Other problems ranging from “too much walk-ing” to the inability of a person’s assistant to be allowedto offer customary assistance. Public policy researchersare called to examine the larger environment experiencedby persons with disabilities when transportation avail-ability, quality, and convenience are considered. Refer-ences are available upon request.

For further information contact:Carol Kaufman-Scarborough

Rutgers University School of Business – Camden227 Penn Street

Camden, NJ 08102Phone: 856.225.6592



THE TORTOISE AND THE HARE EFFECT: AGED“MARATHONERS” BEST YOUNGSTER

“SPRINTERS” WHEN LEARNINGEXPERIENTIALLY

Ashley Goerke, Temple University, PhiladelphiaEric M. Eisenstein, Temple University, Philadelphia

Ayalla Ruvio, Temple University, Philadelphia

SUMMARY

It is common knowledge that developed nations are“graying,” meaning that there will soon be more elderlypeople as a proportion of the population than ever beforein history. The U.S. Census Bureau projects that by theyear 2030, the number of Americans over the age of 65will more than double to over 72 million people – 25percent of the population (U.S. Census Bureau 2008),and the percentages in other developed nations aresimilar (United Nations Department of Economic andSocial Affairs 2009). Moreover, seniors will controlan increasing share of both total wealth and discretion-ary income, trends which will accelerate as current“boomers” age (AARP 2007). Despite the growing im-portance of senior markets both in size and proportionof discretionary income, this consumer segment isrelatively understudied.

In many market scenarios, consumers both young and oldwould benefit if they could learn how some unobservableoutcome of interest is related to observable attributes.For example, one could consider how antique shop-pers appraise items in a wide array of product categories,or how used car buyers learn the complex relationshipsbetween car attributes and a fair price. Efficient learningof attribute-outcome relationships is an important skillfor all consumers. However, increases in the number ofproducts available to consumers in any given category,expansion of the number of consumer-relevant catego-ries, the growing use of auction purchasing formats, andthe increasing complexity of everyday products havemagnified the need for consumers to learn how attributesrelate to price, quality, and other outcomes of interest.Research has found that repeatedly practicing in thepresence of veridical feedback will develop expertise,although both the rate and the asymptote of learningdepend on specific characteristics of the task (Alba andHutchinson 1987; Eisenstein and Hutchinson 2006, 2008;Mellers 1980; Shanteau 1992). In this paper, we compare

aged and young consumers as they learn from experiencein a simulated market, and we examine the psychologicalunderpinnings of their performance.

Our results revealed a surprising “tortoise and hare”effect. Specifically, we found that, as expected, young-sters outperformed the elderly “out of the blocks.” How-ever, aged subjects demonstrated a long-run advantage inlearning. With greater experience, older subjects contin-ued to improve, even with an unanticipated change inenvironment, while youngsters “burned out” and suf-fered degraded performance. Thus, youngsters “sprint”;the aged “marathon,” and the triumph of age over youthis not merely due to worsening performance of young-sters, but also to the surprising continued improvement ofthe older subjects.

Our findings have implications for our understanding oflearning in general and consumer learning in particular.Our results suggest that age-related degradation of foun-dational cognitive capabilities, such as information pro-cessing speed and memory, does not carry over in anexpected fashion to more “real world” tasks in whichlearning is more experiential. In addition, the fact thatolder consumers continue to learn through extendedpractice contrasts sharply with observations made instudies involving young consumers, in which subjectstypically show little performance improvement fromextended training. These contrasting findings suggestthat younger adults may approach learning situationsdifferently than the aged. New insights can also be drawnfrom the superior performance of elderly consumers onprediction tasks under conditions of market change. Thisresult is both interesting and counterintuitive given thevast literature that demonstrates reductions in processingspeed and working memory capacity associated withaging – two components that are highly involved inupdating one’s prior knowledge about the world (Heddenand Yoon 2006). References are available upon request.


For further information contact:Eric M. Eisenstein

Fox School of BusinessTemple University

527 Alter HallPhiladelphia, PA 19122Phone: 215.204.7039



ADOLESCENT BULLYING BIAS AND NEGATIVE CONSUMPTION:THE MODERATING ROLES OF SOCIAL ACTIVITY, TEACHER

SUPPORT, AND LONG-TERM GOALS

Brennan Davis, Baylor University, Waco

SUMMARY

Bullying is increasing among several adolescent seg-ments. For some adolescent victims, bullying biases suchas race represent a threat from an out-group, and accom-panying negative behavior is more externally focused.For others, bullying biases such as sexual orientationrepresent rejection of the self as unacceptable to peers,and accompanying negative behavior is more internally

focused. This paper offers meaningful ways to reducenegative behavior associated with bully victimization.Specifically, social activity helps mitigate internallyfocused behavior, while teacher support helps lowerexternally focused behavior. Finally, long terms goals aresuccessful at reducing most negative behavior associatedwith bully victimization. References are available uponrequest.

For further information contact:Brennan Davis

Baylor UniversityOne Bear Place #98007

Waco, TX 76712E-Mail: [email protected]


PRIVATE INFORMATION IN A SOCIAL WORLD: A POLICYPERSPECTIVE OF CONSUMER PRIVACY ON SOCIAL

NETWORKING SITES

Alexa K. Fox, The University of MemphisMarla B. Royne, The University of Memphis

SUMMARY

While the Internet offers consumers access to a vastquantity of information at minimal effort and cost toenable better, more efficient decision-making (Alba et al.1997), consumers are vulnerable when it comes to theironline information. In fact, many people are uncertainabout (Sheehan and Hoy 1999) how information is col-lected and shared by public entities, beyond the originalpurpose for information collection (Milne and Culnan2004). Despite the rapid growth of social media in recentyears, little academic research has explored consumers’concern for privacy when agreeing to participate in socialnetworking sites. Consumers disclose personal informa-tion to social networking sites in exchange for the oppor-tunity to connect with other members. Because a majorconcern of many consumers is that the Internet is likelyto violate their privacy (Benassi 1999), it is unclearwhether consumers truly understand how their personalinformation is being used by social networking sites andif they are concerned about this usage.

This paper discusses consumers’ concern with their pri-vacy on social networking sites and how this concernaffects consumer acceptance of such sites. Specifically,we discuss how the presentational mode of a socialnetworking site’s privacy policy may affect consumers’awareness of the subsequent effects of sharing theirpersonal information with that site, how this awarenessinfluences understanding and perceived risk of sharingtheir information with that site, and if that perceived riskaffects their perceived usefulness of the site.

Privacy and Perceived Risk

Often, marketers include privacy notices on websites tohelp the consumer decide whether or not to engage withthe website at any level (Culnan and Milberg 1998).Furthermore, consumers use a simple risk calculation indeciding whether to disclose their personal information.If consumers perceive that the estimated benefits ofinformation disclosure outweigh the costs, they will trade

their personal information, even if it puts that informationat risk (Laufer and Wolfe 1977). However, privacy notic-es are often written so the companies that create them willbe covered in case of a legal dispute; they are notdeveloped for consumers to understand them. This maylead consumers to improperly assess what really happensto the information they share with social networkingsites.

Propositions

The readability and explicitness of warnings influenceconsumers’ information processing of such warnings(Lethto and Miller 1986), and privacy notices can beviewed as analogous to warnings because of their similarfunction (Milne and Culnan 2004). Therefore, the fol-lowing proposition is offered:

P1: When privacy policies on social networking sites aredirectly stated, users will be more aware of whathappens to their information when it is shared on asocial networking site.

Consumers are only able to control and protect theironline privacy if they possess the information necessaryto make relevant choices (Nairn and Monkgol 2007) suchas an understanding of online privacy risks (Miyazakiand Fernandez 2001) and what is happening to their data(Milne, Labrecque and Cromer 2009). Therefore, weposit the following:

P2: Users with more awareness of what happens to theirshared information on a social networking site havea better opportunity to understand what happens tothat information.

Individual traits, such as a higher need for privacy, mayaffect consumers’ likelihood to provide accurate infor-mation (Sheehan and Hoy 1999). It is yet to be deter-mined how such traits affect consumer behavior in thecontext of social networks. Thus, the authors propose:


P3: Users who understand what happens to their sharedinformation on a social networking site will per-ceive risk in sharing such information, particu-larly those who are (a) older, (b) more educated,(c) more affluent, (d) female, and (e) have moreonline experience.

Most social networking sites require that a consumerexchange his or her personal information for membershipon the site. If consumers perceive risk in sharing theirpersonal information with a social networking site, theyare less likely to find value in the site. Thus, we proposethe following:

P4: If users perceive that sharing their information withsocial networking sites is a risky behavior, they willfind the social networking site to be less usefuloverall.

Discussion and Implications

If privacy policies on social networking sites are unrea-sonably long or difficult to read, consumers may beunaware of what happens to their shared information,unknowingly putting that information at risk. Some stateshave enacted statutes that deal with websites’ privacypolicies, with the California Online Privacy ProtectionAct (OPPA) being the most comprehensive. States areencouraged to follow California’s example and enacttheir own online consumer protection laws because not

all websites or social networking sites will attract usersfrom California. Furthermore, OPPA does not specifyrequirements for the modality and lexical choice used inonline privacy policies. If privacy policies are written forlegal, not consumer, protection, it may be time for OPPAto be reformed or for an amendment to be made thatcovers social networking sites. Moreover, the law shouldprohibit the use of modality markers, which are oftenused to downplay the frequency of how often the websiteshares users’ personal information (Bodle 2011). Finally,the FTC might consider creating various standardizedprivacy formats for social networking sites to adoptbased on the information to be collected as well as howthe information will be used.

In today’s society, personal information is requested inmany of our everyday activities. This paper begins toshed light on whether or not social networking sites areprotecting their users – and not just themselves – withregard to personal information disclosure policies. Futureresearch stemming from the propositions offered herecan begin to understand how consumers respond todifferent types of presentation modes and their perceptionsabout how their information is used. As social networkingsites continue to gain popularity in the lives of consumers,we must ensure that consumers are aware of and able tounderstand what is happening to the personal informationthey share on social networking sites, so they can protecttheir identities and ensure they can make the best choicespossible to safeguard their personal information.References are available upon request.

For further information contact:Alexa K. Fox

The University of Memphis359 Fogelman College of Business and Economics

Memphis, TN 38152Phone: 901.678.2667



THE ROLE OF ONLINE TALKING IN ACHIEVINGHEALTH BEHAVIOR CHANGE

Bo Liang, University of Utah, Salt Lake CityDebra L. Scammon, University of Utah, Salt Lake City

SUMMARY

Online health communication provides consumers withinstant information (e.g., dynamic resource guides forimproving health) and support (e.g., stress relief strate-gies) (e.g., Donelle and Hoffman-Goetz 2008; Eichhorn2008). Research on the impact of participation in health-related social networking sites (SNS) by consumers islimited, especially with respect to the dynamics of inter-actions among participants.

Participation in SNSs can be described in terms of themessages posted - initial and response posts, and interms of the message senders – discussion threadinitiators and followers. Users can initiate a thread bymaking an initial post (initiators), and others mayextend the thread by posting their responses (followers).After making an initial post, initiators may respond tofollowers’ responses by making subsequent posts inwhich they reveal further thoughts or actions. Previousresearch has shown that the types of social support andthe content of messages transmitted via health-relatedforums are tailored to thread initiators’ expressedneeds (Liang and Scammon 2011).

According to the transtheoretical model of behaviorchange (TTM), as people progress through stages ofbehavior change, their thoughts, feelings and actionschange (Prochaska, Redding, and Evers 2008). In thisstudy, we aim to examine whether health communicationvia SNSs facilitates changes in thread initiators’thoughts, feelings, and actions as reflected in theironline discourses. Specifically, we address the follow-ing two questions:

RQ1: Are there differences between the first and subse-quent posts by discussion thread initiators in termsof initiators’ thoughts, feelings, and actions?

RQ2: Are differences between the first and subsequentposts by discussion thread initiators related tocharacteristics of threads (e.g., sources, the num-ber of posts by initiators)?

We selected advance care planning (ACP)/advancedirective (AD) as the focus for the study. ADs are animportant means by which individuals can communicatetheir wishes for end-of-life treatment, yet for a variety ofreasons many people do not have ADs. Encouragementof ACP is an important public health initiative. Using theTTM, our study aims to investigate the extent to whichonline communication facilitates individual changesregarding AD and ACP.

We used Google’s search engine “discussion” filteringoption, which shows search results from discussion groupsand forums. We collected threaded data for two key-words: advance directive and living will. We collectedthreads in which initiators posted comments related toADs or living wills and make at least one subsequent postin response to followers’ posts. In all, we downloaded 20threads with 197 posts during the period from October 5to October 9, 2011. In this study, the unit of analysis is adiscussion thread.

For each thread, we manually coded the type of sourceinto two categories: health and non-health sites. For thetextual content of each thread, we have two files, one filecontaining the initiator’s first post, and the other filecontaining the initiator’s subsequent posts. We use Gen-eral Inquirer (GI), a text analysis software available foracademic research purposes, to code the textual contentof each thread. GI codes textual data into 182 categoriesand calculates the occurrence of concepts in each mes-sage. Based on the process of change in TTM, we selected33 categories indicating affect/emotion (e.g., emotion),cognition (e.g., thinking), and action (e.g., attempt).Change in valence is one important indicator for decisionbalance (Prochaska and Velicer 2009). We includedcategories for valence such as Positivity and Negativity.We also use categories indicating certainty to representchange in self-efficacy.

For purposes of our analysis, we considered the socialsupport from thread followers as an intervention. Thus,we grouped posts into the pre-intervention group (initialposts) the post-prevention group (subsequent posts bythe thread initiator). In our analyses we compared threads


from pre- and post- intervention groups, threads fromhealth and non-health sites, and threads in which initia-tors made two posts and more than two posts. Overall, onboth health and non-health sites, thread initiators whomade two posts showed a decrease in emotion and meansfor attaining goals, whereas, thread initiators who mademore than two posts showed an increase in gains inenlightment through thoughts and education. On healthsites, thread initiators who made more than two postsshowed an increase in action as reflected in their onlinecommunication with others.

We also used z tests to compare the proportions of thetotal word count for each variable for initial and subse-quent posts within each thread. Results suggest thatpeople who use health sites are more likely than peoplewho use non-health sites to demonstrate changes in theirthoughts, feelings, actions, or valence states in theironline messages; thread initiators had more positiveattitudes or feelings, and fewer negative attitudes orfeelings in their subsequent posts.

Our study provides important implications for publichealth policy makers, health care professionals (e.g.,

physicians, health educators) and marketers. Our studyshows that thread initiators or help seekers progressedthrough stages as they made posts over time in onlineforums. However, the changes thread initiators experi-enced were significantly associated with the types ofonline forums and the number of posts they made. Thissuggests that for a health issue, policy makers and healthcare professionals should tailor promotional or educa-tional contents to different types of online forums. Forexample, individuals who post on health sites may needmore educational guidance to progress through the stagesof change; whereas, individuals who post on non-healthsites may need more promotional support to elevate theirawareness of the urgency of the behavior change. Addi-tionally, since thread initiators who made more postsshowed more changes through the stages of change,educational campaigns encouraging individuals to ac-tively participate in computer-medicated communica-tion, especially by initiating discussions of topics impor-tant to them, may help encourage behavior change. Pol-icy makers and health care professionals may use discus-sion thread moderators to encourage thread initiators tomake more posts in their initiated threads. References areavailable upon request.

For further information contact:Bo Liang

University of Utah1645 Campus Center DriveSalt Lake City, UT 84112



ALCOHOL CULTURE: THE EFFECTS OF SOCIAL MEDIA INRELATION TO ALCOHOL CONSUMPTION

Gwyneth Howell, University of Western Sydney, AustraliaRohan Miller, University of Sydney, Australia

Dijana Apostolovski, University of Western Sydney, Australia

SUMMARY

On November 22, 2012 during end of school year cele-brations in Queensland, Australia, yet another youngperson (aged 17 years) died as the result of misadventurewhere alcohol was involved. This tragedy highlights thepublic and health policy issues surrounding alcohol par-ticularly relating to consumption by the young and vul-nerable and binge drinking. In 2011 the World HealthOrganization found that alcohol related consumptionresults in approximately 4% of all global deaths annually.In Australia where this research took place, approxi-mately 4,000 alcohol related deaths occur each year(Chikritzhs et al. 2003). Today, there are renewed callsfor social marketers and policy makers to identify factorsthat influence high levels of alcohol consumption (Grantand Dawson 1997). The Australian legal drinking is18,alcohol consumption becomes a legitimate activity andbehavioral norms commence. Those aged from 18 to 24years, or Gen Y have very different alcohol consumptiontraits, they communicate differently, they are “DigitalNatives.” Further, smart-phones and new “data” planshave facilitated the rise social media platforms especiallyFacebook’s popularity from about 100 million usersthree years ago to over 750 million users in 2011. Socialmedia not only provides a platform for sharing of contentbut also for discussion of social issues (Madhur Raj,Palak et al., 2010). This suggests Facebook is a strongdriver of contemporary culture in young adults. Bandura(2001) and Miller, Mizon, and Howell, (2012) found thatvicarious cultivation is evident and plays a significantrole in influencing behavior as participants of onlinecommunities seek a degree of incentives (Larose, Mastroet al. 2001). This study explored the real world signifi-cance from young adults’ online Facebook communica-tion in terms of influence and behavior, drawing onMiller, Mizon and Howell’s 2011 study. To meet Univer-sity ethics requirements, research subjects were requiredto opt-in to the study, of the 661 invited 115 opted-in totake part of the study, 19 declined to participate. Thestudy explored that all referenced alcohol consumptionin the past, present or future in Facebook communica-

tions during August and September 2012. The research-ers employed the coding system devised in the develop-ment of the Social Lubrication Model (Miller, Mizon andHowell, 2012). The data collected was analyzed throughaxial coding (Strauss and Corbin, 1990), message-basedcoding (Strijbos et al. 2006) and noting (Kozinets 2010),and was cross-coded by three different researchers tomake sure no intended messages from the research sub-jects were left out and to ensure the validity of theanalysis (Kolbe and Burnett 1991; Milne and Adler1999). The cross coding was over 97% in agreement.

Findings

Implicit within the Lubrication Model and specificallysocial media and word of mouth, participants send andreceive communications. The most common messagewas that they provided a reason to drink excessive amountsof alcohol in the near future. The most common reasonsused by young adults on Facebook to get intoxicated werebirthdays, weekends, days off work, exams finishing,catching up with old friends, etc. But even trivial excuseswere acceptable to justify planning to consume largeamounts of alcohol (e.g. being a Friday, or not having towork on a Sunday whilst hung over). Many young adults(22%) made clear reference they were either drinking orintoxicated. These young adults’ desire to portray them-selves as someone who stays out drinking very late atnight and is thus firmly situated within the social systemof young adult drinkers. Further, they relived the event,posting hang over comments; (18%). Many young adults(94% of the participants) use Facebook to communicatethemselves consuming alcohol or to demonstrate thatthey aspire to get intoxicated. There is pride in ignoringpublic health warning on binge drinking and announcingthat consuming anymore than 4 standard drinks in onesitting is considered binge drinking (Ravens 2008). Thisdemonstrates a lack of effectiveness of previous socialcause-related marketing, it also indicates respondentswere aware of the public health warning, but ignored the‘call to action’ message. Facebook is used as a tool toorganise meetings and events where alcohol consump-tion will take place (10.5%). Quite often intoxicated


experiences are re-lived over Facebook (23%). This styleof communications can be used to piece together theoccurrences of what happened whilst the young adultswhere intoxicated. However, often these types of alcoholrelated Facebook communications are used to prolongthe hedonic experience of getting intoxicated with friends,and, as a tool for young adults’ to portray themselves asdrinkers, and thus enable them to be further establishedwithin the social system. Alco-bullying as indentifiedand described by Miller, Mizen, and Howell (2012)occurred between peers in regard to how much someonecan drink (18). Further, as the interactive nature ofFacebook promotes third party judgments, these postingsare now ranked as of more value than the initial posting.

This research confirms Miller, Mizen, and Howell (2012)Social Lubrication Model provides considerable insightinto the communications undertaken to legitimize a cul-ture of binge drinking using Facebook. This reflects thatculture and online communications are key drivers inbinge drinking. Bandura’s (1986) conclusion that mosthuman behavior is learned by observation through mod-eling is largely still correct, although social media nowfacilitates, accelerates and amplifies peer pressure. Thisresearch has found that young adults drink excessively,

glorify the consumption of alcohol, and drink for publicappraisal by their peers. These respondents illustrate thatyoung adults strive to belong to be same social system astheir peers and fear being the outcast if they do notconform. Today, entry into social systems for the youngadults is either through social learning or accessing itthrough vicarious capability (Bandura, 2001). From apolicy perspective the alcohol culture on Facebook isinfluenced by early adopters, the majority of which aremale. Males needed to prove their prowess in terms ofalcohol consumption more than females, suggesting thatin conveying oneself as a heavy drinker emphasized theirmasculinity and reflected on their leadership status. Healthprofessionals would gain a far greater understanding howyoung adults consume alcohol and how sites such asFacebook contribute in legitimizing the act of drinking toundertake further research into social media use. Addi-tionally, the research indicates peer influence over Face-book is a relevant factor contributing to the contemporaryyoung adult drinking culture. The limitations of thisstudy is a small convenience sample of 115 participants,and many of the participants were University students; aconsumer group that has been shown to consistentlyconsume high levels of alcohol. References are availableupon request.

For further information contact:Gwyneth Howell

University of Western SydneyGreat Western Highway

Penrith, NSW 1797Australia

Phone: +61.2.98525465Fax: +61.2.98525425



IDENTIFYING THE LATENT EFFECTS OF SELECTING SAMPLESUSING EXTREME RESPONSES: LESSER

Alvin Lee, Deakin University, AustraliaDick Mizerski, The University of Western Australia, AustraliaJinchao (Alex) Yang, China Agricultural University, China

Shasha Wang, The University of Western Australia, Australia

SUMMARY

There is a potentially serious problem with the methodof selecting only respondents who choose extremeresponses from a range of responses to filter questions.For example, when selecting for non-susceptible non-smoking respondents who report no intention to smokecigarettes, respondents are required to answer with theextreme responses of “no” and “definitely not” for allsample selection questions about intentions toward smok-ing (cf., Pierce et al. 1998). Research into consumersprone to use alcohol also this sample selection method(Henriksen et al. 2008). This method for choosingresearch cohorts and samples is becoming increasinglypopular because it is argued to measure the trulycommitted (Gilpin, Pierce, and Rosbrook 1997). How-ever, this method of selection becomes a problem whenthe responses required to the filter questions are related toresponse-style biases. In the two examples, the pattern ofonly extreme negative responses at one extreme end ofthe scales appears to match ‘nay-saying’ and extreme-response bias. A sample selection that requires thisextreme response pattern is more likely to increase therecruitment of respondents with response-style biases(Baumgartner and Steenkamp 2001).

Selecting samples of respondents in this manner will notonly increase the levels of response bias in a study, butmay also increase the chance of omitting valid respon-dents from a sample (truncation bias) (Heckman 1979).For example, respondents who are exposed to family andpeer smokers may provide more moderate answers toquestions about their intentions to smoke cigarettes. If thestudy tries to identify peers and family as one of theinfluencers of future smoking, then using a sample selec-tion instrument that requires extreme answers to all of theselection questions about future smoking is likely to biasagainst these possible effects in the population. In thiscase, the potential effects to responses from exposure tofamily and peer smoking will be under-represented in thesample. This group appears to be influenced by peers and

family to drop out of a study because it is about apotentially contentious issue of adolescent cigarette smok-ing. They are also more likely to start smoking and dropout of the study. In both cases, this leads to systematicattrition which further reduces numbers of respondentswith exposure to this the peer and smoker variable. Ifextreme response bias, truncation bias and attrition biasshare the same predictor variables, and are the same as anindependent variable of interest to the study, then thesebiases are likely to aggregate to skew the study’s results(Heckman 2007).

Until now, there has been no easily accessible procedureto identify and track the effects of selecting samples usingextreme response patterns on research results. This arti-cle reports on a method called LESSER (Latent EffectsSelecting Samples of Extreme Responses) for trackingthese effects.

We tested LESSER on the Teen Longitudinal datasetfrom the California Tobacco Studies, using samples fromthis original data using the methods described in thearticle, “Tobacco Industry Promotion of Cigarettes andAdolescent Smoking” (Pierce et al. 1998). This articledescribes a sample selection procedure that requiresextreme responses to assemble a longitudinal adolescentcohort of ‘non-susceptible never-smokers’. Respondentsin this study answered questions about the way informa-tive and normative sources of influence affected theirintentions to smoke. That article reported informativeinfluence (possess a tobacco premium) to be the onlysignificant predictor of increases in intentions to smokecigarettes in that sample.

Using test samples from this original data drawn usingPierce et al.’s methods; we managed to identify theresponse style bias in the data. Respondents who wereselected for study using the extreme-response selectioninstrument gave answers with significantly higher stan-dard deviations than those who were truncated. We thencompared profiles of selected and truncated respondents.


Truncated respondents reported more exposure to familyand peer smokers (normative influence), very few ex-treme response selected respondents had friends or fam-ily that smoked. Both groups reported similar levels ofexposure to cigarette advertising and promotions.

In the third step of our test, we tracked the effects ofattrition on the sample of respondents who were selecteddue to their extreme responses to the selection questions.In this group, attrition can be predicted by reportedexposure to peer and family smokers. Those with norma-tive exposure are more likely to leave the research cohortfor the second survey in the longitudinal study. Thesefindings show that by selecting respondents through asample selection instrument that requires an extremeresponse pattern increases response style bias in the data.The researchers inadvertently omitted a key portion ofresearch candidates with higher levels of exposure to oneof the key variables being studied (normative influence tosmoke cigarettes). The numbers of this key group ofnormatively influenced respondents in the sample is

further reduced through attrition. This aggregation ofbiases greatly reduced the presence of respondents withexposure to normative influence, leaving proportionate-ly more respondents with informative influence in thecohort. It is likely that the study finding of informativeinfluence as the sole predictive factor for increases inadolescent intention to smoke cigarettes actually stemsfrom the series of research biases described.

It is important for researchers to carefully vet all sourcesof bias in their research, especially when making claimsof causality that are used to formulate public policy.Because this method of selecting samples has become theaccepted method for selecting samples to research smok-ing, drinking alcohol, drinking sugary carbonated bever-ages and consuming fast food is worrying. The biases thatlead to incorrect research findings appear to be systemic.When not vetted, these findings may lead to policymakers becoming misinformed. At the very least, thiscould lead to flawed and ineffective consumer protectionpolicies. References are available upon request.

For further information contact:Alvin Lee

School of Management and MarketingDeakin University

221 Burwood HighwayBurwood, Victoria 3125

AustraliaPhone: +61.3.92446966

Fax: +61.3.92519083E-Mail: [email protected]; [email protected]


THE EFFECT OF POLITICAL IDEOLOGY ON REACTIONSTO WARNING LABELS

Mitchel R. Murdock, University of South Carolina, ColumbiaCaglar Irmak, University of South Carolina, Columbia

James F. Thrasher, University of South Carolina, Columbia

SUMMARY

The escalating health problems stemming from unhealthyconsumer behaviors (e.g., overeating, smoking) promptgovernment officials to use a variety of public policytools. For instance, in order to deter smoking, the Foodand Drug Administration (FDA) is attempting to requiregraphic health warnings on all packages of cigarettes(Mears 2012). However, labels can have unintendedeffects (Wansink and Chandon 2006), such as inducingpsychological reactance (Brehm 1966), a motivationalstate focused on restoring a restricted freedom, which canoften have a boomerang effect where the restricted activ-ities become more attractive (Fitzsimons and Lehmann2004).

Research on reactance has shown that warning labelslead to high reactance especially when the label source isdeemed as authoritative (Bushman and Stack 1996). Forexample, college students were less likely to agree witha message advocating teaching machines when the mes-sage came from a prominent professor versus a highschool student (Brehm 1966). While the negative effectof authoritative sources in warning labels has been shownbefore, the process through which it occurs has not beenadequately examined. To do so, we focus on an importantconsumer characteristic—political ideology—to investi-gate how and why warning labels used in public policyinitiatives often become a burden rather than a boon.Political ideology is especially relevant to studying reac-tance to authoritative sources because two key compo-nents of political ideology are an individual’s beliefsregarding authority (Graham et al. 2009) and understand-ing of freedom (Durrheim 1997).

Specifically, in two studies we illustrate that whenauthoritative figures (such as the FDA) act as the sourceof the warning label (vs. no source information is provid-ed), conservatives (but not liberals) are less likely to quitsmoking (Study 1) and more likely to purchase unhealthyfoods (Study 2). These results are important as the FDAmay ironically lessen the impact of its own initiatives ifpeople know that it is the FDA that is behind the warninglabels.

In Study 1, 525 smokers viewed a series of graphic healthwarnings. The source of the graphic warning label (FDAvs. no-source control) was varied between subjects. Inthe FDA condition the health warning stated, “FDAwarning: Cigarettes are addictive” whereas in the controlcondition the health warning simply stated, “Warning:Cigarettes are addictive.” Political ideology was mea-sured by using a modified version of the political atti-tudes scale (Nail et al. 2009). The results showed asignificant two-way interaction between label conditionand political ideology (p < .05) such that in the FDA (vs.control) condition conservatives were significantlyless likely to quit smoking (p < .05) whereas the labelsource did not have an effect on liberals’ quitting inten-tions (p > .70).

In Study 2, 116 participants imagined themselves pur-chasing a pizza in a grocery store. Label (FDA vs. no-source control) was varied between subjects. The pizzahad a front-of-package nutritional label that indicated anunhealthy amount of fat, saturated fat, and salt. In addi-tion to political ideology, we measured participants’ traitreactance (Hong and Faedda 1996). Replicating Study 1,the results showed a significant two-way interactionbetween label condition and political ideology (p < .05).Further, in line with our theorizing, a significant three-way interaction between political ideology, reactance,and label condition (p < .05) showed that the effect wasdriven by conservatives who are high in reactance.

Study 3, which is currently being conducted, investigatesthe underlying mechanism of the effect by measuringdifferent aspects of reactance. Specifically, we test anexplanation based on the implication principle (Brehm1989) such that conservatives associate food-warninglabels with increased government regulation in the foodindustry.

These findings contribute to a growing body of literatureon reactance (Brehm 1966), health communication (Rainsand Turner 2007), and source influence (Kang and Herr2006) by outlining how reactance is affected by differ-ences in consumer groups and source definiteness. Twostudies illustrate across two distinct consumption behav-iors (smoking and eating) and two different intention


measures (quitting intentions and purchase likelihood)that when the FDA is the source of the warning label,conservatives are less likely to engage in the behaviorencouraged by the warning label. Further, the interactionbetween source and political ideology is moderated byreactance such that the effect is observed only for conser-vatives who are high in trait reactance. Thus, we add tothe reactance literature that has demonstrated the nega-tive effect of authoritative sources in warning labels byexamining the process through which such effects occur.

In addition, the findings of this research are significantfor public policy makers who strive to improve theeffectiveness of warning labels on regulated products.Understanding that simple changes in warning labels,such as explicitly stating the authoritative source, cansignificantly reduce their effect among certain popula-tions and possibly encourage individuals to engage in thewarned behavior encourages caution in the way warninglabels are presented to the community. References areavailable upon request.

For further information contact:Mitchel R. Murdock

University of South Carolina1705 College St.

Columbia, SC 29208Phone: 803.777.3176



HOW EFFECTIVE ARE ALTERNATIVE PLAIN PACK DESIGNS?

Philip Gendall, University of Otago, New ZealandJanet Hoek, University of Otago, New Zealand

Damien Mather, University of Otago, New ZealandRichard Edwards, University of Otago, New Zealand

SUMMARY

Smoking’s status as the leading cause of preventablepremature death has led many governments to try andminimize smoking uptake, promote cessation, and thusreduce overall smoking prevalence. Initiatives includerestricting advertising, which has consistently linkedsmoking to aspirational attributes, and regulating point-of-sale marketing, which positions smoking as pervasiveand normal.

However, these restrictions have prompted the tobaccoindustry to focus on branding, which imbues its productswith attractive personalities that hold particular appeal toyoung people. While pictorial warning labels (PWLs)challenge these alluring brand associations, the industryhas responded by altering their brand designs to mini-mize the disruptive visual impact of warnings. Evidencethat brand imagery detracts from PWLs’ impact andsalience has led to renewed calls for so-called ‘plain’packaging, where tobacco packages would feature dis-suasive colors and larger warning images instead ofbrand logos.

Studies documenting the potential effects of plain pack-aging have included qualitative work that exposed howremoval of branding strips smoking of its social cachet.Experimental studies have confirmed that plain packag-ing would diminish brand perceptions, devalue the per-ceived experience of smoking, and reduce misconcep-tions about the relative harmfulness of different cigarettevariants. Choice studies have found that plain packagingmakes cigarettes much less attractive and increases thelikelihood of smokers calling a quit line or making a quitattempt. This evidence formed the basis of Australia’sdecision to introduce plain packaging in 2012 and hasstimulated international interest in the policy. However,tobacco companies have strongly opposed plain packag-ing on the grounds it lacks proportionality, and that other,less ‘radical’, measures could achieve the same outcomeswhile imposing fewer costs.

Marketers have long used colors’ intrinsic meanings toattract consumers’ attention, differentiate their brandsfrom those of competitors, and cue brand perceptions.Theoretically, colors contribute to a brand’s personalityas consumers associate specific colors with particularattributes that, over time, form distinctive heuristics.Changing such associations by substituting the brand’simagery with an aversive color creates dissonance amongexisting consumers and deters future users. This is thereason for tobacco companies’ hostility to plain packag-ing proposals. Evidence that color may have an evenstronger aversive effect than PWLs has raised furtherquestions about how plain packaging works and whetherall components are proportional in their effect.

We tested the relative effects of brand imagery, dissua-sive colors and PWL size and theme, and estimated theeffects on smokers’ and non-smokers’ choice behaviorswhen aversive pack elements were combined with brandattributes. Our sample comprised 1044 young adultsaged between 18 and 30; the sample was stratified byethnicity, gender and smoking status (daily, occasional,former and never smokers). We used a fractional factorialdesign involving two blocks of six choice sets, each setcontaining three cigarette pack images with differentattributes. Respondents were randomly presented withone block of images and within each block the choice setswere also randomly presented. Smokers were asked toidentify the packs they would be most and least likely tobuy for themselves, while non-smokers were asked toidentify the packs they thought would be most and leastlikely to prompt a friend who smoked to quit. The datawere weighted to ensure respondents’ age, gender, eth-nicity and smoking status matched the New Zealandpopulation according to the latest census. Multinomiallogit model parameters unique to smokers and non-smokers were estimated for each classification level ofeach design factor (message, warning, pack branding andbrand) using SAS PHREG.

Pack color and branding level were the most importantattributes. A fully plain pack was significantly less


appealing to smokers than a partly plain pack (featuringa brand logo), and both were significantly less attrac-tive than a fully branded pack (p<.0001). Non-smok-ers wanted their friend to give up smoking, thus thereverse was the case for them; a fully plain pack wassignificantly more appealing than a partly pain pack andboth were significantly more appealing than a fullybranded pack. In both cases the conclusion is the same:replacing brand imagery with a dissuasive color andstandardized font significantly increases the aversive-ness of a cigarette pack. Warning size was the secondmost important attribute: the larger the pictorial warning,the more aversive it was. The 75% warning was signifi-cantly more aversive than the 50% warning, which wasonly marginally more effective than the 30% warning(p<.001 for smokers and p < .05 for non-smokers).

Although aversive colors have been used in tobaccowarning labels to caution users, the use of color todissuade cigarette purchases is a novel and potentiallypowerful policy measure. Our results showed that adissuasive color had a stronger effect on both smokers’and non-smokers’ choice behaviors than PWLs, eventhough the warnings used were novel, and thus likely tohave a strong impact. However, residual brand echoes,such as a logo, significantly enhanced the appeal of acigarette pack, provided smokers with on-going utility,and mitigated the overall effect of the dissuasive color.For policy makers, the findings highlight the proportion-ality of fully plain packs, since these, in combination withvery large PWLs, will most effectively reducing smok-ing’s attractiveness. References are available uponrequest.

For further information contact:Janet Hoek

Department of MarketingUniversity of Otago

P.O. Box 56Dunedin 9054New Zealand

Phone: +643.479.7692E-Mail: [email protected]


DO DISSONANT AD VISUALS CAUSE CONSUMERS TO DISCOUNTPRESCRIPTION DRUG SIDE EFFECTS?

David J. Hagenbuch, Messiah College, Mechanicsburg

SUMMARY

In most video-based advertising, redundant visuals rein-force the ads’ verbal messages. In television commercialsfor prescription drugs, however, narration of the drugs’negative side effects is often paired with dissonant visu-als, or pleasant images, which seem inconsistent with thedrugs’ sometimes very serious risks. Through an empir-ical analysis, this study found that consumers attributesignificantly more positive affect and lower risk to pre-scription drug ads that employ dissonant visuals (imagesthat did not represent an ad’s verbal content) duringnarration of side effects than to those with redundantside-effect visuals (images that accurately reflected anad’s verbal content).

Various studies have investigated the effects of positiveand negative ad content (Bolls, Lang, and Potter 2001;Lang, Dhillon, and Dong 1995; Shadel, Niaura, andAbrams 2002) as well as consumers’ perceptions ofassociated risk (Cox, Cox, and Zimet 2006; Lee and Shin2011). Likewise, certain research has explored ad mes-saging and risk communication in direct-to-consumer(DTC) ads for prescription drugs (Cline and Young 2005;Macias, Lewis, and Tae Hyun 2010; Norris, Bailey,Bolls, and Wise 2012; Macias, Pashupati, and Lewis2007; Cox, Cox, and Mantel 2010; Grenard, Uy, Pagan,and Frosch 2011). More specifically, Glinert (2005)identified several DTC advertising practices that pre-sented potential challenges for consumers, includingnarration about drug risks that contradicted the ads’positive images. These results were consistent with Fox’s(2004) finding that greater comprehension was associ-ated with redundant visuals than with dissonant visu-als. In addition, the main organization responsible foroverseeing prescription drug advertising in the UnitedStates, the U.S. Food and Drug Administration (FDA),has sponsored a variety of studies on DTC drug adsthrough its Office of Prescription Drug Promotion(OPDP). However, none of its studies have found

evidence that inconsistency between risk narration andvisuals influenced people’s feelings toward prescriptiondrugs or their understanding of risk information (U.S.FDA 2011a, 2011b, 2011c).

In light of the extant research, the current study set outto empirically investigate whether consumers wouldexhibit more positive affect toward prescription drugads and rate their side effects as less serious when theads employed dissonant visuals (pleasant images) thanwhen the ads used redundant visuals (negative images).To test consumers’ reactions, the researchers created aseries of original audiovisual ads that featured threehypothetical prescription drugs. The ads were shown totwo different sample groups: a control group and atreatment group. Each sample’s ads were identical exceptfor the pictures shown during narration of the drugs’negative side effects. In keeping with many real prescrip-tion drug commercials, the control group saw ads thatcontained pleasant images (dissonant visuals) duringnarration of the drugs side effects. In contrast, the adsshown to the treatment group contained negative images(redundant visuals) during the side effects. The twosample groups, which were of very similar size andcomposition, together totaled 583 participants and con-sisted of faculty, staff, and students from a private collegeof the liberal arts and applied sciences located in thenortheast United States.

The results of the study affirmed its two main hypotheses:Individuals who saw the ads with the select dissonantvisuals (pleasant images) exhibited significantly morepositive affect toward the ads and rated the drug risks assignificantly less serious than did those who saw the adswith redundant visuals (negative images). These resultssuggest that the pharmaceutical industry’s practice ofpairing narration of negative side effects with pleasantimages may be misleading consumers and violating theFDA’s fair balance standard, which provides good rea-son for public policy change. References are availableupon request.


For further information contact:David J. Hagenbuch

Messiah CollegeOne College Avenue, Suite 3042

Mechanicsburg, PA 17055Phone: 717.766.2511, Ext. 7256



THE PRESENTATION OF QUANTITATIVE BENEFITINFORMATION IN DIRECT-TO-CONSUMER

TELEVISION AND PRINT ADVERTISEMENTSFOR PRESCRIPTION DRUGS

Helen W. Sullivan, U.S. Food and Drug Administration, Silver SpringAmie C. O’Donoghue, U.S. Food and Drug Administration, Silver Spring

Kathryn J. Aikin, U.S. Food and Drug Administration, Silver SpringDhuly Chowdhury, RTI International, Research Triangle Park

Rebecca R. Moultrie, RTI International, Research Triangle ParkDouglas J. Rupert, RTI International, Research Triangle Park

SUMMARY

Understanding of risk and benefit information is crucialto the safe and effective use of prescription drugs. Thisstudy evaluated the effect of including quantitative ben-efit information in various statistical and visual formats(e.g., relative or absolute frequency, bar graphs, tables) indirect-to-consumer (DTC) print and television advertise-ments (ads) on immediate recall, attitude toward the drug,and perceived benefit and risk. The study was guided bythe following research questions: (1) Does presentingquantitative benefit information in a statistical format inDTC ads help people recall quantitative benefit informa-tion in DTC ads? If so, which statistical formats are mosthelpful? and (2) Do visual aids help people recall quan-titative benefit information in DTC ads? If so, whichtypes of visuals are most helpful?

To answer these questions, we designed and imple-mented a randomized, controlled study exposing partic-ipants to a DTC prescription drug ad for a mock drugcontaining quantitative benefit information. Participantssaw either a print DTC ad or a television DTC ad. The adcontained information about either a high-efficacy or alow-efficacy cholesterol drug. This benefit informationabout the drug was presented either in a statistical formator a visual format. The statistical formats tested wereabsolute frequency (for example, 65 out of 100), percent(for example, 65%), relative frequency (for example, 33times more likely), a combination of absolute frequencyand percent, and a combination of relative frequency andpercent. The visual formats tested were pie charts, barcharts, tables, and pictographs. Participants in a controlcondition saw an ad without quantitative benefit informa-tion. Approximately 4,800 participants who had been

diagnosed with high cholesterol responded to the studyvia the Internet.

Findings

Statistical Format

• Participants who did not see any quantitative benefitinformation about the drug were the likely to accu-rately report how well the drug worked.

• There was a match between the kind of quantitativeinformation participants viewed and the kind ofquantitative information participants were able toaccurately report. For instance, participants whoviewed the benefit information as an absolute fre-quency (for example, 65 out of 100), compared withthose who did not see any quantitative benefit infor-mation, were better able to report how well the drugworked as an absolute frequency and a percent butnot as a relative frequency (for example, 33 timesbetter).

• In general, participants who saw the benefit informa-tion presented in two formats (for example, 65 out of100 and 65%) were the most likely to accuratelyreport how well the drug worked.

• The statistical format that participants saw did notaffect their ability to recall the drug’s risks, theirattitude toward the drug, their perceptions of howwell the drug works and how risky it is, or theirintentions to get more information about the drug orto take the drug.


Visual Format

• When viewing print ads, participants who saw a barchart or table, compared with those who saw novisual display, were more likely to accurately recallhow well the drug worked. The bar chart was alsobetter than the pictograph, and the table was betterthan the pie chart at helping participants accuratelyrecall how well the drug worked.

• When viewing television ads, participants who sawany visual display, compared with those who saw novisual display, were more likely to accurately recallhow well the drug worked. The bar chart was alsobetter at helping participants accurately recall howwell the drug worked than the pictograph and thetable.

• The type of visual display that participants saw didnot affect their ability to recall the drug’s risks, theirattitude toward the drug, their perceptions of howwell the drug works and how risky it is, or theirintentions to get more information about the drug orto take the drug.

Conclusions

Our findings demonstrate that participants can accuratelyrecall quantitative benefit information in DTC prescrip-

tion drug print and television ads, and that providing thisinformation does not adversely influence their recall orperceptions of the product’s risk. Overall, presentinginformation using absolute frequency and percent for-mats may be best at helping participants accurately recallhow well a drug works. Presenting a visual aid alsoappears to help participants accurately recall how well adrug works, with bar charts and tables demonstratingadvantages over other visual formats. In general, provid-ing quantitative benefit information to participants en-ables them to see the information and answer questionsabout it correctly, although it does not necessarily changetheir attitude toward the drug, their perception of howwell the drug works and how risky it is, or their intentionsto get more information about the drug or to take the drug.At the same time, including quantitative benefit informa-tion did not have a detrimental effect on the recall of riskinformation. Thus, the inclusion of quantitative benefitinformation in DTC print and television ads may have thepotential to help people make informed decisions aboutspeaking with their health care professional about pre-scription drugs.

A major contribution of this research is that it is the firststudy to systematically examine the addition of quantita-tive information in television DTC ads, a modality thathas not previously been explored in the risk communica-tion literature. References are available upon request.

For further information contact:Amie O’Donoghue

U.S. Food and Drug Administration10903 New Hampshire Ave., Bldg. 51, Rm. 3236

Silver Spring, MD 20993–0002Phone: 301.796.1200



NO GOOD DEED GOES UNPUNISHED: CITING SOURCES INPRINT ADVERTISING CLAIMS

Leslie Koppenhafer, University of Oregon, EugeneCatherine Armstrong Soule, University of Oregon, Eugene

Robert Madrigal, University of Oregon, Eugene

ABSTRACT

The research investigates whether source citation of printadvertisement claims benefits consumers by aiding indeception detection in different risk domains. Threestudies test whether source citations have the intendedeffect of increased consumer trust and reduced deceptionperceptions. Additionally, the potential moderators risklevel and processing style are explored.

INTRODUCTION

Consumers are frequently exposed to advertisements thatmake claims about product performance. When an adver-tisement claims that a product has a “77% improvementover the competition,” for example, does knowing wherethat number came from help consumers form opinionsabout the advertiser’s deceptive intent? If an advertiserfails to explicitly state the evidence for a product claim,would you consider the advertisement to be at leastsomewhat deceptive? Although print advertisement claimsare regulated by the Federal Trade Commission and arerequired to “have evidence to back up their claims,”according to the Bureau of Consumer Protection, thisinformation is not required by federal law to be explicitlystated within the advertisement (www.business.ftc.gov).The “reasonable basis” requirement states that advertis-ers must have objective evidence in the form of consumersurveys, sales data, clinical testing or scientific evidenceusing methods deemed acceptable by experts in the fieldto support the claim. In an effort to increase consumerprotections, many print media publishers are proac-tively and voluntarily going further than the currentlegal requirements. The publishers’ self-regulationpractice now requires advertisers to include supportinginformation for the claim within the body of the adver-tisement. It is important to explore whether this well-intentioned action may have undeserved and detrimentaleffects for marketers.

Source citation is additional text that acts as furtherexplanation of an advertising claim. The absence (orpresence) of a source citation may activate consumers’

deception-protection processes. One possibility is thatwhen provided with legitimate support informationregarding a claim within an advertisement, consumerswill rightfully feel more confidence in the credibility ofthe claim. However, it is also possible that consumersfaced with source information may become more sensi-tized and instead trust the claim less when it is supportedwithin the advertisement. Extra information may act as aflag to consumers, reminding them the advertisement isa persuasive attempt and should be evaluated as such. Afinal possibility is that the additional information doesnot affect the way the claim is perceived and interpreted.We argue herein that the relationship between sourcecitation of advertisement claims and consumer deceptionperceptions is complex. The answer to whether the pres-ence of a source citation causes an advertisement to bejudged as more credible and less deceptive is in fact bothyes and no and is moderated by processing style. Consid-ering work in marketplace deception, persuasion knowl-edge and protection, risk perceptions and dual process-ing, we begin to disentangle the effects of source citationof print advertising claims on deceptive perceptions.

LITERATURE REVIEW

Dimensions of Risk

The concept of perceived risk is an important antecedentto many marketing decisions, not only in deceptiondetection, but also in product judgments, new productadoption, brand/store loyalty and purchase choice (forreviews see Mitchell 1999; Ross 1975). People deter-mine risk level by thinking a priori about the potentialoutcomes of a decision combined with the probabilitiesof each outcome’s occurrence (Dowling 1986). Risk in aconsumer behavior context can be more narrowly de-fined as the uncertainty or ignorance surrounding theproduct, ad or stimuli in question and the related conse-quences (Campbell and Goodstein 2001).

Risk is considered a multifaceted construct and can bedeconstructed into many distinct domains, such as per-formance, social, physical, financial, psychological, time,


etc. (e.g., Bettman 1973; Kaplan, Szybillo, and Jacoby1974). Consumer behavior research uses risk as a defin-ing factor in attitude development and subsequent judg-ments about the deceptive intent of advertisers, as well asconsidering the risk directly related to acquisition of atarget product. Two risk domains especially importantto consumer behavior are financial and physical risk.Because products vary in price from very expensive toinexpensive and others can protect from or cause physi-cal harm, these two domains may be the most influentialin product judgments. Research in consumer behaviorand risk tends to assume that high involvement prod-ucts (expensive or potentially physically harmful) areinherently perceived to contain risk (Dowling 1986),but to our knowledge there is no research investigatingthe link between risk level, risk domain and deceptionperceptions.

Financial risk is the most naturally related to marketingcontexts, as price is one of the most important dimen-sions considered by consumers (risk, quality, etc.)(Monroe 1979). We expect that advertisements repre-senting higher financial risk cause individuals to auto-matically engage in more involved mental processingand those representing lower financial risk will encour-age more superficial processing. Physical risk is also animportant domain for consumer behavior research. Phys-ical risk in a consumer behavior context is often investi-gated in relation to prescription drug advertisements (fora recent example, see Cox, Cox and Mantel 2010).Results from other physical risk studies have been decid-edly mixed for the effect of risk information on productliking and/or purchase intent, which may partially stemfrom trust in the advertising claim and perceptions ofdeceptive intent. The current research considers whetherhigher risk engages more effortful processing and resultsin less perceived deception of an advertising claim that issupported by a source citation.

Dual Processing

Risk level is one critical dimension which causes peopleto process stimuli differently. People are “cognitive mi-sers” (Hastie and Dawes 2010) who exert the minimalamount of effort needed to make a “good” decision,conserving mental energy for more important decisions.Dual processing helps people achieve good decisionswith minimal amount of effort; people naturally use amental energy conservation process and limit theamount of processing needed for minor decisions.People make decisions about how much cognitiveenergy to exert in any given situation. System 1 pro-cessing is quick, instinctual, and implicit and System 2

is effortful, cognitive and deliberative and “rule based”(Evans 2003, Evans 2008). System 1 typically guidesbehavior the majority of the time, with System 2 usedonly when decisions are more complex, difficult andimportant to the individual (Bargh and Chartrand 1999).

The concept of System 1 vs. 2 processing can be appliedto the context of source citation within print advertise-ments. When using System 1, omission of the citation ofthe claim is acceptable, as people want to make quick andeasy decisions and are not looking to incorporate extrainformation. However, when System 2 processing isengaged the additional information provided by the sourcecitation is incorporated in reaching a judgment about theacceptability of the claim.

Advertisements for medications with serious side effects(high risk) are commonly seen next to advertisements forinexpensive personal hygiene products and fast movingconsumer goods assumed to have minimal risk, leadingpeople to shift between using System 1 and 2 processingbased on the level of perceived risk (Baumeister 2008).When perceived risk is high, a consumer viewing anadvertisement should naturally engage in System 2 pro-cessing, with source citations adding credibility to theclaim and resulting in lower deception perceptions. How-ever in a low risk context, a consumer often uses System1 (processing at a more superficial level) and may inter-pret the same additional information differently. Mentalenergy is another factor relevant to print advertisementcontexts, in that even when the product advertised repre-sents a high risk, the viewer is most likely not giving thead full attention and is processing it in a superficialmanner (System 1). Engaging in System 2 processing isone tactic that consumers can use to protect themselvesfrom deceptive messages (Freistad and Wright 1994).

Deception in the Marketplace

Protecting consumers from marketplace deception isbecoming increasingly important in the marketing field.One way to protect consumers is to prevent deceptiveadvertisements. An advertisement should be considereddeceptive if it contains inaccurate information or omitsrelevant information, causing a consumer to make adifferent decision or judgment than he or she wouldotherwise (Boush, Friestad and Wright 2009). Whenadvertisers create content, they are subject not only to theconsumer protection laws of the government, but alsothose guidelines set forth by the channel, publisher, etc.,where the advertisement will appear. When comparedto other advertising mediums, print advertising hasremained relatively under-policed until recently. A com-


pany must have a “reasonable basis,” such as a survey,test, study or experiment that results in “competent andreliable scientific evidence” before running an advertise-ment that makes a claim. Although the FTC requires thatadvertisers must have done the research to back up theirclaims, it does not require them to actually print orotherwise disclose that data within the advertisement(stricter laws regarding certain products, such as pre-scription drugs, are exceptions).

There are currently no federal laws regarding citing ofsources in print advertisement claims (for example: “BrandX’s skin moisturizer is 50% more effective than theleading competitor”). However, two of the world’s larg-est publishing companies, Hearst Publications and TimeInc., have enacted standards requiring advertisers to citethe source of such comparative claims within the adver-tisement in order to assure readers of the accuracy of theclaims. One can expect smaller media companies tofollow suit and these guidelines could realistically beextended to all verifiable claims (not only comparative)and possibly become federal law. Additionally, manycompanies concerned with consumers’ perceptions vol-untarily include source citation information within theadvertisement to verify the legitimacy of claims. Obvi-ously, these rules and actions are established to protectconsumers from bogus claims. Although the intentionbehind the publishing companies’ new requirement isquite honorable, we suggest that the inclusion of sourceinformation within print advertisements may haveunintended negative effects for both consumers andadvertisers.

Research exists regarding the way that consumers pro-cess and react to marketing messages and, more specifi-cally, deceptive marketing messages (e.g., Darke andRitchie 2007). Marketing communications are consid-ered and rightly viewed by consumers as persuasionattempts. Consumers develop and use different tactics inorder to understand a persuasion attempt. According tothe persuasion knowledge model (PKM), consumershave perceptions of their own self knowledge of the topic(product being marketed), their own knowledge aboutpersuasion in general and their beliefs about the actualagent behind the message (Friestad and Wright 1994).Each individual uses these different types of knowledgeto determine the legitimacy of a message. The importanceof consumers’ thoughts about their own abilities to detectdeceptions may also lead them to form opinions aboutwhether a particular persuasion attempt is acceptable ordeceptive (Friestad and Wright 1994).

Ad Credibility

Ad credibility is defined as “the consumer’s evaluation ofthe truth and believability of the advertisement” and isrelated to visuals as well as textual information (Cotte,Coulter, and Moore 2005). Perceived credibility of anadvertisement is an important antecedent of a consumer’soverall judgment. When ads are perceived as more cred-ible, overall attitude toward the ad is more favorable(Petty and Cacioppo 1986) and the ad is more effective(Goldberg and Hartwick 1990). We propose that:

H1: Individuals will perceive an advertisement with asource citation present (vs. not present) as morecredible.

Cotte et al. (2005) find that consumers’ evaluations ofcredibility covary negatively with manipulative intent.Manipulative intent is defined as consumers’ perceptionthat an advertiser is attempting to persuade throughinappropriate, unfair or manipulative means (Campbell1995) Because perceived manipulative intent shouldinfluence judgments about the deceptiveness of an adver-tisement, ad credibility is related to whether the ad isdeemed to be deceptive or not.

Advertisements that make claims without providing thesource of the information are deceptive according tomany definitions (Boush et al. 2009; Kardes, Posavac,and Cronley 2004; Shimp 1979). They are classified as acase of omission, the act of leaving out information thatcould be considered crucial to correct deception identifi-cation - an attempt to hide the true reality (Boush et al.2009). In a print advertisement, marketers may legallyomit the information a reader would need to evaluate thecredibility of a study, such as the context, sample sizeor where and when data was collected and by whom.Consumers may interpret claims where no source in-formation is provided as more deceptive, relative tothose advertisements explicitly identifying the source(whether fine print or normal text). Because the pres-ence of a source citation provides the consumer withmore information and enhances credibility, we expect:

H2: Individuals will perceive an advertisement with asource citation present (vs. not present) as lessdeceptive.

Consumers who devote more cognitive effort to a mar-keting message may make a better decision (i.e., moreaccurately determine whether an ad claim is acceptable or


not) (Friestad and Wright 1994). When a decision is ofminimal importance, it is likely that a consumer will useSystem 1 processing, making decisions using a quick,“gut” instinct to judge (Bargh 2002). When a consumeris more involved in an evaluative experience, he or she islikely to use System 2 processing, slowing down andrationally considering all available information with morecaution. When exposed to a persuasive attempt, such asa print advertisement with a claim, consumers naturallybegin to processes new information using the processingstyle automatically selected by the level of risk involvedand the available mental energy at the time of viewing.

Different processing styles may lead to different interpre-tation of the same information. In the following studies,we explore whether the use of different processing stylesof the same information (source citation of an advertis-ing claim) causes differential judgments. When theconsumer opts to invest minimal effort and make a snapdecision, source citation omission may be acceptable(prevents the mind from unnecessarily “slowing down”).However, when the mind chooses to devote energy andprocess all information, the citation information becomesvaluable, helpful and important. Processing style can bedetermined by the perceived importance of the judgment(a high risk product is more important and requiresSystem 2) or by available mental energy (when an indi-vidual’s mental processes are constrained, she/he willrely on System 1) Therefore, we hypothesize:

H3: Individuals using System Two (vs. System 1) pro-cessing will perceive an advertisement with sourcecitation present (vs. not present) as less deceptive.

STUDY 1: FINANCIAL RISK

Method

In exchange for partial course credit, 110 undergraduatebusiness students (58.4% male, M

age = 21.27 years; SD =

1.50) at a major northwest university participated in a labstudy conducted in fall quarter of 2010 and winter quarterof 2011. The study used a 2 (financial risk: high, low)by 2 (source citation: present, absent) factor betweensubjects experimental design with conditions randomlyassigned. For the risk manipulation, the participants readone of two possible descriptions of the product, stereoheadphones, which were identical except one repre-sented low financial risk (price = $34.99) and one repre-sented greater financial risk (price = $199.99). Next,participants viewed one of two versions of a print adver-tisement for headphones containing a numeric claim(77% improvement in sound quality over competition).The only difference between the two stimuli was one

version contained an asterisk and fine print source cita-tion at the bottom of the advertisement, explaining thebasis for the sound improvement claim (Research isbased on proven scientific laboratory testing methodsconducted by independently certified Listen Inc. andcomprised of Sound Check electroacoustic test and sys-tem and a Bruel & Kjaer head and torso simulator vs. thetop 5 national brands), and the other version contained nosource citation. For copies of stimuli used in all studies,see appendix.

Participants then reported perceived cost of the head-phones and the likelihood of purchase in the productcategory and for the specific product advertised. Next,attitudes towards the product and the advertisement werereported on 3 item, 7 point bipolar scales (bad/good,negative/positive and unfavorable/favorable). Ad credi-bility was measured on a 4-item (realistic, truthful, cred-ible and believable) 5 point Likert scale. Participants alsorated the ad on a 7-point Likert scale from “not deceptiveat all” to “extremely deceptive” and provided open-ended responses regarding how they determined whetherthe ad was deceptive or not. Finally, demographic infor-mation as well as an attention check item (source citationpresence) was collected for all participants. Participantswho were unable to correctly recall the price of theheadphones or whether a source was present or absentfor the advertising claim were removed from furtheranalysis.


Analysis indicates a significant main effect of price pointon perceived financial risk, confirming the manipulationof risk was successful. Participants who viewed theadvertisement for the less expensive headphones, M

low

risk = 9.96, felt that the product was less financially risky

than those who viewed the more expensive version, Mhigh

risk = 15.16, F(1,108) = 87.154, p < 0.001. An analysis of

variance found support for H1. Participants who viewedan advertisement with a source citation present viewed itas more credible, F(1,106) = 15.009, p < 0.001.

An ANOVA indicates partial support for H2 as there is asignificant interaction, F(1,106) = 4.019, p = 0.048. In thehigh risk condition, if the citation is present, participantsuse the information correctly by perceiving it as lessdeceptive. However, when it is low risk, participants donot perceive differences in deception based on sourcepresence or absence.

The results show consumers use the presence or absenceof source citation information correctly when determin-ing credibility; however, this judgment does not always


influence deception perceptions. The presence of a cita-tion is likely a heuristic signaling that the claim is backedup and credible. However, consumers use the presence orabsence of a source citation differently in judging per-ceived deception depending on the perceived level offinancial risk. Participants in high risk conditions may beengaging in System 2 processing, expecting and needingmore information to make a better judgment of deceptiveintent. The additional information enables them to cor-rectly evaluate the advertisement as less deceptive whenthe source citation is present. Analysis of the qualitativeresponses supports this interpretation. A participant inthe high risk, source present condition who found theadvertisement not deceptive reported: “They explainedhow they compiled the data on the bottom of the ad. I haveno reason not to believe what they say.” Meanwhile, aparticipant in the high risk, source absent conditionstated: “It gave a percent for how well it works. But wedon’t (sic) know where that number is coming from or ifit is accurate.” However, it appears that in low financialrisk situations the additional information provided by asource citation is not used to determine deception percep-tions, as no significant difference exists when the sourcecitation is present versus absent. It appears participantsare using System 1 processing when the product is lowrisk and did not need or desire extra information toprocess and judge the deceptive intent advertisement.

STUDY 2: PHYSICAL RISK

To ensure the findings were not limited to a financial riskcontext, Study 2 sought to replicate findings from Study1 in the physical risk domain. The methodology andmeasures were the same as described in Study 1.

Method

In exchange for partial course credit, 124 undergraduatebusiness students (69.3% male, mean age = 21.27 years;SD = 2.83) at the same university participated in a labstudy conducted in spring and summer quarters of 2011.The study used a 2 (physical risk: high, low) by 2 (sourcecitation: present, absent) factor between subjects exper-imental design with randomly assigned conditions. Par-ticipants saw one of two print advertisements for a newvaccine strip, both of which contained a numeric claim(77% of people prefer over injections). The only differ-ence between versions was that one contained nosource citation and the other contained an asterisk andfine print source citation (Based on a survey reportingactual patient levels of pain, anxiety and efficacy whencompared to traditional vaccine injections. Research con-ducted by the Food & Drug Administration in October2010 with 1,016 respondents). For the risk manipulation,the participants read a brief description of the product,

TABLE 1Ad Credibility (Studies 1, 2, and 3)

Cronbach’s ααααα Citation Present (Mean) Citation Absent (Mean) Significance

Study 1 .863 13.96 11.89 p < 0.001Study 2 .889 12.66 10.81 p = 0.005Study 3 .877 14.14 12.89 p = 0.013

TABLE 2Means and Standard Deviations for Deception, Study 1– Financial Risk

Source Absent Source Present Change F Significance

Low Risk ($34.99) 2.93 (1.538) 3.33 (1.569) Δ = 0.40 -0.933 p = .338

High Risk ($199.99) 3.61 (1.499) 2.85 (1.460) Δ = -0.76 3.580 p = .064


either representing low physical risk (safe alternative toinjections) or one representing greater physical risk (num-ber of potential risks when compared to injections).


Participants who viewed the advertisement for the vac-cine with no side effects felt the product was more safe,M

low risk = 4.04 compared to those who viewed the product

with adverse side effects, Mhigh risk

= 2.56, F(1,124) =33.573, p < 0.001. Replicating Study 1 findings, therewas a main effect of source citation presence. Partici-pants perceived ads with citations as more credible com-pared to those without, F(1,121) = 8.221, p = .005,providing further support for H1. Analysis of variancefinds the same pattern of effects from Study 1 partiallysupporting H2. When risk was high, participants per-ceived more deception when an ad lacks source citationalthough only marginally significant F(1,120) = 2.755,p = .100.

These results further support that people use presenceor absence of source citation information differentlydepending on the level of physical risk. Citation presencemakes the ad more credible in both low and high riskcontexts, but only in high risk contexts does that lack ofcitation translate into increased perceptions of deception.Overall, source citation presence significantly influencesthe deception perceptions in high physical risk contexts,as expected with System 2 processing. A source citationprovides reassurance that the advertisement is not decep-tive and its absence indicates some form of deception, asillustrated by participants’ qualitative comments. A par-ticipant in the high risk, source absent condition said:“Where are the statistics from? Who are the 77% ofpeople that want these strips. I’ve never heard of them. Itjust sounds like b.s. There is no proof or evidence.” A

high risk, source citation present participant stated: “Therewas very little room for deception- the ad stated a fact andfollowed it up with its source (FDA).” However, lowphysical risk participants do not seem to use the addi-tional information provided by a source citation in form-ing deception perceptions. When using System 1 pro-cessing, additional information is not expected nordesired.

STUDY 3: PROCESSING STYLES

Study 3 looks at the effects of cognitive processing stylesused by a viewer in the presence or absence of a sourcecitation. We explore whether the effects found in Studies1 and 2 were indeed due to processing style by manipu-lating cognitive load to induce half the participants to useSystem 1 processing, even in a high risk context. Themethodology and measures are identical to those in theprevious two studies.

Method

One hundred fifty-two undergraduate business studentsparticipated in a lab study conducted in spring and sum-mer quarters of 2011 in exchange for partial coursecredit. The study used a 2 (cognitive load: high, low) by2 (source citation: present, absent) between subjectsexperimental design with conditions randomly assigned.Participants were first presented with a seemingly sepa-rate task designed to compare the memory skills ofstudents at their university to other students. Participantswere asked to remember a number (39 or 8382154) andtold they would report it after completing other tasks butthey should do their best to remember the number. Next,participants viewed one of two versions of a print adver-tisement for headphones containing a numeric claim(77% improvement in sound quality over competition).

TABLE 3Means and Standard Deviations for Deception, Study 2 – Physical Risk

Source Absent Source Present Change F Significance

Low Risk(No side effects) 4.46 (1.630) 5.00 (2.075) Δ = 0.54 1.097 .300

High Risk(Significant side effects) 5.54 (2.301) 4.77 (2.349) Δ = -0.77 1.843 .179


The two stimuli are the same as used in Study 1 and highfinancial risk level was held constant (price = $199.99).


Participants under a high cognitive load (longer number)took significantly less time to view the advertisement(M

High load=12.68 seconds) compared to those under a low

cognitive load (shorter number) (MLow load

= 16.13 sec-onds), F(1,146) = 8.031, p = .005. Because high loadparticipants were cognitively busy, they had to resort toSystem 1 processing, even though the high risk levelwould normally encourage System 2 processing. Wewould expect those under a high load to process citationinformation and form judgments in the same pattern asthose in the low risk groups in previous studies due toprocessing style. Again, we find support for H1 whereads with source citations are perceived to be more cred-ible, F(1,149) = 6.270, p = .013. However, the previouslysupported pattern of results in risk level findings does notemerge in this study. We expected those using System 2processing (low load group) would find ads with sourcecitation absent more deceptive than those which con-tained citations, whereas those using System 1 wouldperceive no differences in deception. We find no maineffect of processing style or citation presence, and theinteraction was not significant F(1, 149) = .111, n.s.,failing to support H3. Although there were no differencesin perceived deception magnitude for either cognitiveload conditions, the reaction time data does suggest thatparticipants are using different processing styles to eval-uate the advertisements. Future research should focus onuntangling the impact of these different processing styleson perceived deceptive intent.

GENERAL DISCUSSION

The findings from these studies are meaningful for con-sumers, regulators, practitioners and print media publish-ers. In real contexts, advertisements are often sand-wiched between articles and other advertisements and theact of viewing them rarely requires or achieves theconsumer’s full attention. The research extends theoret-ical understanding of consumer deception perception bydemonstrating presence or absence of a source citation isa cue consumers use differently depending on perceivedrisk level. The findings provide optimism regardingconsumers’ abilities to protect themselves from decep-tion and lead to interesting future research questions.

As these studies demonstrate, the manner in which con-sumers process the presence or absence of a sourcecitation in a print advertisement is not straightforward. In

financial and physical risk domains, participants’ decep-tion perceptions are moderated by perceived risk. Thepresence of the source citation adds credibility to theclaim but only reduces perceived deceptive intent in highrisk contexts. This is good news for consumers becausewhen it matters most (high risk), they tend to make betterjudgments about the ads. However, when people per-ceive low levels of financial and physical risk, they seemunable to correctly identify deceptive intent. The pres-ence of a source citation in a low risk context is additionalinformation that does not influence deception percep-tions, indicating consumers are not willing to exert theextra mental energy required to make the best possibledecisions in low risk situations. Additionally, the fact thatconsumers under low cognitive load are unable to cor-rectly identify deceptive ads better than those who aredistracted is concerning, especially given that print ad-vertisements are 1) often for “low risk” products and 2)viewed in a superficial way. These findings suggest thatproviding consumers with source citation information isnot helpful in certain contexts.

The implications of these studies are also important forconsumer advocates, magazine publishers and advertis-ers. All advertising claims require substantiation, butpresently advertisers are not legally required to printsource citation within the body of the advertisement.Some publishers are beginning to mandate the practice,which requires extra work, changes to formatting, influ-ences the type of claims advertisers use in print adver-tisements and may cause inconsistencies across adver-tising campaigns. These well-intentioned publishinghouse policies aim to help consumers distinguishbetween deceptive and non-deceptive claims. However,as these studies show, how consumers use source cita-tions to make judgments of a print advertisement claimis not as straightforward as may have been assumedand that the information does not always lead to moreaccurate judgments.

For consumers, different processing systems lead todifferential assessments of deception perceptions. Eventhough all advertisers in these magazines are required tocomply with the new regulations, advertisers of high riskproducts differentially benefit. For practitioners, in somecases, such as print advertisements for high risk products,citing the source for a claim is beneficial and aids con-sumers in making a better judgment. However, for lowrisk products consumers are not using the information toinfluence decisions regarding deceptive intent. Publish-ers should search for alternative methods to aid consum-ers in better judgments about deceptive intent in printadvertisements for all products, not just those perceivedas high risk. Magazine publishers should take these


findings of differential impacts into account when creat-ing and reviewing internal policies.

This subject is also relevant to policy makers. It iscommendable that print advertising publishers areattempting to self-regulate advertisements in order to aidconsumers in making better decisions. However, giventhe differential benefits dependent on risk, this policy isnot uniformly beneficial and deserves more attention andresearch. Policy makers should seek to find a way tocommunicate information in advertisements that trulyaids consumers in making the best decision possible,regardless of the level of perceived financial or physicalrisk.

LIMITATIONS AND FUTURE DIRECTIONS

There are limitations to these studies, some of which areaddressed in planned future studies.

One important avenue to explore is whether the sourcecitations are used as a simple heuristic. In our studies, allcitation information simply provided support for thecredibility of the claim. However, often the fine print inads contains warning information or qualifies the claim.Other ads use fine print with information that is seem-ingly neutral such as “when used as directed.” It isimportant to see if the observed relationship extends to allfine print information or if the content of the fine print(positive, negative or neutral) will have differentialeffects.

The placement and structure of the source citations couldalso be viewed as a limitation. Although many advertise-ments use an asterisk following the claim with the sourceinformation in fine print at the bottom, the possibilityexists that the fine print itself is what is perceived asdeceptive, rather than the actual citation information.Advertisers are able to (and often do) cite source informa-tion within the text of the claim. It would be interesting todetermine whether the font and placement or the actualinformation is what is driving perceptions. We intend toconduct future research exploring how different presen-tation styles and type of information presented affects anindividual’s deception perception.

Although the research agenda is still in progress, it doesextend the current theoretical understanding of consumerdeception perception. It is not a simple story where thepresence or absence of a source citation is what consum-ers use to guide deception perceptions. If it were thatsimple, magazine publishers would be absolutely correctin requiring advertisers to put the support informationbehind a claim statistic in an advertisement. However,given that consumers use that information differentlybased on the perceived risk level indicates the publishers’requirement could be detrimental to marketers and doesnot always aid consumers in making an optimal decision.Our studies indicate more research is needed to fullyunderstand the mechanisms consumers use to processand judge product claims in print advertisements andalso how those mechanisms differ across different riskdomains.

REFERENCES

Bargh, John A. and Tanya L. Chartrand (1999), “TheUnbearable Automaticity of Being,” American Psy-chologist, 54 (7), 462–79.

____________ (2002), “Losing Consciousness: Auto-matic Influences on Consumer Judgment, Behaviorand Motivation,” Journal of Consumer Research, 29(2), 280–85.

Baumeister, Roy F., Erin A. Sparks, Tyler F. Stillman,and Kathleen D. Vohs (2008), “Free Will in Con-sumer Behavior: Self-Control, Ego Depletion, andChoice,” Journal of Consumer Psychology, 18, 4–13.

Bettman, James R. (1973), “Perceived Risk and Its Com-ponents: A Model and Empirical Test,” Journal ofMarketing Research, 184–90.

Boush, David, Marian Friestad, and Peter Wright (2009),Deception In The Marketplace. New York: Psychol-ogy Press.

Bureau of Consumer Protection, The Federal TradeCommission. Retrieved November 2, 2011, fromhttp://business.ftc.gov/documents/bus35-advertis-ing-faqs-guide-small-business.

Campbell, Margaret C. (1995), “When Attention-Get-ting Advertising Tactics Elicit Consumer Inferencesof Manipulative Intent: The Importance of Balanc-ing Benefits and Investments,” Journal of ConsumerPsychology, 4 (3), 225–54.

____________ and Ronald C. Goodstein (2001), “TheModerating Effect of Perceived Risk on Consumers’Evaluations of Product Incongruity: Preference forthe Norm,” Journal of Consumer Research, 28 (3),439–49.


Cotte, June, Robin A. Coulter, and Melissa Moore (2005),“Enhancing or Disrupting Guilt: The Role of AdCredibility and Perceived Manipulative Intent, Jour-nal of Business Research, 58, 3 61–68.

Cox, Anthony D., Dena Cox, and Susan P. Mantel (2010),“Consumer Response to Drug Risk Information:The Role of Positive Affect,” Journal of Marketing,74 (4), 31–44.

Darke, Peter R. and Robin J. B. Ritchie (2007), “TheDefensive Consumer: Advertising Deception,Defensive Processing, and Distrust,” Journal ofMarketing Research, 44 (1), 114–27.

Dowling, Gramhame R. (1986), “Perceived Risk: TheConcept and its Measurement,” Psychology & Mar-keting, 3,193–210.

Evans, Jonathon (2003), “In Two Minds: Dual-ProcessAccounts of Reasoning,” Trends in Cognitive Sci-ences, 7 (10), 454–59.

____________ (2008), “Dual-Processing Accounts ofReasoning, Judgment and Social Cognition,” AnnualReview of Psychology, (59), 255–78.

Friestad, Marian and Peter Wright (1994), “The PersuasionKnowledge Model: How People Cope withPersuasion Attempts,” Journal of ConsumerResearch, 21 (1), 1–31.

Goldberg, Marvin E. and Jon Hartwick (1990), “TheEffects of Advertiser Reputation and Extremity of

Advertising Claim on Advertising Effectiveness,”Journal of Consumer Research, 17 (2), 172–9.

Hastie, Reid K. and Robyn M. Dawes (2010), RationalChoice in an Uncertain World: The Psychology ofJudgment and Decision Making. Second Edition.Thousand Oaks, CA: SAGE Publications.

Kaplan, Leon B., George J. Szybillo, and Jacob Jacoby(1974), “Components of Perceived Risk in ProductPurchase: A Cross-Validation,” Journal of AppliedPsychology, 59 (3), 287–91.

Kardes, Frank R., Steven S. Posavac, and Maria L.Cronley (2004), “Consumer Inference: A Review ofProcesses, Bases and Judgment Contexts,” Journalof Consumer Psychology, 14 (3), 230–56.

Monroe, Kent B. (1979), Pricing: Making ProfitableDecisions. New York City: McGraw-Hill.

Petty, Richard E. and John T. Caccioppo (1986), Com-munication and Persuasion: Central and PeripheralRoutes to Attitude Change. New York: Springer-Verlag.

Ross, Ivan (1975), “Perceived Risk and ConsumerBehavior: A Critical Review,” Advances in ConsumerResearch, 2, 1–20.

Shimp, Terence A. (1979), “Social-Psychological(Mis)Representations in Television Advertising,”Journal of Consumer Affairs, 13, 28–40.

APPENDIX


For further information contact:Leslie Koppenhafer

Lundquist College of BusinessUniversity of Oregon

1208 University of OregonEugene, OR 97403

Phone: 910.620.9922E-Mail: [email protected]


THE EFFECT OF HEALTH RISK PRESENTATION FORMAT ONCONSUMER PERCEPTIONS AND CHOICE

Jesse R. Catlin, Washington State University, Tri-CitiesElizabeth Gelfand Miller, University of Massachusetts, Amherst

Ann M. Mirabito, Baylor University

SUMMARY

Individuals are increasingly being asked to take a larger,more proactive role in their own healthcare (Laine andDavidoff 1996). However, there are numerous barriersthat consumers face when navigating this change, includ-ing the processing and understanding of complex riskinformation (Scammon et al. 2011). For example, anindividual may need to decide between two drug regi-mens that offer different likelihoods of success anddifferent probabilities of experiencing side effects. Tomake these decisions effectively, consumers must be ableto understand and appropriately use risk information.However, communicating complex numeric informationis not an easy task on its own, not to mention the fact thatmany consumers have low numeracy skills and/or areuncomfortable with numeric expressions of risk (Anckeret al. 2006). Thus, there is a pressing need to identifyeffective ways to communicate risk information.

However, while a large literature examining the presen-tation of risk information exists, there is little consensusregarding which formats are most effective. Risk can beconveyed in a number of ways that provide incidenceand/or base rate information. As such, numerically, num-bers can be expressed as frequencies (e.g., 1 in 5 or 2 in10) or percentages (e.g., 20%). Risk information isalso frequently displayed graphically. Graphical for-mats include icon arrays (which depict risk using indi-vidual icons, such as dots or stick figures), bar graphs,stacked bar graphs (which use two bars on top of eachother to depict both the risk incidence and non-inci-dence), pie charts, and line graphs, among others. Nu-merous studies examining the influence of various pre-sentation formats exist, but results are often mixed and itremains difficult to determine which formats are mosteffective (e.g., see Lipkus 2007). Further complicatingthis assessment is that researchers have studied numerousdependent variables and results often differ depending onmetric (e.g., verbatim knowledge, behavioral change,appeal).

To help synthesize these findings and provide a clearerset of recommendations for health communication, wepropose and test a more generalized theoretical frame-work for organizing and understanding prior resultswithin the literature. In doing so, we focus on threecritical goals relevant for consumers participating in theirown healthcare: increasing factual knowledge accuracy,increasing understanding, and influencing behavior. Theconstruct of factual knowledge can be connected toprevious literature referring to “verbatim knowledge,” orthe ability to specifically report the exact numeric infor-mation presented. Understanding can be related to aseparate construct, “gist knowledge,” which refers to theoverall impression or meaning taken away by the individ-ual based on the information presented (cf., Hawley et al.2008).

To test this framework, we conducted a 2 (base rateformat: graphical, numeric) x 2 (risk incidence format:graphical, numeric) between-subjects experiment tosystematically examine how variations in graphicalvs. numeric representations of incidence and base rateinformation will impact outcomes related to accuracyand comprehension, gist knowledge, and behavioralintention. Given task consistency theories, we expect thatnumeric information will lead to increased verbatimknowledge. Based on findings related to informationsalience, we expect that graphical formats which makecomparisons more salient will increase gist knowledge.Additionally, we propose that, as the first step in process-ing risk information, verbatim knowledge will drive thehigher level gist knowledge, while gist knowledge willexert more influence on behavioral intentions.

Consistent with these hypotheses, we find that numericformats resulted in greater accuracy than graphical for-mats as measured by verbatim knowledge (i.e., the abilityto directly report numbers from the provided informa-tion). Contrary to expectations, we also find that numericformats outperformed graphical formats with respect togist knowledge. Finally, we find that gist knowledge


drives treatment choices, but that verbatim knowledgemediates the effect of format on gist knowledge. As awhole, these findings suggest that risk incidence infor-mation should be provided numerically to increase accu-racy. However, our findings also provide evidence thatgraphical elements can increase salience of certain infor-mation. Presenting risk incidence information graphical-ly and base rate information numerically led to the lowestaccuracy when the accuracy question required partici-pants to access base rate information in addition to riskincidence information. This result could have occurred ifthe more salient graphical information increased thelikelihood participants would overlook the base rateinformation. Conversely, providing both types of infor-mation numerically, so that they were equally salient, ledto increased accuracy in some conditions.

Surprisingly, we did not find that graphical formats led toimproved gist knowledge over numeric formats. Thisconclusion requires further investigation, however, asthe gist questions followed the verbatim questions in oursurvey. Consequently, gist knowledge may have beenimproved for all participants because they had to firstconsider the actual numbers. However, this procedurealso suggests that one way to improve gist knowledge(comprehension) is to ensure that participants are able toread and understand the actual numbers presented. More-over, although display format affected verbatim knowl-edge, it did not have a direct effect on choice. The lack ofdirect effect should be investigated further. Finally, ourfindings suggest that choice is driven by gist knowledge,which is influenced by verbatim knowledge. Referencesare available upon request.

For further information contact:Ann M. Mirabito

Hankamer School of BusinessBaylor University

Waco, TX 76798–8007Phone: 254.710.3914



ECO-LABELING: APPROACH OR AVOIDANCE?

M. Paula Fitzgerald, West Virginia University, MorgantownJoshua D. Dorsey, West Virginia University, Morgantown

SUMMARY

Literature

Researchers studying conservation-oriented labeling havegenerally framed eco-labels in a positive light. As aresult, it is predicted that these labels will evokeapproach behaviors by consumers. Consumers are saidto approach, or move toward, products and services thatfacilitate the positive achievement of their goals and tomove away from, or avoid, those that inhibit goals and,thus, result in punishment. The underlying rationale ofapplying eco-labels to products is that they can “guideconsumers towards a more environmentally friendly pur-chasing behavior,” (Heinzle and Wustenhagen 2011,p. 60). Eco-labels, therefore, are designed to create ap-proach behavior, indeed the dominant theoretical fram-ing of eco-labels is that such labels encourage consumersto pay a premium for products carrying a credible label(Gulbrandsen 2006). In this paper, we explore the possi-bility that, for at least some consumers and some eco-labels, the eco-label may “stigmatize” a product. Workby Teisle, Rubin and Noblet (2007), alludes to the possi-bility that some labels have the ability to engender avoid-ance behavior. They theorize that aspects of the label,such as the credibility of its source, as well as aspects ofthe consumer, including his/her perceptions regardingthe product quality compromises required for positiveenvironmental effects, affect consumer approach/avoid-ance of a product. We posit that, under certain circum-stances, an eco-label may actually act as a stigma – aphysical mark on a product that identifies the product asundesirable or flawed, leading to unwarranted evalua-tions of the stigmatized product (Ellen and Bone 2008;Kasperson, Jhaveri, and Kasperson 2001).

Model/Results

We present our model using a relatively new eco-label,the US EPA WaterSense label, as the focal point of ourdiscussion. EPA partners (plumbing manufacturers and

retailers) offer the WaterSense label on plumbing fix-tures that are third-party certified to be at least 20 percentmore water efficient than the average product in thecategory and to “perform as well as or better than theircounterparts” (EPA 2012). We study two constructshypothesized to create approach behaviors: value on theenvironment and consumer’s belief that she/he workshard to conserve water. These constructs reveal goalsconsistent with conserving natural resources and, there-fore, are likely increase intentions to use the WaterSenseeco-label. However, skepticism, negative product at-tributes, perceptions that the source is untrustworthy, andambiguity in label meaning should, based on stigmatheory, result in avoidance behaviors. Data from 567homeowners living in one of four water-stressed U.S.cities (Atlanta, GA, Austin, TX, Baltimore, MD, andMinneapolis, MN) was used. Respondents viewed one oftwo different WaterSense labels – one of which is thecurrently-used version and a second, re-designed versionwhich replaced the words, “Meets EPA Criteria” with“Quality Performance • Conserves Water.” Then con-sumers completed an open-ended question, “If you sawthe label shown above on a plumbing fixture (toilet,faucet or showerhead), what would it mean to you?”Respondents also provided their intentions to look for thelabel the next time they shop for plumbing fixtures.Respondents provided information about their environ-mental values, water conservation effort, skepticism,product attitudes, and source untrustworthiness. Regres-sion was used to test the hypotheses.

Importantly, our results show that consumers tend toavoid eco-labels when they have negative attitudestoward the environmentally-oriented models of the prod-uct, when they do not trust the source, and when the labelis ambiguous. Our results suggest at least some individ-ual and label factors have the ability to reverse thisrelationship. In conclusion, although prior research tendsto use the underlying assumption that eco-labeling resultsin approach behaviors, we find that eco-labels can alsolead to avoidance. References are available upon request.


NOTE

This publication, Eco-Labeling: Approach or Avoid-ance, was developed under Cooperative Agreement No.X7-83410001 awarded by the U.S. Environmental Pro-tection Agency. It has not been formally reviewed by

EPA. The views expressed in this document are solelythose of M. Paula Fitzgerald and Joshua D. Dorsey andEPA does not endorse any products or commercial ser-vices mentioned in this publication. The authors thankUS EPA for funding this study.

For further information contact:M. Paula Fitzgerald

West Virginia UniversityP.O. Box 6025

Morgantown, WV 26505Phone: 304.293.7959



SCALE AND METRIC DESIGN AS CHOICEARCHITECTURE TOOLS

Adrian R. Camilleri, Duke University, DurhamRichard P. Larrick, Duke University, Durham

SUMMARY

Choices are never presented in a vacuum: rather, alterna-tive choice options are always presented within a con-text. Importantly, in many cases the context – the “choicearchitecture” – reflects the meta-decisions made by thedesigner of the context – the “choice architect” (Johnsonet al. 2012; Thaler and Sunstein 2008). A key principleassociated with the choice architecture metaphor is thatthere is no neutral choice context and therefore thoseresponsible for framing decisions will influence choicesin all cases. Motivated by this belief, scientists andpolicy-makers have become increasingly interested inlearning how to best arrange the choice architecture inorder to help mitigate the threat of various social dilem-mas, most recently the threat of climate change (Newelland Pitman 2010; Weber and Stern 2011).

Energy labels, and in many cases the standards thataccompany the labels, are often considered to be the bestavailable tools for governments to manage energy-efficiency policies and climate-change-mitigation pro-grams (Gillingham, Newell, and Palmer 2009; Sternet al. 1987; Wiel and McMahon 2003). Given that theconsumption of fossil fuels is a major contributor ofgreenhouse gas emissions, one natural target of choicearchitecture intervention are vehicle choices via fueleconomy label design.

One important feature of the fuel economy label is thechoice of metric or metrics to include. Most existinglabels report some metric associated with fuel efficiencyor consumption; however, there is also an option ofproviding a cost estimate. There is reason to expect thatproviding estimated fuel costs will influence consumerpreferences more than fuel consumption informationeven though these two metrics are simple translations ofone another. Payne (1982), for example, has arguedstrongly for the case that decision-makers often formpreferences through task-contingent strategies that aretied to the representation of a problem (Payne, Bettman,and Johnson 1992). Similarly, a number of researchershave proposed that task, strategy, and information sys-tematically interact to produce “compatibility” effects

(Fischer and Hawkins 1993). Given the assumption thatconsumers have a primary motivation to minimize thetotal costs of purchasing and running their vehicle, andtend to underweight the latter, we hypothesized that thosepresented with a gas cost metric would make a greaterproportion of fuel-efficient choices than those presentedwith a gas consumption metric.

A second important feature of the fuel economy label isthe scale to express metrics upon. Most existing labelsreport fuel economy metrics upon a “per 100” distancescale. Interestingly, there is a growing body of research incognitive psychology and marketing showing that rescal-ing otherwise identical information can systematicallychange preferences (Burson, Larrick, and Lynch 2009;Gourville 1998; Pandelaere, Briers, and Lembregts 2011).The consistent finding in this literature is that decision-makers tend to perceive differences as larger when theyare expressed on an expanded scale (such as per year) thanwhen they are expressed on a contracted scale (such as perweek). Larger differences in turn prompt greater reli-ance on that dimension in choice, thereby increasingpreference for the option favored on that dimension.Utilizing typical driving behavior information gath-ered during pilot work, we hypothesized that a greaterproportion of fuel efficient choices would be made bydecision-makers presented with a lifetime “per 100,000miles” scale than those presented with either an annual“per 15,000 miles” scale or the routine “per 100 miles”scale.

In order to test these two hypotheses, we conducted anonline experiment where we presented participants with anumber of hypothetical binary choices between two vehi-cles that traded off on price and fuel economy. We alsomeasured individual difference variables relevant to carpurchase decisions: environmental attitude, discount rate,and cognitive reflection. Stronger preferences for themore fuel-efficient vehicle were expected for those whoexpected to drive more, who held pro-environment atti-tudes, and exhibited lower discount rates (in this case, ofthe later savings associated with the fuel efficient option).


With regards to metric, we observed that decision-makers tended to select the more fuel-efficient vehicleswhen fuel economy was expressed in terms of the esti-mated cost of gas rather than the amount of gas con-sumed. This metric effect is compatible with contin-gency explanations of information processing, such asthe theory of cognitive fit (Vessey 1991; 2006) and thescale-compatibility hypothesis (Tversky et al. 1988).In general, metrics that match the problem-solving pro-cesses required to form a preference have the greatestinfluence on choice.

With respect to scale, we observed that decision-maker’stended to select the more fuel-efficient vehicles when thefuel economy metric was expressed on an expanded,lifetime scale. Interestingly, we observed a “U” shapedfunction in which the proportion of fuel efficient optionswas smallest at 15,000 miles and largest at 100,000miles; 100 miles was somewhere in between. Thisresult partially replicates past research on scale-induced preference reversals changes (Burson et al.

2009; Gourville 1998; Pandelaere et al. 2011). The “U”shaped function was partially mediated by a similar U-shaped function for expected driving behavior: thosewho were presented with the 15,000 label during theexperiment subsequently expected to drive less in thefuture than those presented with either the 100 or 100,000scale.

As expected, those who held pro-environmental atti-tudes were more likely to select more fuel-efficientvehicle options. However, in contrast to our expecta-tions, implied discount rate did not appear to moderatechoices.

The important message from our work is that policy-makers should provide consumers with meaningful met-rics that help consumers assess their goal progression.Many consumers would appear to do better in this regardif provided with estimate fuel cost information expressedupon an expanded, lifetime scale. References are avail-able upon request.

For further information contact:Adrian R. Camilleri

The Fuqua School of BusinessDuke University100 Fuqua Drive

Durham, NC 27708E-Mail: [email protected]


THE ROLE OF CENTRALITY OF GREEN ATTRIBUTES INGREEN PRODUCT EVALUATION

Andrew D. Gershoff, University of Texas at AustinJudy K. Frels, Audencia Nantes School of Management, France

SUMMARY

Concern about the limited natural resources of the planetand an interest in careful stewardship of those resourcescontinue to influence the decisions of both firms andconsumers. Firms are building environmentally friendlysupply chains, including recycled material in their prod-ucts, and developing manufacturing methods that avoidpolluting their surrounding environments.

Consumers’ reactions to firms’ investments have beenmixed. While some consumers espouse enthusiasm formore environmentally-friendly offerings, a recent Gal-lup poll suggests that many stop short of actually pur-chasing such products (Dunlap 2010). Customers havebeen found to suffer from “green fatigue” (Platt andRetallack 2009), to be suspicious of “green washing”(www.greenwashingindex.com), and to be unwilling topay extra for environmentally-conscious products (Bhat-tacharya and Sen 2004).

However, a more fundamental question underlies thesereactions: What makes a consumer perceive a product asgreen or not green to start with? Before a product’senvironmental friendliness can have any influence on aconsumer’s purchase decision, the consumer must seethe product as green – or not.

Consider a consumer looking at two brands of papertowels in a supermarket. The information provided indi-cates that the environmental savings achieved by includ-ing recycled material in both brands is exactly the same,but one brand includes recycled materials in the packag-ing of the towels, while one includes recycled materialsin the towels themselves. Do consumers perceive thesetwo products as equally green? Or consider two laptopcomputers, one with an environmentally friendly moth-erboard and the other with an environmental friendlypower cord. The use of recycled materials in the twoproducts saves the same amount of plastic material fromentering landfills, but do consumers perceive a greenperipheral element (the power cord) as contributing as

much “greenness” to a product as a green component thatis more essential to the product (the motherboard)?

In this research, we look at factors that influence aconsumer’s perception of the environmental friendlinessor greenness of a product. We explore this from theperspective of centrality theory (Sloman, Love, and Ahn1998) which states that certain features of a product aremore central – more essential, more integral – to definingwhat a product (or natural object) is and to what categoryit belongs. One facet of centrality involves high- versuslow-level construals (Trope, Liberman and Wakslak 2007)and so we examine the role construal plays in the consum-er’s evaluation process and the consumer’s categoriza-tion of a product as green or not green. We also look athow different elements of a product are presented to theconsumer (holistically or separately) and also examinewhere in the supply chain the “greenness” of a productoccurs. We test the effect of these variables across mul-tiple experiments that span four different product catego-ries – mattresses, paper towels, laptop computers, andwashing machines. The results of our four studies suggestthat products with identical reductions in their environ-mental impact may have their environmental friendlinessevaluated very differently by consumers depending onthe centrality or non-centrality of the product featuresthat are green. We find that differences in green percep-tions between products with central or peripheral greenattributes are attenuated when a high-level or distantconstrual is triggered.

The implications of this for public policy makers andfirms are potentially significant. From a public policyperspective, our results suggest that consumers’ evalua-tion process of green claims may be influenced by factorswhich are not related to green outcomes. In our study, thegreen claims were very specific and clearly identical, yetthe “peripheralness” or centrality of the green feature hada significant effect on the consumers’ assessment ofgreenness. This suggests that consumers’ ability to assessenvironmental impact is influenced by factors not relatedto environment outcomes. From the perspective of firms,


they may wish to consider carefully their green invest-ments and their communication of these efforts. If thesame environmental impact can be achieved throughboth a central feature and a non-central feature, ourfindings suggest that for positive consumer evaluations,

the investment is best made in the central feature. If thatis not possible, then encouraging a more abstract con-strual may alter the consumer’s processing and achievepositive results from an investment in a peripheral fea-ture. References are available upon request.

For further information contact:Judy K. Frels

Audencia Nantes School of Management8 route de la JoneliereNantes, France 44312

Phone: +1.202.621.0028E-Mail: [email protected]


AN EXPLORATORY STUDY OF BEHAVIOR PLACEMENTS INTELEVISION SITCOMS

Christine M. Kowalczyk, East Carolina University, GreenvilleJennifer Martinez, The University of Memphis

SUMMARY

Drive a hybrid vehicle. Recycle plastic beverage bottlesin the workplace. Change light bulbs to CFLs. Theseenvironmental behaviors have been featured on NBCtelevision programs via a new concept called a behaviorplacement, which was developed “to sway viewers toadopt actions they see modeled in their favorite shows”(Chozick 2010). The subtle messaging brings main-stream concerns, like environmental issues, to the public.NBC has realized the value in this messaging since 75percent of consumers consider themselves to be “envi-ronmentalists” (Saad 2006). However, 2009 estimatesshowed 160 million consumers have yet to purchase agreen product (Stengel 2009). A general concern for theenvironment does not translate into green purchases(Kaiser et al 1999), thus indicating an attitude-behaviorgap in consumers’ green behaviors. This study attemptsto address this gap by evaluating whether a behavioralmessage embedded in the storyline of a TV show caninfluence consumers’ behavior intentions.

Product placements (PPL), “the practice of placingbranded products in the content of mass media program-ming” (Russell 2002), have shown to influence behavior.The concept of PPLs has evolved into behavior place-ments, designed to influence viewers to adopt specificactions they see in their favorite shows (Chozick 2010).While NBC tends to focus on non-offensive or non-controversial messages in its behavior placements, thequestion remains how viewers respond to these types ofmessages in their programs. Viewers watch television tobe entertained, or escape from reality. Putting “lessons”into a program may have the opposite effect and alienateviewers from the program or network, but they mayprovide a unique communication opportunity for publicpolicymakers to influence pro-social behaviors. Thus,the following hypotheses are advanced:

H1: Attitudes toward the environmental behavior place-ment are positively related to environmental behav-ior intentions.

H2: Perceived realism of the environmental behaviorplacement in the television show has a positive

relationship with (a) attitudes toward the environ-mental behavior placement and (b) environmentalbehavior intentions.

H3: Skepticism of the environment behavior placementhas a direct but negative relationship with attitudestoward the environment behavior placement.

H4: Environmental consciousness has a direct relation-ship with (a) attitudes toward the environmentalbehavior placement and (b) environmental behaviorintentions.

To test these hypotheses, an environmental behaviorplacement was identified from the 2009-2010 season ofNBC’s The Office. In the episode, one of the maincharacters dressed up as a caped superhero who isobsessed with recycling and related different ways torecycle in the context of the show’s storyline. The TVsegment containing the behavior placement was edited toa reasonable length for research purposes. Respondentswere recruited through an online panel service with afinal sample of 248 usable responses. The questionnaireincluded a series of measures to assess perceptions ofenvironmental consciousness, perceived realism, skepti-cism, attitudes toward the behavior placements and be-havior intentions. Reliability and validity of the scalemeasures adequately fit the measurement model in M-Plus. The structural model found full support for the firstthree proposed hypotheses and only partial support forhypothesis four.

To our knowledge, this study is first of its kind tounderstand the growing use of environmental behaviorplacements in TV programs. The study found attitudestoward the environmental behavior placement had apositive effect on behavioral intentions. If positive feel-ings toward an environmental behavior featured in a TVprogram can be garnered from the exposure to just thebehavior, maybe more consumers will be inspired toparticipate in pro-environmental activities like recycling,taking public transportation and buying locally grown.These messages may be more influential than publicservice announcements from the government or advo-cacy groups.


In addition, perceived realism of the environmentalbehavior placement in The Office had a positive effect onattitudes toward the specific environmental behaviorplacement segment and environmental behavior inten-tions, suggesting that perceived realism may be a criticalelement in strengthening attitudes and intentions for thebehavior placement in a scripted TV show. While theenvironmental behavior does not feature a product orbrand, this study supports similar relationships in the PPLliterature which state that products in movies and televi-sion add a realistic element and produce positive attitudesand intentions. It appears consumers may have beenskeptical about the environmental behavior placement,which resulted in weakened attitudes toward the place-ment, supporting the literature on skepticism and adver-tising. Finally, it appears that individuals’ feelings aboutthe environment do not translate to positive attitudestoward an environmental behavior in their TV program;however, their feelings about the environment do influ-ence their behavioral intentions toward the environment.

As PPLs become more prevalent in TV shows, theireffectiveness may be questioned by consumers and pub-lic policymakers. Building a message (i.e., encouraging

recycling) into a storyline through a behavior placement,however, is less obtrusive and may have a greater influ-ence on behaviors. This exploratory study sought tobetter understand attitudes and perceptions toward abehavior placement in a specific TV program, an area thathas received little attention in the academic literature.Also, the study expanded the concept of PPL to includebehavior placements, a unique form of PPL now fea-tured on NBC. This study is one of the first to examinethis new approach and provide theoretical and mana-gerial implications.

Featuring “good” messages is not a new concept, butNBC has taken an active role in packaging and marketingthe concept not only to its viewers but also potentialadvertisers. As PPLs become more pervasive and jarringin TV programming with potential regulations by theFederal Communications Commission, behavior place-ments provide a unique opportunity for persuasive envi-ronmental and pro-social messages to be embedded intoTV shows. This will allow marketers, advertisers andeven public policymakers to better target their pro-socialcommunication messages. References are available uponrequest.

For further information contact:Christine Kowalczyk

East Carolina University3125 Bate Building

Greenville, NC 27858Phone: 252.737.4572



AN EMPIRICAL STUDY OF THE DETERMINANTS OF GREENSHOPPING BEHAVIOR IN BRAZIL

Helenita R. da Silva Tamashiro, Faculdade Reges de Ribeirão PretoJosé Augusto Giesbrecht da Silveira, FEA-USP

Claudia Rosa Acevedo, FGV-SP

SUMMARY

Introduction

As explained by Firat (2009), pressures from environ-mental organizations, competition between firms andacademic discussions are factors that have contributed tothe emergence of environmental awareness from 1980,especially due to climate change, droughts and otherevents resulting from global warming. It is from thisperspective that this study investigates the influence ofEcological Knowledge (EK), Ecological Concern (EC)and Ecological Affection (EA) on Ecological CosmeticsShopping Behavior (ECSB) in Brazil.

Understanding the consumers of green products (andtheir characteristics) has been the keynote of market-ing studies (Goncalves-Dias et al. 2009). Further,research studies that seek to investigate the relationshipbetween environmental factors and consumer behav-iors are still embryonic in Brazil. Therefore, this studyprovides a unique opportunity to broaden the discussionrelated to the shopping behavior of consumers withregard to environmental issues. Thus, the aim of thisresearch was to investigate the influence between Eco-logical Knowledge (EK), Ecological Concern (EC) andEcological Affection (EA) on the consumer’s shoppingbehavior.

Methodological Design of the Empirical Research

In order to empirically investigate the proposed model,research was conducted with consumers in Brazil.Women were defined as potential respondents of thestudy, because they represent the major consumer groupin terms of cosmetics (ABHPPC 2012). Thus, femaleconsumers were surveyed, aged 18 and over, living in theurban area of Ribeirão Preto-SP. The choice of this agegroup was because accurate data on the female popula-tion of other age groups could not be found. Altogether800 questionnaires were distributed. Of this total, 500questionnaires were considered in the research, whichcorresponds to a rate of 62.5 percent. In the data analysis,

we used Cronbach’s alpha coefficients and the multivari-ate technique of exploratory factor analysis for datacleansing. The dimensions observed in the theory on theresearched subject demonstrate theoretical support forthe empirical results found in this study, showingsimilarities with the theoretical input, given the corre-lation between the independent variables and depen-dent variable.

Discussion of the Results and Closing Remarks

By the end of this research, it was possible to note that theproposed scope had been achieved. As to the relationshipbetween EK and ECSB, although Tanner and Kast (2003)and Kollmuss and Agyeman (2002) suggest that empir-ical tests have not agreed on the direct relationshipbetween environmental knowledge and pro-environmen-tal behavior, the results found in this research contradictsuch premises, as they show the existence of a one-waycorrelation. It is believed that this correlation is asso-ciated with respondents’ high levels of knowledgeabout environmental issues: 7.39 (Factor 1) and 8.20(Factor 2). By analyzing the results concerning the corre-lation between EC and ECSB, it was noted that they donot eliminate the ambiguity of the dimensionality of ECremarked by Xiao and Dunlap (2007). If analyzed fromthe perspective of the arithmetic means of EC (7.20 and8.77 for EC1 and EC2, respectively), it can be said thatsuch results are very positive, because, on average, theECSB of the searched sample is not as much (6.70).

The results of the correlation between EA and ECSBwere in line with the premises of Dispoto (1977) andMaloney and Ward (1973), as explained in the theoreticalreferences. According to these authors, even with littleknowledge about the environment, people have a strongemotional connection with it. The congruencies hereinare also extended to the premises of Schouten (1991),which conceives emotional factors as one of the mostinteresting components of consumer behavior. Com-pared with the other independent variables, EC1 pre-sented the strongest correlation with the dependentvariable (r = 0.540), followed by EK1 (r = 0.301) andEA (r = 0.276). It should be noted, however, that


although there is an association between ECSB and EK,it is a tenuous relationship, meaning that people are awareof environmental problems, but present an ecologicalbehavior that does not match their levels of knowledge inthis respect.

An important contribution glimpsed in this study refers tothe references that can serve as an inspiration for futureresearch studies in the marketing area, more specifically,strategic planning, consumer behavior and marketingcompound management, among others. However, therewere some limitations that may be overcome in future

studies. First, although the indicators used were based onthe literature and duly adapted to the specific object ofthis study, the 11-point scales used, which ranged from 0to 10, were grounds for complaints by some respondentswho felt they were too extensive. Thus, some assump-tions may be proposed in future studies in order tobroaden academic knowledge and obtain differentresults. Second, future studies could establish an analysisthat also considers men as consumers, since that literatureindicates that the male audience is a potential cosmet-ics consumption group. References are available uponrequest.

For further information contact:Claudia Rosa AcevedoRua Duarte da Costa1170, Alto da Lapa

São Paulo SP, 05080–000Phone: 55.11.3834.2222



BRINGING IT BACK TO LIFE: REPURPOSING PRODUCTS TOMAKE THEM NEW AGAIN

S. Todd Weaver, Point University, West PointKristin Scott, Minnesota State University, Mankato

SUMMARY

A new consumption market is emerging in the U.S., onethat is generally labeled repurposing or upcycling.Although definitions vary, repurposing involves takingsomething that is no longer useful or needed and convert-ing it so that it is once again valuable to a consumer. Theemergence of this new market could hold significantbenefits for the environment. When something is recy-cled, it becomes something of lesser quality and con-sumes energy in the process (Wang 2011). In compari-son, repurposing is a closed-loop system where materialsare continuously kept in the production-consumptioncycle and maintain their status as resources instead ofbeing down-graded through recycling. Repurposing isalso more sustainable than purchasing environmentallyfriendly new products because all new products, whethersustainably manufactured or not, require additional nat-ural resources and energy production in the manufactur-ing process.

Despite the potential importance of the repurposing mar-ket, very little research has investigated consumer andbusiness behavior in this industry. The current researchexplores the motivations of business owners in the indus-try and the barriers and opportunities they face. A total ofeight in-depth interviews were conducted with busi-nesses that included industrial products, jewelry, artobjects, reclaimed lumber and repurposed furniture.Business owners were located nation-wide, includingstates such as California, Colorado, Minnesota and NewYork.

Results

The business owners interviewed revealed interestinginsights about the emerging repurposing industry, theirmotivations, and the motivations of their consumers.Three themes emerged. First, the repurposing market isfar from an established, well-defined industry. Rather, itis still emerging and fragmented, with diverse and con-tested definitions and many types of producers. Amongthe words and phrases used to describe their industryinclude: craft, recycled, repurposed, restyled, reused,reclaimed, recreated, rescued, used materials, and land-

fill divergence. These accounts indicate that the defini-tion of the market is still being contested and it appears atthis point to be a fragmented aggregation of narrowly-focused specialty markets.

The second theme to emerge was that both producers andconsumers in the market for repurposed items have amixture of self-interested and altruistic motives for par-ticipating in this market. In spite of the environmentalbenefits of repurposing, most informants did not reportidealistic motivations for starting their businesses. Rath-er, their work in the market for repurposed goods wasprimarily a way to earn a living and to achieve personalwell-being. At the same time, however, their motiveswere not purely selfish. Informants also reported motiva-tions to help others or to sustain the natural environment.

The final theme in this study concerns institutional barri-ers and facilitators of the emerging market for repurposeditems. According to informant accounts, several barriersand challenges are endemic to the repurposing market.First, when used or discarded goods serve as the inputs tothe business process, finding a reliable source of thesegoods can be challenging. Another challenge faced bybusinesses in this market is the lack of a successful,replicable business model in the industry. A third barrierfaced by businesses in the repurposing market is uncer-tainty over the market value of what they sell. Becausethere aren’t well-established markets for repurposedgoods, there often isn’t an accepted market price orcompetitor price to use as a reference. The final challengereported by informants concerns the operational diffi-culties presented by reprocessing used or discardedmaterials.

In spite of the disadvantages of operating a repurposingbusiness, the informants remained optimistic about thefuture of the industry as a result of three key opportunitiesin the market environment. First, repurposing businessescan capitalize on inputs (i.e., goods or materials to berepurposed) that are frequently cheap or even free. Sec-ond, the market takes advantage of current shifts inconsumer attitudes toward going green. A final facilitatorfor repurposing businesses is that the act of repurposing


and purchasing repurposed goods creates an experientialvalue for customers that new goods do not.

Discussion

The results of this study provide a number of importantinsights about repurposing and the market for repurposedgoods. One important, over-arching theme in the data isthat the repurposing industry is still emerging and uncer-tain. This early stage of development in the market isconsistent with research on institutional theory (Hum-phreys 2010; Navis and Glynn 2010; Scott 1995) andcollective action frames (Snow and Benford 2005). Thisresearch suggests that new markets and industries seeklegitimation via establishment in the cultural-cognitivevernacular of a society, consistency with societal norms,and adherence to public regulations (Scott 1995). Theinformants for this study placed the repurposing marketearly in all three forms of legitimation. In terms ofcultural-cognitive legitimacy, informants expressed alack of agreement about the definition of repurposing andhow it differs from similar concepts like recycling and

upcycling. Congruence with societal norms offers a sec-ond avenue for legitimation (Scott 1995), and the repur-posing industry has favorable prospects because it isalready in-line with several consumer and industry trends.The final legitimating factor is regulation; businesses andmarkets gain legitimacy by adhering to laws and regula-tions governing commerce. Currently there is no regula-tion or legislation specific to the repurposing industry,which presents a significant possible role for publicpolicy in promoting and facilitating markets for repur-posed goods.

The repurposing industry is an emerging consumer mar-ket with important potential benefits for consumers andsociety. Public policy could play an instrumental role inlegitimizing the industry and promoting the productionand consumption of repurposed goods. The establish-ment of robust repurposing markets would not onlyprovide economic and experiential value to producersand consumers, it would also promote a more environ-mentally-friendly form of consumption. References areavailable upon request.

For further information contact:S. Todd WeaverPoint University

507 West 10th StreetWest Point, GA 31833Phone: 706.385.1096



CROSSING THE RECYCLING BIN BOUNDARY: THE EFFECT OFPRODUCT DISTORTION AND SIZE IN CONSUMER

DISPOSAL DECISIONS

Remi Trudel, Boston UniversityJennifer Argo, University of Alberta, Edmonton

SUMMARY

Two billion tons of trash is thrown away each year, withthe United States throwing away more trash than anyother country (Wilcox 2008). Given the substantiveeffect recycling has our environment, a better under-standing of the consumer behavior behind recycling is animportant endeavor. Unfortunately, research in this areais scant at best with the majority of the working focusingon the implications of consumer-related characteristics(e.g., attitudes and personality) and promotional-relatedvariables (e.g., message framing) on recycling tenden-cies without considering the role of the product’sattributes themselves. In the present research we seek toprovide initial insight into the implications of variousproduct attributes on the likelihood that the product willbe recycled as compared to trashed. Our basic thesis isthat if the consumption process distorts the productsufficiently from its original form and function then itwill be deemed as less useful and in turn be more likelyto be thrown in the garbage (as opposed to recycled). Totest our thesis, we explore two product attributes that areoften affected during the consumption process – size andaesthetics – to determine their implications for recyclingdecisions.

To make sense of the numerous products that exist,consumers assign products to different classes or catego-ries. While research on categorization assumes that aproduct does not change its category once it has beenproduced (Malt, Ross, and Murphey 1995; Moreau,Markman, and Lehmann 2001), in reality products oftengo through a number of physical changes during and afterconsumption that “distort” the product from its originalform and function and may influence categorization. Webelieve that product distortion during the consumptionprocess does change consumers categorization of a prod-uct pre and post consumption. When product distortionalters a product’s characteristics substantially and dis-tances the changed product from the original, the new

product form is assigned to the disposal category withsimilar prototype attributes (Smith and Minda 1998).

In the present research we examine product distortionand the implications of changes in two attributes on thelikelihood a product will be garbaged (vs. recycled):size and aesthetics. As mentioned above, if a productis distorted during consumption, distance is createdbetween the original product and its new form andfunction. If the changes are in attributes that define adisposal category, then the product will be assigned tothe disposal category with the most similar prototypeattributes. We believe that size and aesthetics have asignificant impact on perceived usefulness of a productand that usefulness is a category defining attribute forrecyclables and trash. If the consumption process leads toan end product that is too small and/or in a product that isunwhole, damaged, or broken, it will no longer be per-ceived as typical to its category, which will decreaseits perceived usefulness. Because the product will beassigned to the disposal category with the most similarprototype attributes, we propose that products withlow perceived usefulness will be assigned to the gar-bage category and subsequently trashed (as opposed torecycled).

Employing a field and five experimental studies we setout to investigate the implications of product distor-tion on the likelihood a product will be recycled. In allof our studies, actual recycling behavior is our depen-dent variable.

Our initial field study involved searching through therecycling and garbage bins of twenty-two faculty assis-tant offices from a large northeastern US university. Thefield study found that the size of the product matters:participants were more likely to recycle a piece of paperthat was large than one that was small. Using a con-trolled laboratory environment, in study 2 we foundthat distorting a product by making it smaller substan-tially decreased recycling behaviors. Participants recy-


cled an undistorted product more often even when itsmass was exceeded by the mass of a distorted product’ssmaller parts. We also found that product distortioninfluenced categorization of the product and made itmore typical of garbage. In Study 3, we set out togeneralize the results found in the first studies by extend-ing our investigation to another product – aluminumcans. Consistent with what we found with paper, partic-ipants recycled the can more often when it was largeversus small, and even more so when it was not dam-aged. The results from Study 2 and 3 also show that itdoesn’t seem to matter whether the product is distortedby the consumer (Study 2) or by someone else (Study 3).Studies 4–6 demonstrated the graded nature of the gar-bage category and found support for the role of useful-ness in disposal decisions. Study 4 provided additionalsupport that it is not the act of distorting the productalone that makes it more like garbage, rather the newlydistorted product’s final form and function. As a demon-stration of the garbage category’s graded structure, par-ticipants only trashed the paper when it was sufficientlysmall enough in the end. We found that even after it hadbeen distorted, if the end product was large enough, itcontinued to be classified as recyclable and disposed of

accordingly. Studies 5 and 6 found direct evidence for therole of usefulness allowing us to conclude that size andaesthetics have a significant impact on recycling behav-iors due to changes in the perceived usefulness of theproduct. In particular, when product distortion leads to alarge product that is damaged, the decrease in the per-ceived usefulness of the product leads to it being per-ceived as more typical of garbage (vs. a recyclable) andsubsequently trashed. Finally, increasing the salience ofthe usefulness of a product can reverse consumers’ ten-dencies to trash a small object.

It us our hope that our research stimulates a discussion ofpost-consumption and recycling decision making thathas largely been ignored. Given the significant amount ofwaste that is generated in society, and the detrimentaleffect it has on the environment it is critical to understandconsumers’ recycling behaviors. This research seeks toprovide some insight into how certain product character-istics – size and aesthetics – influence the likelihood thatconsumers will or will not categorize a product as gar-bage and subsequently their decision to do their part forthe planet and recycle. References are available uponrequest.

For further information contact:Remi Trudel

School of ManagementBoston University

595 Commonwealth AvenueBoston, MA 02215

Phone: 617.358.3316Fax: 617.353.4098



WHEN ENVIRONMENTAL MESSAGES SHOULD BE ASSERTIVE:THE MODERATING EFFECT OF EFFORT INVESTMENT ON

CONSUMERS’ RECYCLING INTENTIONS

Tae Hyun Baek, Indiana University – Southeast, New AlbanySukki Yoon, Bryant University, Smithfield

Seeun Kim, University of Tennessee, Knoxville

SUMMARY

Considerable effort has gone into developing and imple-menting green campaigns to persuade people to act in anenvironmentally responsible manner. In particular, envi-ronmental campaigners routinely utilize assertive lan-guage that is characterized by increased use of impera-tives (i.e., commands and orders) rather than proposi-tions or indirect suggestions (Miller et al. 2007). Whilepast research has shown that assertive language yieldslower compliance with the advocated behavior withsome boundary conditions such as language intensity andexpectations (Bensley and Wu 1991; Dillard and Shen2005; Quick and Considine 2008; Quick and Stephenson2007; Wilson and Kunkel 2000), we propose that theeffort invested in receiving the request can moderate thepersuasiveness of message assertiveness, altering com-munication expectations for assertive language. The cur-rent work builds on prior findings that the effort investedin attaining the desirable outcome is associated with itsexpectations (e.g., Kim and Labroo 2011; van Dijk, vander Pligt, and Zeelenberg 1999). Thus, the main purposeof this study is to explore whether the effect of messageassertiveness varies as a function of the effort invested incomplying with environmentally-friendly requests.

In addition, this study highlights perceived issue impor-tance and self-efficacy as underlying explanatory factorsfor the interactive effects of message assertiveness andeffort investment on recycling intentions. Prior work hassuggested that perceived issue importance plays a medi-ating role in the process of compliance with assertivegreen messages (Kronrod, Grinstein, and Wathieu2012). It is clear that self-efficacy, defined as “beliefin one’s capabilities to mobilize the motivation, cogni-tive resources, and courses of action needed to meetgiven situational demands” (Wood and Bandura 1989,p. 408), involves cognitive judgments of personal abilityto comply with the advocated behavior (Bandura 1993).The extent to which people alter their perceptions of self-efficacy will vary as a function of how much effort they

put forth in given endeavors and the outcomes theyexpect their efforts to produce (Bandura, 1984). Self-efficacy has also been shown to translate into pro-envi-ronmental behaviors (e.g., Axelrod and Lehman 1993;Lindsay and Strathman 1997). It is expected that combin-ing message assertiveness with effort investment leads togreater perceived issue importance, which translates intogreater self-efficacy. Such a sequential pathway willresult in more positive recycling compliance.

We tested our hypotheses using a 2 (effort investment:high versus low) X 2 (message assertiveness: assertiveversus non-assertive) between-subjects design. A total of249 college students at a northeastern U.S. universityparticipated in exchange for extra credits. Following theinduction procedure used by Zhang et al. (2011, experi-ment 2), we manipulated the degree of effort investmentusing the content of the recycle pledge. On the otherhand, the assertive message contained imperatives andcontrolling terms (e.g., “Recycle what you can: You haveto recycle plastic containers, paper, cardboard, alumi-num and steel cans”), whereas the non-assertive messageincluded terms that emphasize the autonomy of actions(e.g., “It’s worth recycling what you can: You couldrecycle plastic containers, paper, cardboard, aluminumand steel cans”). Except for the message assertivenessmanipulations, all other aspects of the ad stimuli wereinvariant with respect to size, layout, and background.

The findings show that individuals with high-effortinvestment show higher recycling intentions whenexposed to an assertive message versus a non-assertivemessage. In contrast, a non-assertive message is moreeffective than an assertive message for those with low-effort investment. Accordingly, our results suggest thatenvironmental advertising persuasion is enhanced whenmore assertive message is in line with the extent to whichpeople put in more effort to comply with green requests.Furthermore, the interaction between message assertive-ness and effort investment is partially mediated by per-ceived issue importance to predict self-efficacy, which in


turn leads to behavioral intentions to recycle. From atheoretical standpoint, the current research takes theimportant conceptual step of untangling the underlyingmechanism through which the interaction between mes-sage assertiveness and effort investment influences com-

pliance with green requests. From a practical standpoint,understanding the combined impact of message asser-tiveness and effort investment helps green marketersfine-tune their environmental message strategies. Refer-ences are available upon request.

For further information contact:Tae Hyun Baek

Department of Communication StudiesIndiana University – Southeast

4201 Grant Line RoadNew Albany, IN 47150–6405

Phone: 812.941.2031Fax: 812.941.2529



ETHICS AUDIT REVEALS “LITTLE DUTCH BOY” POLICYLEAKS: CREDIT CARD ACT (2009) FAILS

COLLEGE STUDENTS

Connie R. Bateman, University of North Dakota, Grand ForksMary K. Askim-Lovseth, University of North Dakota, Grand Forks

SUMMARY

Regulation debates surround the efficacy of Title III ofthe 2009 Credit Card Accountability and Reporting Dis-closure (CARD) Act which is intended to protect thewell-being of young adult consumers from predatorycredit card marketing practices on campus. The CreditCARD Act was put into effect in response to concerns byCongress and the President that certain practices in thecredit card industry were neither fair, nor transparentor accountable to consumers. Protection areas includeextension of credit, increases in credit lines, reportingrequirements, pre-screened credit offers, privacy protec-tion, and free gifts. Under the Act, college students(under the age of 21) are considered a protected classbecause not only may they lack financial literacy, but dueto their environment and decision making processes, maybe ill-equipped to manage credit. Current marketplaceeffects indicate the Credit CARD Act’s restrictions maynot have proven as successful as hoped for (e.g., easycredit; unmanageable debt loads; suffered grades; forceddropouts; privacy threats; a similar number of agree-ments between bank issuers and colleges; creative post-Act credit terms and strategies that protect corporatefinancial returns). Calls have been made for researchersand policymakers to find ways to inform future amend-ments to the Act.

Ethics Audit

We assert that key ethical principles (Sirgy, 2008) givevaluable insight into policy leak risks and provide afoundation for ethically informed amendments concern-ing key stakeholder groups. These principles are thefoundation of an ethics audit conducted on the currentrequirements of the Credit CARD Act.

Consumer Sovereignty Ethical Concerns

Applying Consumer Sovereignty to Title III SEC 301(Truth in Lending Act): Extension of Credit to CollegeStudents reveals ethical concerns relating to dishonest

reporting of financial information (irrational consumer)and lack of self-education (financial illiteracy). ApplyingConsumer Sovereignty to Title III SEC 302 (Fair CreditReporting Act): Protection of College Students fromPrescreened Credit Offers reveals ethical concernsrelating to student ignorance regarding the relation-ships between opting in to credit offers, multiple creditreport pulls, credit scores, and the long-term ramifica-tions of poor credit scores. It also reveals concerns aboutlack of self-education on the consumption process as itrelates to properly reacting to pre-screened credit offerssuch as how to discern whether a need for credit exists,search for unbiased and accurate information about cred-it (in general) and credit card offers, and compare andcontrast pre-screened credit offers; and understand whatthe relationship between credit report ‘pulls’ and creditscore is, and how credit scores work and why they areimportant for the student’s future.

Social Justice Ethical Concerns

Applying Social Justice to Title III SEC 302 (Fair CreditReporting Act): Protection of College Students fromPrescreened Credit Offers reveals ethical concerns relat-ing to social injustice for the segment of college studentswho are financially literate, self-educated and sovereign,and financially qualified to receive credit, but not offeredcredit because although they opt in to the offers, their dateof birth is not provided on the credit report, therefore thecredit card issuer may not follow-through with validationof age, causing an injustice to sovereign consumers underthe age of 21.

Non-Malfeasance Ethical Concerns

Applying Non-Malfeasance to Title III SEC 303 (Truthin Lending Act): Issuance of Credit Cards to CertainCollege Students (those with financially qualified co-signers) reveals ethical concerns relating to a co-signer whom may request an increase in the card’slimit without the cardholder’s knowledge and subse-quently, the co-signer may make charges on the ac-count, again without the cardholder’s knowledge and


harming the student. Applying Non-Malfeasance to TitleIII SEC 304 (Truth in Lending Act): Privacy Protectionfor College Students reveals an inducement-related con-cern relating to misplaced brand perception halo effectsresulting from free gifts received from the credit cardissuer; thus potentially harming the consumer psycho-logically by facilitating a perception of the card brandthat is based more on the “feel good” gift and based lesson an intellectual analysis of the card and its terms.Another inducement-related concern exists relating toconfusion regarding the loosely interpreted definition ofcampus ‘borders’, which may allow credit card market-ing to occur in high traffic routes of students. Title IIISEC 304 also reveals disclosure-related concerns relat-ing to limited and obscure student access to the credit cardagreements made between their college and credit cardissuers.

Stakeholder Ethical Concerns

Applying Stakeholder Theory to Title III SEC 305(Truth in Lending Act): College Credit Card Agree-ments reveals ethical concerns relating to the lack ofpriority put on college students as a key stakeholdergroup needing full disclosure. Colleges should provideeasily accessible and visible access (via a high traffic areaon their website) to students regarding (1) the college’scredit card agreements and affiliate program contracts;(2) filed complaint reports for credit card issuers, such asprovided by the Consumer Financial Protection Bureau;

(3) all credit card marketing practices involving theircampus or student body; and (4) all financial literacyprograms and resources made available by or through thecollege.

Quality of Life / Human Development Ethical Con-cerns

Applying Quality of Life to Title III SEC 305 (Truth inLending Act): College Credit Card Agreements revealsethical concerns relating to the government, as a whole,falling short of the desired quality of life / human devel-opment “end goals” for underage college students andtheir relationship with credit cards and debt. Specifically,concerns relating to the credit card / debt consumptioncycle and its effects on perceived quality of life, lifesatisfaction, overall happiness or subjective well-being,economic or financial health, social health, and percep-tion of a positive consumptive environment.

The ethics audit reveals that policy leaks still existregarding the Credit CARD Act, continue to threatenyoung adult consumer wellbeing, and may fuel a water-shed of predatory practices. To mitigate policy leak risks,policymakers are encouraged to use ethics audits toinform amendments and help establish a moral floor forthe Act’s marketplace effects. The authors make specificrecommendations for ethically informed amendments tothe Credit CARD Act. References are available uponrequest.

For further information contact:Connie R. Bateman

University of North Dakota293 Centennial Drive, Gamble 118

Grand Forks, ND 58202Phone: 701.777.4201



MOTIVATING CONSUMERS TO GET OUT OF DEBT FASTER

Keri L. Kettle, University of MiamiRemi Trudel, Boston University

Gerald Häubl, University of Alberta

SUMMARY

Imagine a consumer, Dave: like many consumers withproblematic levels of revolving debt, Dave carries bal-ances on five credit cards. Let’s suppose that the balanceon each credit card is $3000, and that Dave has justearned an extra $1000 that he wishes to use to pay downhis credit card debt. How should he allocate those repay-ment dollars? One possibility for Dave is to allocate anequal amount of money ($200) to each debt account andpay down the accounts simultaneously. Another possi-bility for Dave is to allocate the entire amount ($1000) toa single account, and thus focus on paying down theaccounts one at time. Is it possible that paying down oneaccount rather than five accounts would influence Dave’ssubsequent motivation to repay his debt?

We assume that most indebted consumers have the long-term goal of paying off their debt and accumulatingwealth. Achieving such a goal requires persistent self-control (Hoch and Lowenstein 1991; Vohs and Baumeis-ter 2004), which can be enhanced by the perception ofgreater progress toward goal attainment (Baumeister2002; Fishbach, Dhar and Zhang 2006). We propose thatpaying down debt accounts sequentially (versus simulta-neously) enhances perceived progress towards subordi-nate goals. That is, Dave will feel as though he has mademore progress by paying $1000 into a single accountrather than $200 into each of five accounts.

Prior work on the dynamics of self-regulation suggeststhat the perception of greater progress towards subordi-nate goals can lead to reinforcement and increased moti-vation to achieve a superordinate goal by signaling goalcommitment (Fishbach and Dhar 2005; Fishbach et al.2006; Huang and Zhang 2011). Making greater progresstoward successfully paying down smaller individual ac-counts – thereby enhancing perceived progress towardachieving more proximal subordinate goals – providesconsumers the opportunity to evaluate whether the goalof becoming debt free is attainable and worth pursuingfurther, and to infer higher commitment on the basis ofgreater perceived progress. For Dave, this suggeststhat perception of greater progress (from allocating

$1000 into a single account rather than $2000 acrossfive accounts) will motivate greater subsequent efforttoward his focal goal – getting out of debt.

We test our proposition with a unique consumer debtdataset and three experiments. Using credit card transac-tion data from indebted consumers and three experi-ments, we show that paying down accounts sequentially(versus simultaneously) leads consumers to infer greaterprogress toward their debt repayment goal, and thusincreases their motivation to repay their debt.

We obtained a large dataset from an online financialguidance company. The data span a 36-month period,and aggregate monthly credit card information includingpayments made and outstanding account balances. Basedon this, we used a dataset with a total of 68858 rowsrepresenting a member/year-month/account for a total of2522 members, with an average of four open credit cardaccounts (mean = 3.98, median = 3.00) per member andnearly $10,000 in total credit card debt (mean = $9858.03,median = $5433.96).

We measured each consumers’ tendency to engage insequential debt repayment, and examined whether thattendency has an impact on subsequent debt and subse-quent debt repayments. As hypothesized, we find thatconsumers who use a more sequentially oriented repay-ment strategy have less debt in the following period, andwe find that the use of a more sequentially orientedrepayment strategy in the previous month has a positiveinfluence on one’s debt repayment in the subsequentmonth. These effects are robust to varying levels ofindebtedness, credit card spending, and repaymentamounts.

Experiment 1 was designed to examine the effect ofdifferent debt repayment strategies on individuals’ per-ceptions of progress and motivation to repay debt. Wetold participants that they had five credit card accounts,each with a starting balance of $1200, and that their goalwas to repay all of their credit cards in a year. We thenrandomly assigned participants to a scenario in whichthey were told that, in the past month, they had managedto repay a certain amount of debt: we independently


manipulated the total amount of debt repaid (either $500or $1000) and that amount was allocated across their fivedebt accounts. In support of our theoretical account,individuals who had been assigned to a more sequentialdebt repayment strategy indicated that they had madegreater progress toward repaying their debt, indicatedgreater subsequent motivation to get out of debt, andperceived progress mediated the effect the debt repay-ment strategy on motivation.

Experiment 2 was designed to examine the effect ofdifferent debt repayment strategies on individuals’ moti-vation to repaying debt using a new measure of motiva-tion – work output. Participants began the experiment indebt (divided into five accounts of equal value) whichthey had to pay off before they were able to keep anyearnings. Participants could repay their debt faster byincreasing the amount of work they put in. Half of theparticipants were randomly assigned to either a simulta-neous repayment (their earnings were allocated equallyacross accounts) or sequential repayment (their earningswere allocated to accounts one at a time). The other halfof the participants were permitted to choose one of thetwo strategies. The results indicate that low self-controlconsumers worked harder to pay down their debt whenthey paid into their debt accounts one at a time rather than

simultaneously, and that this effect persisted whetherthey were randomly assigned to debt repayment strategyor chose it themselves.

In experiment 3, we examined how goal attainabilityaffects the interplay between self-control and debt repay-ment strategy by manipulating the amount that partici-pants could earn for their work. In support of our theoret-ical account, when the magnitude of these rewards was solow that it was nearly impossible for them to pay off theirdebt, low self-control individuals were not be motivatedby paying down their accounts one at a time; by contrast,when the rewards were sufficiently large for them torepay their debt, low self-control individuals producedsignificantly greater work output when paying downtheir debts sequentially rather than simultaneously, as inexperiment 2.

Consumer debt is a self-control issue of great importance.Using a unique consumer debt dataset and three experi-ments, our research suggests that indebted consumerscan be motivated to repay their debt more quickly byhaving them pay down their debt accounts sequentially(versus simultaneously). This research has importantpractical implications for indebted consumers and publicpolicy. References are available upon request.

For further information contact:Keri L. Kettle

University of MiamiCoral Gables, FL 33134

Phone: 305.284.6849E-Mail: [email protected]


SUSCEPTIBILITY TO THE CREDIT CARD EFFECT: A NEWCONCEPT AND MEASUREMENT SCALE

Sandra Awanis, University of Manchester, United KingdomCharles C. Cui, University of Manchester, United Kingdom

SUMMARY

In today’s vogue toward a cashless society, using creditcards as “plastic” money is now for most people thedominant payment form to replace cash money (Feinberg1986; Soman 2003). Credit cards are regarded as aspending stimulus that allows greater ease of spendingthan cash. Indeed, it has been extensively documentedthat some consumers display greater willingness to spendand make greater value purchases (a phenomenon called“credit card premium”), have weaker recollections ofpast credit expenses, and overvalue their availablefunds when using credit cards than cash in an other-wise identical purchase situation (Chatterjee and Rose2012; Raghubir and Srivastava 2008; Soman 2001, 2003).We refer to these inclinations as the credit card effect. Weconjecture that individuals become susceptible to thecredit card effect when their cognitive capabilities, psy-chological traits and financial sophistication come shortto comprehend complex credit card terms and foreseepost-payment negative effects. A review of the extantliterature reveals that there is lack of a concept to capturecredit card users’ exposure to such susceptibility to thecredit card effect. This paper contributes to the literatureby introducing a new concept of susceptibility to thecredit card effect (SCCE) and its measurement scale,which can be employed for better understanding of thisvulnerable behavioral inclination and wider issues incredit card misuse.

Susceptibility to Credit Card Effect

An individual’s susceptibility to the credit card effectreflects one’s incapacity to make rational, deliberativevaluation concerning their credit card usage. Such inca-pacity may be manifested in an individual’s partialitytoward cognitive biases, psychological aptitude towardimmediate or future consequences, and the lack of finan-cial sophistication.

A rational and strategic use of credit card requires aconsumer’s full awareness of available income and con-scious planning to make credit card repayment (includinghow much to repay and when to make repayments) inorder to achieve maximum utility and minimum financial

charges. However, consumers tend to make computa-tional or planning errors since some credit card cuesexhibit heuristic properties capable of influencing cogni-tive bias. Stewart (2009) showed that the presence ofminimum payment option led people to “anchor” theirdecision frame and make smaller repayments. In addi-tion, people’s capability of rational and strategic use ofcredit card may be undermined by their psychologicalaptitude toward immediate or future consequences.According to Wiener et al. (2007), psychological factorsaffect consumers’ ability to make rational credit cardusage deliberation in three ways. First, consumers mightbe able to accurately anticipate the high costs of overus-ing credit card, but still perceive the immediate payoffmore favorably than any long-term effects. Second, con-sumers might “choose” to ignore the long-term costs ofcredit card since the act of parting with actual money is adistant outcome. Third, consumers might inaccuratelypredict the emotional consequences of their actions,which mislead them to emotionally downplay the poten-tial high costs of repayment at the time of purchase(Wiener et al. 2007). Moreover, we argue that a consum-er’s financial sophistication acquired from credit cardknowledge and usage experiences help consumers toshape their cognitive ability to comprehend complexinformation inherent in credit card terms and aggressivesolicitations. For example, Lee and Hogarth (1999) con-tend that consumer knowledge and experience influencetheir ability to evaluate and distinguish between APR andCIR (contract interest rate).

Methodology

Our study focuses on young credit card users in Sin-gapore, Malaysia, and the U.K. These countries representdifferent degrees of credit card issuance and consumerprotection regulations, which systematically influencethe way people use their credit cards (Navarro-Martinezet al. 2011; Stewart 2009). We followed Castro, Kellison,Boyd, and Kopak’s (2010) recommendation for integra-tive mixed methods (IMM) to establish a multi-itemmeasurement scale of SCCE. IMM helps to establish ameasurement scale that offers rich, “deep structure”embodying individual as well as structural differences incredit card use. We began by conducting exploratory


interviews followed by surveys to young credit card users(18–25 years old) in Singapore (n = 6), Malaysia (n = 6)and the U.K. (n = 6). Based on the results of thematicanalyses, we developed 10 initial item pool. We furtherpurified the scale through further pilot test and EFA. CFAand multigroup invariance tests were subsequently con-ducted in a new set of samples in Singapore (n = 6),Malaysia (n = 6) and U.K. (n = 6) and we establisheddiscriminant, construct, and convergent validity. Fur-ther, we established that SCCE positively relates to thetendency to develop revolving credit card debt, whichprovide additional validity of the measure. The resultsindicate a 6-item measure of SCCE that is valid andreliable across the three countries.

Discussion

The concept and measure of SCCE provide an effectivedevice for credit card providers and regulators in their

market research and business auditory to identify themost vulnerable credit card users without obstructingfinancial freedom for young consumers. Subsequently,this approach can be utilised as a socially responsiblemarketing programme that empower consumers to bemore considerate in their credit card purchases. An exam-ple from the UK is the provision of text messages viamobile phone to remind people of their credit balanceswhenever they have reached a self-imposed limit. TheSCCE instrument may provide an integral part of SCCEdiagnostic facilities supported by other educational pro-grammes, such as provision of financial managementclasses and financial socialisation efforts to inform andempower young consumers to more effectively use creditcards. By understanding their own susceptibility, youngconsumers can improve their immunity to the credit cardeffect and improve knowledge and skills to makebetter credit card purchase and budgeting decisions.References are available upon request.

For further information contact:Charles C. Cui

Manchester Business SchoolUniversity of Manchester

Booth Street WestManchester, M15 6PB

United KingdomPhone: +44.161.3063461

Fax: +44.161.3063167E-Mail: [email protected]


TEMPTATION AND PREVENTION PROVIDED BY THE GAMBLINGINDUSTRY: MAIN AND INTERACTIVE

EFFECTS ON GAMBLERS

Linda Lemarié, HEC MontréalJean-Charles Chebat, HEC Montréal

SUMMARY

For industry officials and gambling defenders, gamblingis an effective way to create jobs and to raise revenuewithout raising taxes. However, these benefits come witha hefty social and economic price tag. The growing oflegalized gambling brings an increase in the number ofaddicted gamblers with what it all means: higherindebtedness and bankruptcy rates as well as increaseddivorces, suicides, and gambling-related crimes (Blaszc-zynski and Nower 2002). Statistics estimate that 816,000Canadian adults currently experience moderate to seri-ous gambling behavior problems (Wood and Williams2009). The reasons why gamblers become addicted arecomplex. However, non-profit organizations, public agen-cies, and researchers are focusing their attention on apossible cause of problem gambling development, thatis, gambling advertising (Binde 2007; Griffiths 2005;McMullan and Miller 2009; Sklar and Derevensky 2011).

In response to these accusations, some gambling com-panies have developed their own prevention programdesigned to educate people, especially gamblers, aboutthe harmful effects of gambling. In Quebec, currentgambling advertising comes from a monopolistic state-owned gambling corporation. This corporation sup-ports financially a non-profit organization whichintends to prevent problems related to excessive gam-bling behavior. In December 2010, this non-profit orga-nization broadcasted its annual awareness campaign.Obviously, gambling advertising did not stop during thiscampaign. This situation has been criticized by anti-gambling activists as potential conflict of interests andsome questions have emerged: what are the effects ofthese two types of advertisements? Are current industry-funded anti-gambling campaigns able to counteract theeffects of gambling advertising and to develop a moreresponsible attitude and behavior toward gambling? Ourgoal was to address these issues.

We tested two opposing theses regarding the effects ofpro- and anti-gambling ads both funded by the gambling

industry: the tainted fruit theory (Christenson 1992;Bushman and Stack 1996) versus the two-sided per-suasion paradigm. The tainted fruit theory posits thatanti-gambling ads will offset the effect of commercialadvertising and decrease positive gambling attitude andintention. On the contrary, the two-sided persuasionparadigm posits that anti-gambling ads messages willmake commercial gambling ads even more efficient andincrease positive gambling attitude and intention.

Turning to method, we first conduct a pilot study tochoose the gambling ads for the main study. First, acontent analysis of one hundred forty-seven gamblingads broadcasted on TV during the last five years wasperformed. Then, the set of message characteristic itemswas reduced through a factor analysis and the ads wereclassified in several homogeneous groups through acluster analysis. Five clusters or group of ads emerged:escape, feelings, thrill of winning, socialization and gam-bling skills. For the main study, we chose to focus on twoclusters: the Escape and the Gambling Skills clusters.The reason for this choice is twofold. First, reference togambling expertise, control or gambling ability has beencriticized by the literature as a way to enhance gambling-related problems (Blaszczynski and Nower 2002). Sec-ond, the use of gambling as a way to escape dailyproblems has been presented as a risk to develop gam-bling-related problems (Blaszczynski and Nower 2002).The three ads closest to the center in each of these twoclusters were preselected. In a pre-test, we asked 32undergraduate students to evaluate these six ads in termsof reference to expertise and/or to escapism and weretained. The ad which scored the most on expertise andthe ad which scored the most on escapism were retainedfor the main study.

In Study 1, we tested the first hypothesis regarding theimpact of gambling advertising on gamblers’ attitudesand intent. The design was a 2 X 1 randomized factorialdesign. The manipulated factor was the presence versusabsence of gambling ads. We used the two gambling adsselected during the pilot study. The main dependentmeasures were attitude toward the gambling industry,


attitude toward gambling, and gambling intent. Ninety-three adult gamblers from Quebec, randomly recruitedthrough a major online marketing research service, filledthe questionnaire. After completing a brief questionnaireabout their gambling behavior (gambling frequency,average money spent on gambling each year), the sub-jects were randomly assigned to one of the two condi-tions. The stimulus advertisements were shown at thebeginning of the well-known TV show and filler ads wereincluded in two other separate commercial breaks inorder to enhance realism. After viewing the televisionshow, the subjects completed the main questionnairewith the dependent measures (i.e., attitude toward gam-bling, attitudes toward the gambling industry, gamblingintent).

The results of this first study showed that exposure togambling ads enhances both gamblers’ attitude towardgambling and their gambling intent through the activa-tion of positive attitudes toward the gambling industry.Subjects exposed to a gambling ad tend to have morepositive attitudes toward the activity of gambling andtoward the gambling industry which led to higher gam-bling intent.

Study 2 tested our two latest hypotheses regarding theimpact of industry anti-gambling messages on gam-blers attitudes and intention. We hypothesized signif-icant interactions of pro-gambling ads and anti-gam-bling ads. The tainted fruit theory predicts that anti-gambling ads significantly reduce the effects of pro-gambling ads. The two-sided paradigm predicts thatindustry-funded and sponsored anti-gambling adsenhance positive attitudes toward the gambling industry,attitude toward gambling, and gambling intention.

We used the 93 subjects from Study 1 and recruited anadditional 97 adult gamblers through the same majoronline marketing research service for a total of 190 adultgamblers. Subjects in the anti-gambling ad conditionwere exposed to one of the two anti-gambling ads.Subjects in the mixed condition were exposed to one ofthe two anti-gambling ads followed by one of the gam-bling ads (randomly selected). The procedure and mea-sures were similar to Study 1, except for the inclusion ofanti-gambling advertisements.

The results supported the tainted fruit theory in thepresence of gambling ads whereas in the absence of

gambling ads, the results supported the two-sided persua-sion paradigm. It is reasoned that, in the presence ofgambling ads, the anti-gambling message primes a neg-ative image of gambling behavior, which cancels theimpact of gambling ads on gambling intent. In theabsence of gambling ads, anti-gambling messagesfunded and sponsored by the gambling industry havean opposite effect: they act as a gambling ad. Theyparadoxically enhance the gambling industry’s imageand gamblers’ gambling intent, which confirms the two-sided paradigm.

Study 1 suggests a two-step effect of gambling advertis-ing. In the first step, the gambling ads enhance gamblers’positive attitude toward gambling activities and improvethe gambling industry image. In the second step, positivebeliefs toward the gambling industry increase peopleintent to gamble.

Study 2 shows that anti-gambling ads cancel the effect ofpro-gambling ads on gambling intent if gamblers areexposed to them prior to gambling ads. However, itshould be noted that this effect appears only when theanti-gambling message competes directly with the com-mercial message and that the anti-gambling messageimpacts only on gambling intent. In contrast, in theabsence of gambling ads, the industry-funded anti-gam-bling message impacts positively on attitude toward thegambling industry and on intent to gamble. This findingis of utmost importance. It shows that such anti-gamblingmessages can serve the gambling industry’s interestsunder the apparent purpose of protecting gamblers fromincreasing their gambling behavior.

One limitation of the present research is that we did notmeasure the link between exposure to gambling ads andactual gambling behavior. Another limitation concernsthe manipulation itself. It is interesting to note thatindustry-funded anti-gambling messages may offset theimpact of gambling ads on gambling intent. In Study 2,the presence of an industry-funded anti-gambling mes-sage presented just before the gambling ad decreasedgambling intent. However, this situation is, in reality,quite unusual. Gambling ads and anti-gambling messag-es have seldom to compete directly. Future researchcould strive to replicate these results in more realisticsettings to examine whether these results were influencedby the experiment setting itself. References are availableupon request.


For further information contact:Linda Lemarié,HEC Montréal

3000, Chemin de la Côte-Sainte-CatherineMontréal, Canada, H3T 2A6

Phone: 001.514.340.6431Fax: 001.514.340.6097



THE RELATIONSHIP BETWEEN SOCIAL MARKETING POLICIESON PUBLIC HEALTH AND EPIDEMIOLOGICAL SITUATIONS:

AN EXPLORATORY RESEARCH BASED ON TELEVISIONADVERTISEMENTS PREVENTING HIV/AIDS INFRANCE, GERMANY, PORTUGAL, AND ITALY

Beatriz Casais, University of Porto, PortugalJoão F. Proença, University of Porto, Portugal

SUMMARY

Social Marketing has been growing importance in the 21st

century and it has being increasingly adopted all over theworld (Andreasen 2003; Andreasen 2002). It consists ofadapting commercial marketing tools to the promotion ofsocial change (Dann 2010). Social marketing has beeneffective in the encouragement of behavior changes,especially those that are related to public health (Evansand McCormack 2008; Keller and Lehmann 2008; Steadet al. 2007). Though effective, the question is how socialmarketing is done to promote healthier behaviors along-side the epidemiological dynamics (Vega and Roland2005).

HIV/AIDS has been an important cause of death and itstransmission is caused by unsafe behaviors (Merson et al.2008). It is an important topic for social marketing(Lefebvre 2011).

Prevention efforts are negative correlated with HIV prev-alence and incidence rates (Merson et al. 2008). TheAIDS Program Effort Index (Merson et al.) is based on ananalysis of 54 countries and reveals that the nations withthe highest prevalence rates have on average lower APIscores than the ones with relatively low prevalence rates(USAID et al. 2003). This reveals the importance ofprevention efforts in epidemic dynamics.

One of the most important and effective tools of socialmarketing in Public Health is television advertising(Abroms and Maibach 2008; Block and Keller 1995;Della et al. 2008; Hastings et al. 1998). The effortsfocusing on HIV prevention constitute extensive archivesof media campaigns, generally considered effective(Mattson and Basu 2010; Noar et al. 2009).

As the control of infection is strategically linked to strongprevention plans, that should include social marketing

actions (EU 2009; UN 2001), this paper analyzes therelationship between HIV prevention advertisementsbroadcasted on TV in France, Germany, Portugal, andItaly and the epidemic characteristics in these four Euro-pean Union countries. The paper identifies the targets andmessages of those advertisements and discusses therelationship between the management choices and thedescription of HIV epidemics in each country. Theobjective of the research is to discuss the use of this toolof prevention, the relationship between the social mar-keting choices and the epidemiological situations andevaluate the social marketing policies and investmentsthat have been done, improving also future social market-ing strategies.

Considering that the majority of advertisements broad-casted on Television, even the oldest ones, are digitalizedand shared online (Paek et al. 2010; Spigel 2009), datacollection was done in the internet. The internet is themost accessible and convenient tool to collect this kind ofdata, because of the rising routine of sharing videos indifferent websites and social media and the increasingneed to digitalize old media archives in the web as abackup of analogical data (Paek et al. 2010; Spigel 2009).

The database was built by in-depth research into thenational virtual media archives of the countries belong-ing to the European Union (EU) since the 1st of Januaryuntil the 30th of June 2012, when the researchers foundsaturation in the searching process. We searched for datain: (a) institutional websites and Facebook profiles of thegovernmental institutions coordinating HIV/AIDS pre-vention; (b) institutional websites and Facebook profilesof NGOs members of “Aids Action Europe” in each EUcountry; (c) online media archive INA.fr; (d) onlinemedia archive culturepub.fr; (e) video-sharing websitesYoutube, Vimeo, dailymotion and Google videos. Thesearching expressions in video sharing websites were“HIV AIDS Prevention advertisement” in English fol-lowed by the name of each EU country also in Englishbecause of the fact it is the most important international


language and it could help to find videos. Then, wetranslated the expression to each of the twenty-threeEuropean official languages.

After the collection of 539 television advertisementsdeveloped in 21 EU countries between 1986 and 2011,we selected 375 (69.6%) that belonged to only fourcountries: 146 are from France (38.9%), 115 from Ger-many (30.7%), 76 from Portugal (20.2%) and 38 fromItaly (10.1%). We coded and classified the 375 dataaccording to the country and year of broadcast, the source– Governmental, NGO or a partnership of both -, thechanging behavior message identifies and the main tar-get. We classified and analyzed data through Nvivosoftware. The classification allowed a cross analysis ofthose variables, showing the trends of social marketingchoices.

The results show that the relationship between socialmarketing policies regarding the production of HIV pre-vention television advertisements and the epidemiologi-cal situation in each country is not consistent. Moreover,when there is a relationship between marketers’ choicesand epidemics, the advertisements are in general part ofa reactive prevention policy.

Television advertisements target especially general peo-ple and young people, suggesting an avoidance of target-ing the most vulnerable populations to HIV. It may

suggest a concerning to reduce the stigma attached tothem, but contradicts what is known about marketingsegmentation, even in social marketing (Fine 1980).

We may conclude that social marketers need to adaptbetter advertisement messages to the specificities of eachepidemic in order to use this tool to build an activeprevention public policy instead of a reactive policy. Thisresearch also concluded that the internet tool to collectadvertisements is suitable for some countries, but not toothers.

The discussion may be useful for management inpublic sector and non-profit organizations, concern-ing the evaluation of the social marketing policies andthe analysis of investments to improve future socialmarketing strategies. The results are connected withthe analyzed data, whose process of collection wasdescribed. Future research may analyze communica-tion appeals in order to understand possible relation-ships with the epidemiological dynamics. Referencesare available upon request.

NOTE

The author João F. Proença thanks Fundação Ciência eTecnologia (FCT Portugal) for financial support to AD-VANCE by the Multi-Year Funding Program for R&DUnits.

For further information contact:Beatriz Casais

EconomicsUniversity of Porto

Rua Dr. Roberto Frias4200-464 Porto

PortugalPhone: 00.351.22.55.71.100

Fax: 00.351.22.55.05.050E-Mail: [email protected]


POSITIVE MARKETING TO PROMOTE HEALTHY FOODS ANDADVANCE FOOD WELL-BEING

Melissa G. Bublitz, University of Wisconsin – OshkoshLaura A. Peracchio, University of Wisconsin – Milwaukee

SUMMARY

What differentiates marketing for an unhealthy snackfood from those efforts designed to promote a healthyalternative? Advertising efforts that promote food tochildren often use appeals that are “hedonistic, includingtaste, humor, action-adventure, and fun” (Hastings et al.2003). Food marketing to adults frequently uses emotionto emphasize pleasurable food experiences (Geuens et al.2011). While marketing communications for healthyproducts tend to focus on informing consumers of thenutrition and health benefits of consuming the product,advertising efforts for hedonic foods and beverages oftenfocus on the sensory experience of consuming a product(Hastings et al. 2003). While the marketing practicesused to promote healthy versus hedonic products differ,research implies that when low-involvement, utilitarianproducts (e.g., healthy foods) employ humor or sexappeal, these unexpected advertising strategies mayimprove brand attitudes. Yet, producers and marketers ofhealthy foods often focus on cognitive information, suchas nutrition information and other dimensions of health-fulness, and neglect affective appeals when they promotetheir products to consumers.

In this paper we present an organizing framework formarketing communications with a focus on how to trans-late positive marketing insights into an opportunity topromote healthy foods and beverages. For example,creative ad campaigns like those used to promote Califor-nia Raisins in the 1980’s or those that use celebrityendorsements as part of the long running “Got Milk!” adcampaign adopt industry best practices to promote ahealthy food or beverage in a positive manner. In thispaper, we examine how positioning strategies, emotionalappeals, advertising schedules, and alternate media maybe used by the producers, growers and marketers ofhealthy foods to advance food well-being.

Combining pleasurable taste cues with a specific healthclaim may help promote healthy foods by adopting a dualmessage strategy to position a product as both tasty andhealthy. For example, apple producers could use visual

techniques to promote their product not only as a “crisp,refreshing snack” but also as one that is a “good source offiber” echoing the advertising approach of many pack-aged snack foods. Targeting health conscious consumerswith messages designed to satisfy their desires for hedon-ic indulgences while simultaneously helping them pur-sue their health goals, producers may use positive mar-keting to change the expectations and appeal of healthyproducts.

While we typically think of hedonic taste appeals inrelation to less healthy products, it would also be quitepossible to use the same approach to describe a healthyfood. For example, Stoneyfield Organic employed bothsexual innuendo and humor in their 2012 Super Bowladvertisement to introduce Oikos, a new brand of Greekyogurt. Using descriptors such as “incredible creami-ness” and “made from fresh, delicious, organic milk” thecompany creates indulgent taste expectations for theirproduct (www.stoneyfield.com). Creative ads like theOikos commercial work in two ways. First, they captureconsumers’ attention (Smith et al. 2007) by creating adesire to postpone closure, motivating consumers toprocess the ad (Yang and Smith 2009). In addition,creative ads trigger “positive emotions that transfer tointention, Eugenes … even when consumer involvementis low” (Yang and Smith 2009, p. 945). When consumersare in a positive affective state, research indicates theyattend more to visual information as well as the bene-fits of a product (King and Janiszewski 2011). Emo-tional responses and enhanced taste perceptions mayalso be achieved by using product descriptions thatexplicitly describe the sensory experience consumerscan expect when they consume a healthy product (e.g.,Elder and Krishna 2010; Lwin et al. 2010). Producers ofhealthy food products should use these creative tech-niques to attract attention, transfer positive feelings, andenhance advertising effectiveness.

When making food choices, consumers use informationthat is most easily accessible within their own memory toform attitudes about a product or brand (Menon et al.2003). One way to make the affective attributes of healthy


foods more accessible in the minds of consumers is toengage in higher levels of ad repetition and repetitionacross media vehicles. Repetition creates familiarity andpositive affect through incidental exposure (Ewing et al.1999) and increases demand for the product. As advertis-ers explore media alternatives and new technologies,executing repetition across different media types such astelevision and the web create synergy that enhances theeffectiveness of ads (Chang and Thorson 2004). Thepower of repetition achieved via alternative mediaincluding social media and mobile advertising mayalso help place the cost of repetition within reach for themarketers of healthy foods.

By examining the marketing practices for hedonic foodsand beverages, entities trying to promote healthy foodchoices may discover opportunities to entice consumerstoward healthier options. Beyond educating and inform-

ing consumers, efforts to promote healthy eating shouldintegrate a positive approach to emphasize taste as well asother sensory properties likely to trigger an affectiveresponse and increase the desirability of healthy prod-ucts. Using mainstream appeals that draw upon emotionsand emphasize the sensory dimensions of a food togetherwith its health benefits may attract the attention of con-sumers and impact food choices. Many health advocatescriticize the marketing and advertising efforts of foodmanufacturers and the restaurant industry for contribut-ing to the obesity epidemic by increasing demand for lesshealthy foods. However, by examining the successfuladvertising practices used by these commercial entities,marketers of healthy alternatives together with healthadvocates and policymakers may learn from their com-mercial practices and identify new pathways to positivelypromote healthy food choices and advance food well-being. References are available upon request.

For further information contact:Melissa G. Bublitz

College of BusinessUniversity of Wisconsin – Oshkosh

00 AlgomaOshkosh, WI 54901

Phone: 920.424.1013Fax: 920.424.7413



CHOICE, CONTROL, AND THE PARADOX OF SMOKINGDURING PREGNANCY

Janet Hoek, University of Otago, New ZealandNinya Maubach, University of Otago, New Zealand

Heather Gifford, Whakauae Research Services, New ZealandStephanie Erick, Tala Pasifika, New Zealand

Richard Edwards, University of Otago, New ZealandRhiannon Newcombe, Health Promotion Agency, New Zealand

SUMMARY

Numerous studies have documented the health risks ofsmoking during pregnancy; these include increased riskof miscarriage, low birth weight babies, stillbirth, andsudden infant death (SIDS). This well-established evi-dence has led public health researchers and health pro-moters to encourage pregnant women who smoke to quit.However, despite these efforts, many women continue tosmoke during and after pregnancy, putting themselvesand their children at risk.

Pregnant women who smoke typically come fromcommunities where smoking is commonplace and theirbehavior more accepted; this perceived normalcyreduces the urgency that commonly prompts andmotivates quitting. Some pregnant women may also seequitting as segregating them from their social network ata time when they require greater support. Further, theymay see other women smoking while pregnant, and maythemselves have been born to smokers; comparisonsbetween their own experience of living among smokersand the lack of perceived harm may support beliefs thatsmoking poses few, if any, risks to unborn children.Thus, despite the wider social disapproval many preg-nant women report experiencing, acceptance of smokingwithin their immediate social network reassures themmay reduce their motivation to quit.

Cessation interventions must recognize the ambivalencemany pregnant women feel about smoking, which issimultaneously stigmatized, commonplace, and seen asbeneficial. Messages need to counter views of smokingas a reward, challenge the self-exempting beliefs preg-nant smokers construct, and provide stimuli that promptand support quit attempts while also avoiding reactance.Cognitive dissonance theory provides a framework forexploring this ambiguity as it recognizes the tensiongenerated when people’s beliefs conflict with theirbehavior and the potential for behavior change. Women

who are pregnant and smoking may experience acutetension following their frequent contact with health pro-fessionals, who will advise them to quit, and if they attractstrong social disapproval. Some resolve this tension byquitting but others discount the risks of smoking and usethe apparent health of other smokers’ babies as evidencethat smoking is not inevitably harmful. These variedoutcomes raise questions about how social marketersmight use the tension created by denormalization, knowl-edge of smoking’s harms, and advice from health profes-sionals, to promote cessation among pregnant women.

We addressed these questions by exploring how womenrationalize smoking during pregnancy and their responseto cessation messages designed to elicit varying levels ofaffect. The two phase study first involved in-depth inter-views with 13 women who were pregnant and smoking,or who had recently given birth and smoked during theirpregnancy. Participants discussed smoking’s role in theirlives and their perceptions of images presented in aphotosort task. Themes identified informed cessationadvertisements that were tested in a second phase, com-prising in-depth interviews with a further 21 women whowere either pregnant or who had recently given birth andwho had smoked while pregnant.

We identified two dominant metaphors in the Phase onedata: choice and control. Participants tried to assertsmoking as a choice over which they maintained control,even though none had made an active choice to becomea smoker. They used choice metaphors to frame smokingand quitting as a “choice” only they could make and todistance themselves from stereotypes of addicted smok-ers. Depictions of smoking as a choice placed theirbehavior within a rational framework, which separated itfrom emotional consequences and enabled counter-argu-ments. However, for all their efforts to establish smokingas a choice, participants struggled to maintain this priv-ileged position of control and autonomy; the resultingdissonance weighed heavily on many and, ironically,reflected the lack of power they felt.


The dominance of control and choice metaphors in par-ticipants’ discourse suggested three potential cessationmessage themes; the first two themes aimed to arousenegative affect by illustrating the effects of smoking onbabies who had no choice in being exposed to toxins, andthe consequences children face when their parents areharmed by smoking. The final theme recognized smok-ers’ autonomy, invited them to quit and called on them torecognize children’s right to a smokefree life. The firsttwo themes allowed “choice” to be imbued with moreaffect-laden attributes that could challenge the reasonedpositions smokers had constructed. The third approachavoided a didactic tone but attempted to engage withsmokers’ reasoning, provided opportunities to adoptsmokefree behaviors, and challenged smokers to con-sider others’ rights.

Phase Two participants reacted strongly to messagesdepicting the first two themes, which graphically depict-ed harms caused by smoking and showed the distress of

children who could become tobacco orphans. Their emo-tionally-laden responses highlighted children’s lack ofchoice and few participants embarked on counter-argu-ments. However, while participants said they resentedreceiving quitting advice, which they thought impingedon their rights, messages that appealed to their auton-omy were less effective and more likely to stimulaterationalizing.

In summary, messages depicting harm cut through ratio-nalized defenses and elicited strong reactions that partic-ipants found hard to dismiss. Messages reframing choicefrom a child’s perspective were thought more likely toprompt cessation than rational themes, and were alsomore resistant to counter-argument. However, evidenceof rationalization suggests that social marketing, whileimportant, is not a panacea and should not pre-emptaction in other domains. References are available uponrequest.

For further information contact:Janet Hoek

Department of MarketingUniversity of Otago

P.O. Box 56Dunedin 9054New Zealand

Phone: +643.479.7692E-Mail: [email protected]


IS THE PRICE RIGHT? POLICY IMPLICATIONS ASSOCIATEDWITH SCANNER DATA INACCURACY

David M. Hardesty, University of KentuckyAnthony Miyazaki, Florida International UniversityRonald Charles Goodstein, Georgetown University

Dhruv Grewal, Babson UniversityPraveen Kopalle, Dartmouth University

SUMMARY

There was widespread publicity surrounding price scan-ning accuracy in the United States in the 1980s and1990s. Since then, the U.S. Federal Trade Commission(FTC) has, for budgetary or other reasons, shifted itsefforts to other issues leaving individual state regulatingagencies to determine whether scanner accuracy will bemonitored and enforced, all while other countries havebegun more fervent investigations into price scanneraccuracy and regulation (e.g., CCGD 2008; Governmentof South Australia 2008). Here, we evaluate the degree ofscanner inaccuracies existing across different retailertypes. Data from the most recently released inspection

period conducted by officials from the U.S. Weights andMeasures offices in thirty-six states and the U.S. VirginIslands and a more recent data set garnered in a statewidestudy are used to establish the extent of inaccuracy. Theempirical findings demonstrate that systematic errorsstill plague the data and vary across retailer types thoughtheir levels have declined compared to prior results.Next, survey results are summarized illustrating howscanner inaccuracies impact consumers’ reactions tostores. Specifically, buyers express negative attitudes,lower patronage intentions, and more negative word-of-mouth intentions when overrings are more prevalent thanunderrings. Finally, implications for public policy arediscussed. References are available upon request.

For further information contact:Ronald Charles Goodstein

Georgetown University519 Hariri Bldg

Washginton, D.C. 20057Phone: 202.687.8009



PRODUCT CONTENT RATINGS IN THE ENTERTAINMENTINDUSTRY

Joseph Matthes, University of Nebraska – LincolnLes Carlson, University of Nebraska – Lincoln

SUMMARY

Currently in our society children, teenagers, and youngadults have legal access to a plethora of different types ofmedia on a daily basis. These potentially influentialsources include television, movies, music, video games,computer games, the Internet, books, and much more.Due to the natural existence and subsequent artisticemulation of mature subject matters involving violence,harsh language, nudity, and drug use; the content avail-able through these assorted forms of media varies greatly,ranging from perfectly acceptable to extremely inappro-priate for younger audiences.

In response to societal concerns about undo exposure toyouths, several ratings systems were voluntarily devel-oped by various media organizations in order to ideallyprovide accessible, easy to understand information toadults (ESRB 2011a; MPAA 2011a; RIAA 2011). Suchratings are intended to assist parents in choosing what isappropriate for their children to consume without havingto independently research every single entertainmentproduct that is made available. However, even if someparents maintain high levels of discipline with theirchildren, kids are still potentially able to have friendswith a greater level of access to such products. Theproblem of keeping adult only forms of media out ofyoung hands is one shared by parents, businesses, gov-ernment, and society as a whole.

Due to concerns about exposure to youths, the regula-tion of various entertainment industries is becomingan increasingly important topic in American society(Anders 1999; Collier, Liddell Jr., and Liddell 2008;Federal Trade Commission 2000, 2001a, 2001b, 2002,2004; Grier 2001; Hemphill 2003; Hoy and Andrews2006). As concerns continue to grow about the potentialeffects of violent and adult-oriented entertainment onminors, it is possible that there will be additional calls for

increased regulation. It is the public responsibility ofeach entertainment industry to stringently self-regulatetheir content in order to avoid government intervention(Gupta and Lad 1983). Specifically, inappropriate prod-uct content must be kept away from individuals that aredeemed too young to safely consume the product. Thisincludes regulation of content ratings, as well as adver-tisements and product placements within these forms ofentertainment. Furthermore, such adult-oriented prod-ucts should not be promoted through outlets that purpose-fully target underage consumers.

Several research questions are evaluated: (a) What is thestatus of product regulation in the entertainment indus-try? (b) What are the risks associated with adult-orientedmaterial in the entertainment industry? (c) What types ofratings systems are preferred by consumers? (d) Whichtypes of ratings systems are most effective? (e) Should allforms of entertainment media use a compatible ratingssystem?

Though some research has been completed in this area,this paper develops a broader analysis of the issues frommarketing management and public policy perspectives.First a brief history of the motion picture, recorded music,and electronic game industries is provided along with areview of their current entertainment ratings systems.Then these three industries and their self-regulatory ef-forts are examined in terms of how they are affected bygovernment regulation (e.g., Federal Trade Commission2000, 2001a, 2001b, 2002, 2004), the potential negativeeffects resulting from consumption of violent media byyouths (e.g., Anderson and Bushman 2001; Andersonand Dill 2000; Bushman and Anderson 2002), as well aspast, present, and prospective legal actions (e.g., Anders1999; Collier, Liddell Jr., and Liddell 2008; Liptak 2011).Lastly survey results that propose and test multiple stan-dardized ratings systems are analyzed and discussed.References are available upon request.


For further information contact:Joseph Matthes

University of Nebraska – LincolnCBA 330a

P.O. Box 880492Lincoln, NE 68588–0492

Phone: 402.578.9936Fax: 402.472.9777



THE POLITICAL CEO: AN EVENT STUDY COMPARINGCONSUMER ATTRIBUTIONS OF CEO BEHAVIOR

Jessica Mikeska, University of Nebraska – LincolnElise Johansen Harvey, University of Nebraska – Lincoln

SUMMARY

The treatment of top management in the existing litera-ture has emphasized leadership style, personality, and therelationship between behavior and performance. Thepopular press and media, instead, provide coverage of thepublic image of the Chief Executive Officer (CEO)including CEO hobbies, political views, and other behav-iors unrelated to their position in top management. Thediscrepancy in coverage leaves marketing researchersand consumers with differing perspectives on CEObehavior, which leaves us with a gap in the existingliterature.

This study attempts to understand consumer attribu-tions of CEO political behaviors and if consumers usepersonal, but publicized political beliefs of CEOs toevaluate the coupled firm and the firm’s products. Theobjectives of the study are to establish a relationshipbetween negative political behavior of key members ofthe firm and consumer evaluations of the firm, toextend the use of attribution theory in marketing toinclude consumer explanations of events unrelated tofirm strategy or operations, and to compare and con-trast both the consumer attribution process and firmconsequences of volitional and situational events. Wedefine a volitional event as an instance when an organi-zational member publicly volunteers his/her personalwishes, preferences, or value and a situational event as anevent that only once or on an irregular basis and informsthe public about operational problems rather than firmvalues. Drawing from attribution theory, volitional eventswere expected to be attributed as stable and, thus,attributed internally, whereas situational events areexpected to be attributed as unstable, and thus attri-buted externally. Further, negative affect among con-sumers with internal attributions was expected to begreater following a negative volitional event as comparedto a negative situational event. Among these consumers,purchase intentions were expected to cease to a greaterdegree following a negative volitional event as comparedto a negative situational event.

We sample and code approximately 1,500 social mediaimpressions on the Twitter platform total surroundingtwo events in the fast food industry related to distinctareas in public policy: (1) Chick-fil-A CEO Dan Cathy’spublic volunteering of political beliefs about the defini-tion of marriage (volitional event) and (2) Taco Bell’spublic response to a class-action lawsuit suggesting pro-motions about beef menu items are false (situationalevent). Both events are shown to be considered negativeby consumers and the media. Social media was chosen asthe data source for the study for a number of reasonsincluding its viability as a source for measuring brandimpressions, its accessibility and permanence in its rolein electronic word-of-mouth, the prevalence of use amongconsumers, the ability to gain a nearly real-time glimpseinto consumers’ experiences of marketplace events, whichalso avoids problems with respondent memory, and theability to avoid validity threats such as response bias dueto the naturally occurring data without researcher in-volvement. The social media data were gathered by theresearchers in a time frame surrounding the events thataccounted for both pre-event and post-event consumerresponses. The data were coded by one research and oneindependent coder for affect, consumer attribution, andconsumer response (future purchase intent). Interrraterreliability Kappa coefficients range between .58 and 1.0,and all reflect at least acceptable agreement betweenraters.

The results of the study suggest that consumers utilizeinformation about CEO behavior in their firm evalua-tions. We find consumers attribute CEO political behav-iors, though unrelated to firm operations, toward the firm.In other words, spillover can occur from top manage-ment’s personal, political behavior to the firm itself. Ourfirst objective involved establishing a relationship be-tween political behaviors of key members of the firm andconsumer evaluations. Our findings do not offer strongsupport for a relationship between event type and affect,suggesting negative affect is likely to follow both voli-tional and situational events similarly. Our second objec-tive involved extending attribution theories to includeconsumer explanations of events unrelated to firm strat-


egy or operations. Based on analysis social media data, itappears that consumers are more likely to attribute disap-proval of CEO political behavior internally (toward thefirm) than externally (toward forces outside the firm).Our hypotheses suggesting attribution type (i.e., internal,external locus of cause) will influence affect towardthe firm, however, involved mixed support. On the onehand, attribution type does not appear to influenceaffect toward the firm, but on the other hand, no Tweetsof the situational event indicated decisions to no longerreturn to Taco Bell while all Tweets of the volitionalevent indicated decisions to no longer return to Chick-fil-A. Our final objective involved comparing and contrast-

ing both the consumer attribution process and firm con-sequence of volitional and situational events. We foundthat a product lawsuit (situational) is more likely toinvolve internal attributions than a firm’s public responseto CEO political behavior (volitional). Consumers mayexperience more negative affect because they trustedpublic policy to regulate product-related operations as ina product lawsuit, but have no such public policy expec-tations when it comes to CEO political behavior. Overall,this study confirms that media representations of theindividual CEO are useful to the public in predicting andunderstanding the abstract firm. References are availableupon request.

For further information contact:Elise Johansen Harvey

University of Nebraska – Lincoln330 College of Business Administration

Lincoln, NE 68588Phone: 402.472.3279



AUTHOR INDEX

Aaron Garvey, M. 3Acevedo, Claudia Rosa 68Aikin, Kathryn J. 46Albrecht, Carmen-Maria 7Apostolovski, Dijana 36Argo, Jennifer 72Askim-Lovseth, Mary K. 76Awanis, Sandra 80Baek, Tae Hyun 74Ball, Dwayne 23Bateman, Connie R. 76Bauer, Hans H. 7Bearden, William O. 5Bergner, Kai 7Bolton, Lisa E. 3Bublitz, Melissa G. 87Camilleri, Adrian R. 15, 62Carlson, Les 23, 92Carroll, Ryall 17Casais, Beatriz 85Catlin, Jesse R. 58Chebat, Jean-Charles 82Chowdhury, Dhuly 46Cui, Charles C. 80da Silva Tamashiro, Helenita R. 68da Silveira, José Augusto Giesbrecht 68Davis, Brennan 31DeMotta, Yoshiko 1Dorsey, Joshua D. 60Edwards, Richard 42, 89Ehrich, Kristine R. 9Eisenstein, Eric M. 29Erick, Stephanie 89Fitzgerald, M. Paula 60Fox, Alexa K. 32Frels, Judy K. 64Gendall, Philip 4Gershoff, Andrew D. 64Gifford, Heather 89Goerke, Ashley 29Gogia, Tamara 7Goodstein, Ronald Charles 91Gould, Stephen 1Grewal, Dhruv 91Hagenbuch, David J. 44Hardesty, David M. 91Harvey, Elise Johansen 94

Häubl, Gerald 78Hoek, Janet 42, 89Howell, Gwyneth 36Inman, J. Jeffrey 21Irmak, Caglar 5, 40Irwin, Julie R. 9Johnson, Eric J. 15Juan, WenYen 21Kaufman-Scarborough, Carol 27Kettle, Keri L. 78Khare, Adwait 21Kim, Seeun 74Kopalle, Praveen 91Koppenhafer, Leslie 48Kowalczyk, Christine M. 66Kumar, Minu 11Larrick, Richard P. 15, 62Lee, Alvin 38Lemarié, Linda 82Liang, Bo 34Lin, Chung-Tung 21Luchs, Michael G. 11Madrigal, Robert 48Martinez, Jennifer 66Mather, Damien 42Matthes, Joseph 92Maubach, Ninya 89Mikeska, Jessica 23, 94Miller, Elizabeth Gelfand 58Miller, Rohan 36Minton, Elizabeth 13Mirabito, Ann M. 58Mitakakis, Chrissy 19Miyazaki, Anthony 91Mizerski, Dick 38Moultrie, Rebecca R. 46Murdock, Mitchel R. 40Naylor, Rebecca Walker 9Newcombe, Rhiannon 89O’Donoghue, Amie C. 46Peracchio, Laura A. 87Proenca, João F. 85Robinson, Stefanie Rosen 5Rosenbaum, Mark S. 25Royne, Marla B. 32Rupert, Douglas J. 46Ruvio, Ayalla 29


Scammon, Debra L. 34Scott, Kristin 70Sen, Sankar 1Soule, Catherine Armstrong 48Sprott, David E. 7Sullivan, Helen W. 46Sweeney, Jillian 25Thrasher, James F. 40

Trudel, Remi 72, 78Ungemach, Christoph 15Vallen, Beth 17, 19Wang, Shasha 38Weaver, S. Todd 70Weber, Elke U. 15Yang, Jinchao (Alex) 38Yoon, Sukki 7

Date post:	09-Oct-2020
Category:	Documents
Upload:	others
View:	0 times
Download:	0 times

Volume 23 The Changing Role of Policy in Consumer Well-Being · 2013 Marketing & Public Policy...

Documents