Proceedings of the

Society for Consumer Psychology 2004 Winter Conference

February 19th – 21st, 2004 San Francisco, California

Amar Cheema, Scott Hawkins, and Joydeep Srivastava, Editors

ii

© Copyright 2004, Society of Consumer Psychology All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form, by any means, electronic, mechanical, photocopying, recording, or otherwise, without prior written consent of the publisher. Amar Cheema, Scott Hawkins, and Joydeep Srivastava, Editors

i

List of Reviewers

Lisa Abendroth , Boston University Rohini Ahluwalia , University of Kansas On Amir , Yale University Eduardo Andrade , University of Florida Jennifer Argo , University of Alberta Dan Ariely , Michigan Institute of Technology Raj Arora , University of Missouri-Kansas City Debra Basil , University of Lethbridge Michael Basil , University of Lethbridge Namita Bhatnagar , University of Manitoba H. Onur Bodur , Concordia University Carolyn Bonifield , University of Vermont David Boush , University of Oregon Terry Bristol , Arizona State University Merrie Brucks , University of Arizona Anne Brumbaugh , Wake Forest University Frederic Brunel , Boston University Margaret Campbell , University of Colorado, Boulder Rich Campbell , California State University, Bakersfield Amitav Chakravarti , New York University Joseph Chang , University of Regina Alex Chernev , Northwestern University Magdalena Cismaru , University of Regina Sooraj Commuri , University of Missouri-Columbia George Comstock , Syracuse University Robin Coulter , University of Connecticut Sandor Czellars , University of Florida Darren Dahl , University of British Columbia Kristin Deihl , University of South Carolina Stephanie Dellande , Chapman University Kalpesh Desai , State University of New York, Buffalo Utpal Dholakia , Rice University Aime Drolet , University of California, Los Angeles Jane Ebert , University of Minnesota Jennifer Escalas , University of Arizona Sasha Fedorikhin , University of Southern California Carrie La Ferle , Michigan State University Elizabeth Gelfand-Miller , Boston College Jim Gentry , University of Nebraska - Lincoln Andrew Gershoff , Columbia University Bryan Gibson , Central Michigan University John Godek , University of Oregon Inge Goh , National University of Singapore Susan Jung Grant , University of Colorado, Boulder Jennifer Gregan-Paxton , University of Delaware Curt Haugtvedt , Ohio State University Timothy Heath , Miami University of Ohio Paul Herr , University of Colorado, Boulder Susan Hogan , Emory University Heather Honea , San Diego State University Lee Yih Hwai , National University of Singapore Jeff Inman , University of Pittsburgh William James , Hofstra University Christopher Joiner , George Mason University

ii

Lisa Klein , Rice University Rob Kozenitz , Northwestern University Partha Krishnamurthy , University of Houston H. Shankar Krishnan , Indiana University Aparna Labroo , University of Chicago James Leigh , Texas A&M University Donlad Lichtenstein , University of Colorado, Boulder Peggy Sue Loroz , Gonzaga University Tina Lowrey , University of Texas – San Antonio Nicholas Lurie , University of North Carolina Bob Madrigal , University of Oregon Naomi Mandel , Arizona State University Susan Mantel , Indian University-Purdue University, Indianapolis Anna Mattila , Pennsylvania State University Michael McCarthy , Miami University of Ohio Norma Mendoza , University of Arkansas Tom Meyvis , New York University Vikas Mittal , University of Pittsburgh Sonia Monga , University of Minnesota Page Moreau , University of Colorado, Boulder John C. Mowen , Okhlahama State University Anirban Mukhopadhyay , Columbia University Nada Nasr , Bentley College Joe Nunes , University of Southern California Gillian Oakenfull , Miami University of Ohio Christine Page , Skidmore College Vanessa Patrick , University of Southern California Lisa Penaloza , University of Colorado, Boulder Steve Posavac , University of Rochester Rajagopal Raghunathan , University of Texas - Austin Priya Ragubir , University of California, Berkeley Sekar Raju , State University of New York, Buffalo Suresh Ramanathan , University of Chicago Rebecca Ratner , University of North Carolina Anne Roggeveen , Babson College Julie Ruth , Rutgers University Sridhar Samu , University of Manitoba Robert Schindler , Rutgers University Ann Schlosser , University of Washington Baba Shiv , University of Iowa L. J. Shrum , University of Texas – San Antonio Jennifer Siemens , University of Dayton Subramanian Sivaramakrishnan , University of Manitoba Sanjay Sood , University of California, Los Angeles John Story , University of Texas – San Antonio Suresh Sundaram , Loyola University Ana Valenzuela , San Francisco State University Debora Viana , University of Maryland, College Park Madhu Vishwanathan, University of Illinois at Urbana-Champaign Kathleen Vohs , University of British Columbia Lan Xia , Bentley College Manjit Yadav , Texas A&M University Gary Ye , Towson University Shi Zhang , University of California, Los Angeles

iii

Table of Contents and Conference Program List of Reviewers………………………………………………………………………………………………......ii Table of Contents and Conference Program………………………………………………………………………iv

THURSDAY, FEBRUARY 19th

REGISTRATION 5:00 pm - 7:30 pm

CONFERENCE RECEPTION & POSTER SESSION

Grand Ballroom 6:00 pm – 7:30 pm

Reception Sponsored by Earlbaum P.01. Anthropomorphized Electronic Recommendation Agents: Their Effects on Persuasion and Attriution of Responsibility…………………………………………………………………………………………………….. 1 Yang Dai, University of California, Los Angeles Praub David, The Ohio State University P.02. The Antecedents, Consequences and Moderators of Sterotype Activation Within and Advertising Context…………………………………………………………………………………………………………….2 Edith F. Davidson, University of Tennessee David W. Schumann, Universitty of Tennessee P.03. How Individuals Cope With Complexity In Online Environments: An Empirical Investigation Using A Goal Theoretic Framework………………………………………………………………………………………..3 Reetika Gupta, Baruch College Sucheta Nadkarni-University of Nebraska-Lincoln P.04 Consumer’s Perceptions of Sellers’ Quality Claims In online Auctions With Quality Uncertainty…….......5 Ying Lan, University of Mississippi Jianwei Hou, University of Mississippi P.05. Understanding Spanish-Speaking Citizens’ Experiences With Media and Advertising In a Non-Spanish-Speaking Geographic Area………………………………………………………………………………………..7 Annalisa Libermann, Miami University Sabrina M. Neeley, Miami University P.06. Ulterior Motives Vs. Irrational Suspicion: Flattery As a Cue In Sales Transactions....................................8 Kelley J. Main, University of British Columbia Darren W. Dahl, University of British Columbia Peter R. Darke, University of British Columbia P.07 The Impact of Regulatory Focus On How Consumers Replace Possessions………………………………..9 Kelley J. Main, University of British Columbia Lisa Sinclair, University of Winnipeg P.08. Nonverbal Decoding And Effective Customer Service.……………………………………………………10 Ereni Markos, Emerson College Nancy Upton (Puccinelli) Suffolk University

iv

P.09 “Will I Like It? Will I Use It?” :The Role of Desirability And Feasibility In Advance Buying Decisions………………………………………………………………………………………………………….11 Vanessa M. Patrick, University of Southern California Paul A. Jackson, University of Southern California P.10. Impulse Buying, Depletion and Strategies For Restraint………………………………………………….12 Aparna Rao, University of Michigan Zeynep Gurhan-Canli, University of Michigan P.11. The Effect of Brand Knowledt On Consumer Information Search………………………………………...13 Mototaka Sakashita, Kobe University, Japan P.12. The Influence Of Materialism And Indivicualism On Price Perception And Shopping Behaviors………..15 Kittichai Watchravesringkan, University of Arizona Jennifer Yurchisin University of Arizona P.13. How Perception Of Package Size Changes Influences Consumer Choice………………………………...17 Michael Wiles, Indiana University

FRIDAY, FEBRUARY 20th

REGISTRATION 8:00 am - 5:00 pm

CONTINENTAL BREAKFAST

7:00 am - 8:15 am

SESSION 1 Friday, 8:15 am - 9:30 am

1.1 Special Session: Consumer Behavior Goes to Wall Street: New Insights into the Psychology

of Investing Location: Union Square Chair: Michel Tuan Pham, Columbia University Moderator: Michal Strahilevitz, University of Arizona

1.1.1 Up Close and Personal: An Individual Level Analysis of the Disposition Effect…………………...……..18

Ravi Dhar, Yale University Ning Zhu, University of California, Davis

1.1.2 Promotion and Prevention Across Mental Accounts: Promotion and Prevention across Mental Accounts: When Financial Products Dictate Consumers’ Investment Goals………………………………..………………20

Rongrong Zhou, Hong Kong University of Sciences and Technology Michel Tuan Pham, Columbia University

1.1.3 Once-Burned-Twice-Shy, Counter-Factual Reasoning, and the Repurchasing of Stocks by Individual Investors………………………………………………………….……………………………………………….22

Terrance Odean, University of California, Berkeley Michal Strahilevitz, University of Arizona Brad Barber, University of California, Davis

v

1.2 Competitive Session: Branding Location: North Beach Chair: Mike Bastin, University College Northampton

1.2.1 Consumer Response to Performance-based Persuasion Attempts to Rebuild Weak Brands………………23

Julie A. Ruth, Rutgers University Anne York, University of North Carolina

1.2.2 The Effect of Brand Commitment on the Size of the Consideration Set........................................................24 Sekar Raju, State University of New York at Buffalo H. Rao Unnava. Ohio State University

1.2.3 Spillover of Negative Information on Brand Alliances…………………………………………………….26 Nicole L. Votolato, Ohio State University H. Rao Unnava, Ohio State University Sekar Raju, State University of New York at Buffalo

1.3 Competitive Session: Decision Making - I Location: Telegraph Hill B Chair: Susan Mantel, IUPUI 1.3.1 Forgetting and False Remembering: The Role of Imagery on Errors of Omission Versus Commission….28 Ann E. Schlosser, University of Washington

1.3.2 Thinking about Money: The Ultimate Equalizer?.........................................................................................30 On Amir, Yale University

Dan Ariely, Massachusetts Institute of Technology Ziv Carmon, INSEAD

1.3.3 Towards an Understanding of Benefit Depreciation: An Experimental Approach……………………..…32 Jennifer C. Siemens, University of Dayton Norma A. Mendoza, University of Arkansas

BREAK 9:30 am - 9:45 am

SESSION 2 Friday, 9:45 am - 11:00 am

2.1 Competitive Session: Inference Processes Location: North Beach Chair: Sabrina Neeley, Miami University of Ohio

vi

2.1.1 When Belief and Evidence Conflict: Factors that Mitigate Consumers’ Selective Use of Price-Quality Date……………………………………………………………………………………………………………….34

Tracy Meyer, University of Cincinnati James Kellaris, University of Cincinnati Frank Kardes, University of Cincinnati Maria Cronley, Miami University of Ohio Steven Posavac, University of Rochester

2.1.2 Positioning High-tech Hybrids: Factors that Facilitate and Inhibit the Use of Multiple (versus Single) Category Inference Strategy………………………………………………………………………………………35 Jennifer Gregan-Paxton, University of Delaware Steve Hoeffler, University of North Carolina Min Zhao, University of North Carolina

2.1.3 The Unbiasing Effect of Biasing Cues……………………………………………………………………..36 Wouter Vanhouche, University of Florida Stijn M.J. van Osselaer, Erasmus University 2.2 Special Session: The Influence of Goals on Consumer Choice Processes Location: Union Square Chairs: Steve Hoeffler, UNC-Chapel Hill Jesper Nielsen, University of Arizona 2.2.1 Effects of Nonconscious Goal Priming on Consumer Choice Behavior…………………………………..38

Baba Shiv, University of Iowa Joel Huber, Duke University Tanya Chartrand, Duke University

2.2.2 Nonconscious Priming of Decision Goals………………………………………………………………...40 Patricia W. Linville, Duke University Jesper Nielsen, University of Arizona Steve E. Hoeffler, UNC – Chapel Hill

2.2.3 Selecting versus Rejecting: Quest for Compatibility………………………………………………………41 Margaret G. Meloy, Penn State University J. Edward Russo, Cornell University Irwin Levin, University of Iowa

2.2.4 In Search of Consumer Goals……………………………………………………………………………..43

Kurt A. Carlson, Duke University 2.3 Competitive Session: Information Processing Location: Russian Hill Chair: Tina Kiesler, California State University, Northridge 2.3.1 Processing Abstract versus Concrete Brand Information: The Influence of Relationship Norms in Consumers Choice of Processing Strategies……………………………………………………………………44

Pankaj Aggarwal, University of Toronto Sharmistha Law, University of Toronto

vii

2.3.2 Classical Conditioning of Implicit Product Attitudes……………………………………………………..46

Bryan Gibson, Central Michigan University 2.3.3 The Influence of Need for Cognition on Temporal Construal Theory…………………………………….48 Kiwan Park, University of Michigan Joseph R. Priester, University of Michigan 2.4 Special Session: Emotion and Cognitive Processes in Older Adults: Consequences for

Consumer Information Processing and Decision Making Location: Telegraph Hill B

Chairs: Carolyn Yoon, University of Michigan Cathy Cole, University of Iowa Moderator: Cathy Cole, University of Iowa

2.4.1 The Moderating Impact of Aging on Responses to Emotional Advertising Appeals……………………………………………………………………………………………………………49

Patti Williams, University of Pennsylvania Aimee Drolet, UCLA

2.4.2 Trying to Discredit Information Can Make it Seem True to Older Adults…………………………………51 Ian Skurnik, University of Toronto Carolyn Yoon, University of Michigan Norbert Schwarz, University of Michigan

2.4.3 The Shrinking Decision Process of Older Consumers: Another Avenue to Repeat Purchase…………….53 Raphaëlle Lambert-Pandraud, Negocia Gilles Laurent, HEC

BREAK 11:00 am - 11:15 am

DISTINGUISHED SPEAKER

Professor Antoine Bechara University of Iowa

Title: “The bright and dark sides of emotion in decision-making:

Lessons from patients with broken brains”

Union Square Friday, 11:15 am - 12:15 pm

viii

LUNCHEON

SCP PRESIDENTIAL ADDRESS Professor Dipankar Chakravarti

University of Colorado

AWARD PRESENTATION SCP/Sheth Dissertation Award and Runners Up

SCP/Sheth Award for Early Contribution in Consumer Psychology JCP Young Contributor Award and Runners Up

Grand Ballroom

Friday, 12:15 pm - 1:45 pm

SESSION 3 Friday, 2:00 pm - 3:15 pm

3.1 Special Session: Extending our Understanding of Visceral Factors: Mechanisms,

Boundaries, and Context Sensitivity of the Impact of Hunger and Sensory Pleasures on Decision-making and Behavior

Location: Telegraph Hill B Chair: Laurette Dubé, McGill University 3.1.1 Regulatory Feelings and Their Relationships to Behavior: The Case of Hunger………………………..55

Laurette Dubé, McGill University Catherine Paquet, McGill University

3.1.2 Effects of Distractions while Tasting a Food Sample: The Interplay of Informational and Affective Components in Subsequent Choice……………………………………………………………………………….57

Baba Shiv, University of Iowa Stephen M. Nowlis, Arizona State University

3.1.3 On the Limitations of Pleasure Intensity in Shaping Behavior over Time………………………………...59 Jordan L. LeBel, Concordia University

3.2 Competitive Session: Comparative Advertising and Attribute Alignability Location: Russian Hill Chair: Selin A. Malkoc, University of North Carolina 3.2.1 Comparative Advertising: Effects of Attribute Typicality and Structural Alignability on Brand Evaluations and Similarity Judgment…………………………………………………………………………………………..61 Xiaoli Nan, University of Minnesota Sela Sar, University of Minnesota

3.2.2 When Two Brands Attack Each Other: A Comparative Advertising Examination………………………...62

Nicole L. Votolato, Ohio State University H. Rao Unnava, Ohio State University Payal Karumbiah, Ohio State University

ix

3.2.3 Purchasing Now, Or Later: Interactive Effect of Time and Attribute Alignability………………………..64 Selin A. Malkoc, University of North Carolina Canan Ulu, Duke University Gal Zauberman, University of North Carolina

3.3 Competitive Session: Pricing Location: Union Square Chair: Lan Xia, Bentley College 3.3.1 Switching between Pricing Schedules: A Dynamic Advantage for Flat Rates……………………………66 Tom Meyvis, New York University Jinhong Xie, University of Florida

3.3.2 Penny Wise and Pound Foolish: The Left Digit Effect in Price Encoding………………………………...68

Manoj Thomas, New York University Vicki Morwitz, New York University

3.3.3 Reference, Other Trust, and Perceived Price Fairness………………………………………………….....70 Lan Xia, Bentley College Kent Monroe, University of Illinois

3.4 Special Session: SCP-Sheth Dissertation Proposal Competition Location: North Beach Co-Chairs: Maria Cronley, Miami University of Ohio and Sabrina Neeley, Miami University of Ohio 3.4.1 Winner: In Need of a Favorable Conclusion: The Role of Goal-Biased Reasoning in Consumer Judgments and Evaluations…………………………………………………………………………………………………...72

Gustavo de Mello, University of Southern California Advisor: Professors Deborah MacInnis and David Stewart

3.4.2 Runner-Up: Unintended Purchase Opportunities: Conflict, Choices, and Consequences…………...…..74

Anirban Mukhopadhyay, Columbia University Advisor: Professor Gita Johar

3.4.3 Runner-Up: It's the Type of Thought that Counts: Consequences of Biased versus Objective Processing for Attitude Certainty……………………………………………………………………………………………..76

Derek Rucker, Ohio State University Advisor: Professor Richard Petty

BREAK 3:15 pm - 3:45 pm

SESSION 4 Friday, 3:45 pm - 5:00 pm

x

4.1 Special Session: Consumer Response to Unexpected or Counter-persuasive Information and Refutation Communications: New Directions

Location: North Beach Chair: Dobrina G. Teifel, Cornell University

4.1.1 Understanding Theory-Based Inference Making when Theories and Data Conflict: The Role of Theory Confidence………………………………………………………………………………………………………...77

Dobrina G. Teifel, Cornell University Douglas M. Stayman, Cornell University

4.1.2 A Meta-cognitive Analysis of Resistance to Persuasion and Attitude Certainty…………………………...78 Zakary L. Tormala, Indiana University Richard E. Petty, Ohio State University

4.1.3 Perceptual Similarity and Belief Change…………………………………………………………………..80

Gita V. Johar, Columbia University Anne Roggeveen, Babson College

4.2 Competitive Session: Post-Purchase Behavior and Attitudes Location: Telegraph Hill B Chair: Christine Page, Skidmore College 4.2.1 Effects of Anger, Regret, Distributive Justice and Social Comparison on Postpurchase Behaviors……..82

Carolyn Bonifield, University of Vermont Catherine A. Cole, University of Iowa

4.2.2 Experiencing Suspenseful Drama: The Sequencing of Emotions and Retrospective Judgments………….84 Robert Madrigal, University of Oregon Colleen Bee, University of Oregon Monica LaBarge, University of Oregon

4.2.3 The Impact of Social Satisfaction on Consumer Satisfaction Judgement -- A Telic/Autotelic Perspective of Consumer Satisfaction…………………………………………………………………………………………....87 Xiaojing Yang, Indiana University H. Shanker Krishnan, Indiana University

4.3 Special Session: Across Time and Space: The Influence of Regulatory Focus On Information Processing and Judgment

Location: Union Square Chairs: Jing (Alice) Wang, Northwestern University Angela Y. Lee, Northwestern University

4.3.1 Variations in Consumer Impatience……………………………………………………………………....89

Haipeng Chen, University of Miami Sharon Ng, University of Minnesota Akshay Rao, University of Minnesota

xi

4.3.2 What Will The Future Bring? Subjective Assessments of Future Probabilities Based on Prior Outcomes in Different Domains…………………………………………………………...........................................91 Ashwani Monga, University of Minnesota Akshay R. Rao, University of Minnesota

4.3.3 Effects of Regulatory Fit on Information Search and Persuasion………………………………………..93 Jing (Alice) Wang, Northwestern University Angela Y. Lee, Northwestern University

JOURNAL OF CONSUMER PSYCHOLOGY EDITORIAL BOARD MEETING

Russian Hill Friday, 5:00 pm - 6:00 pm

RECEPTION Grand Ballroom

Friday, 6:00 pm - 7:00 pm

SATURDAY, FEBRUARY 21st

REGISTRATION 8:00 am - 3:00 pm

CONTINENTAL BREAKFAST

7:00 am - 8:15 am

SOCIETY FOR CONSUMER PSYCHOLOGY EXECUTIVE BOARD MEETING

Russian Hill Saturday, 7:00 am – 8:15 am

SESSION 5 Saturday, 8:15 am - 9:30 am

5.1 Special Session: Embodied Cognition: Perceptual Processing in Online and Offline

Consumption Environments Location: Union Square Chair: Alan J. Malter, University of Arizona

5.1.1 Perceptual Symbols and Conceptual Representation of Products and Services………………………….94

Ling-Ling Wu, National Taiwan University Alan J. Malter, University of Arizona Lawrence W. Barsalou, Emory University

5.1.2 Body Boundary Aberration and Body Esteem as Influencers Of Consumer Attitudes and Intentions……96 Ellen Garbarino, Case Western Reserve University José Antonio Rosa, Case Western Reserve University

xii

5.1.3 Mental Simulation and the Effectiveness of Online Sales Aids for Experiential Products……………….98

Alan J. Malter, University of Arizona José Antonio Rosa, Case Western Reserve University

5.2 Competitive Session: Promotion Effectiveness Location: North Beach Chair: Priya Raghubir, University of California, Berkeley 5.2.1 The Impact of Coupon Value, Purchase Recipient, and Time of Redemption On What I Think They Think of Me: An Investigation of Metaperceptions In Coupon Redemption Situations…………………………………..100 Jennifer J. Argo, University of Alberta Kelley J. Main, University of British Columbia

5.2.2 Spatial Positioning: The Value of Center-Stage………………………………………………………….105

Ana Velenzuela, San Francisco State University Priya Raghubir, University of California, Berkeley

5.2.3 What's up Doc? DTC Pharmaceutical Advertising and Patient Drug Requests………………………...107 Rhonda W. Mack, College of Charleston Julia E. Blose, College of Charleston Rita A. Balaban, College of Charleston 5.3 Competitive Session: Decision Making - II

Location: Telegraph Hill A Chair: Maria Cronley, Miami University of Ohio 5.3.1 The Goal-Driven Shopper: The Influence of Goal Setting and Goal Completion on Consumer Spending………………………………………………………………………………………………………... 114 Leonard Lee, Massachusetts Institute of Technology Dan Ariely, Massachusetts Institute of Technology 5.3.2 Tis Better to Give Than to Receive: Preference Estimates Conditioned on Own and Others' Preferences…………………………………………………………………………………………………….. 116 Andrew D. Gershoff, Columbia University Susan M. Broniarczyk, University of Virginia

5.3.3 Shooting the Principal while Embracing the Agent: How Responsibility Acceptance Influences Principal and Agent Evaluations when Confronted with Bad News………………………………………………………118

Susan Hogan, Emory University

BREAK

9:30 am - 9:45 am

xiii

SESSION 6 Saturday, 9:45 am - 11:00 am

6.1 Special Session: New Theoretical Developments in Preference Construction Location: Telegraph Hill A Chair: Nathan Novemsky, Yale University 6.1.1 Preference Fluency and its Effect on No-choice, Compromise, and Attraction Effects…………………125

Nathan Novemsky, Yale University Ravi Dhar, Yale University Itamar Simonson, Stanford University Norbert Schwarz, University of Michigan

6.1.2 The Psychology of Matching and the Prominence Effect…………………………………………………127 Shane Frederick, Massachusetts Institute of Technology Eldar Shafir, Princeton University

The Distinction Effect: Mis-prediction and Mis-choice Due to Joint-Separate-Evaluation Inconsistency…………………………………………………………………………………………................128

Christopher K. Hsee, University of Chicago Jiao Zhang, University of Chicago

6.2 Competitive Session: Social Marketing Location: Union Square Chair: Mike Basil, University of Lethbridge 6.2.1 Implicit Attitudes toward Green Consumer Behavior………………………………………………….....130 Delphine Vantomme, Ghent University Maggie Geuens, Ghent University Jan De Houwer, Ghent University Patrick De Pelsmacker, Ghent University

6.2.2 Adolescents' Influence Tactics and Parental Yielding……………………………………………………132 Aviv Shoham, University of Haifa Gregory M. Rose, University of Washington - Tacoma Vassilis Dalakas, Berry College Maja Makovec Brenčič, University of Ljubljana 6.2.3 A Test of the Robustness of Cause-Related Marketing……………………………………………………133 Debra Z. Basil, University of Lethbridge Michael D. Basil, University of Lethbridge 6.3 Special Session: The Effect of Contextual Factors on Consumption Location: North Beach Chair: Nitika Garg, University of Pittsburgh

xiv

6.3.1 Feeling the Urge to Splurge: The Effect of Valence and Arousal of Consumer Mood States on Impulsive Consumption……………………………………………………………………………………………………..139

Alexander (Sasha) Fedorikhin, University of Southern California Vanessa Patrick, University of Southern California

6.3.2 The Influence of Incidental Affect on Consumers’ Food Intake…………………………………………..141 Nitika Garg, University of Pittsburgh Brian Wansink, University of Illinois at Urbana-Champaign Jeff Inman, University of Pittsburgh

6.3.3 Do Products Have Cooties? The Law of Contagion in Product Evaluations……………………………143 Andrea C. Morales, University of Southern California Gavan J. Fitzsimons, Duke University

6.4 Competitive Session: Advertising Location: Russian Hill Chair: David Luna, Baruch College 6.4.1 Excitation Transfer Theory and Advertising Effects Attributable to Ad Placement Order, Suspenseful Programming, and Ad Execution Style………………………………………………………………………….145

Colleen Bee, University of Oregon Robert Madrigal, University of Oregon

6.4.2 Fear Appeals in Traffic Safety Advertising - The Moderating Role of Context, Advice, Duality and Anxiety…………………………………………………………………………………………………………..147 Wim Janssens, University of Antwerp Patrick De Pelsmacker, University of Antwerp Liesbeth Bulkmans, University of Antwerp

6.4.3 The Impact of Context Induced Feelings on Effectiveness of Ad Appeals Evoking the Same Feeling or a Different feeling of the Same Valence…………………………………………………………………………...149

Tine Faseur, Ghent University

BREAK

11:00 am - 11:15 am

DISTINGUISHED SPEAKER

Professor Arthur B. Markman University of Texas

Title: “The Dynamics of Preference and the Structure of

Consumers' Goals”

Union Square Saturday, 11:15 am - 12:15 pm

xv

LUNCHEON BUSINESS MEETING

AWARD PRESENTATION 2004 SCP Fellows Awards

SCP/Sheth Award for Scientific Achievement in Consumer Psychology

SCP SCIENTIFIC ACHIEVEMENT AWARD WINNER ADDRESS Professor John Lynch

Duke University

Grand Ballroom Saturday, 12:15 pm - 1:45 pm

SESSION 7 Saturday, 2:00 pm - 3:15 pm

7.1 Competitive Session: Persuasion Processes Location: Telegraph Hill A Chair: Nidhi Agarwal, New York University 7.1.1 Should it be About Them or Me? Emotional Underpinnings of the Persuasiveness of Health Messages .151

Nidhi Agarwal, New York University Jennifer Aaker, Stanford University Geeta Menon, New York University

71.2 Public Commitment: A Key Motivator in Long-Term Customer Compliance…………………………….153

Stephanie Dellande, Chapman University Prashanth Nyer, Chapman University

7.1.3 Nurturance, Religiosity and Appearances: The Effect of Personality Traits on Responses to Corporate Philanthropy……………………………………………………………………………………………………..155 Debra Z. Basil, University of Lethbridge Michael D. Basil, University of Lethbridge Deanne Weber, Porter Novelli International

7.2 Competitive Session: Socio-Cultural Influences Location: Union Square Chair: Robin Coulter, University of Connecticut 7.2.2 Renegotiating Who I Am: The Effects of Changing Cultural Ideologies and Interpersonal Influence on the Self…………………………………………………………………………………………………………........157

Camelia Micu, University of Connecticut Robin A. Coulter, University of Connecticut Linda L. Price, University of Nebraska

xvi

7.2.2 Personal Values for Customer Service……………………………………………………………………159 Nicole Hafey, Central Queensland University Tony Ward, Central Queensland University 7.2.3 Psychological Antecedents of Impulsive and Compulsive Buying: A Hierarchical Perspective…………168 Tao Sun, Emerson College Guohua Mark Wu, San Jose State University Seounmi Youn, Emerson College 7.3 Special Session: Guilty Pleasures and Guilty Sins: How They Affect Consumer Behavior

and Misbehavior Location: North Beach Chairs: Michal Strahilevitz, University of Arizona Sooyeon Nikki Lee, New York University 7.3.1 Take Aversion vs. The Joy of Giving: A Case of Guilt-Based Asymmetry……………………………….175

Michal Strahilevitz, University of Arizona

7.3.2 A Little Something for Me, and Maybe for You Too: Promotions that Relieve Guilt…………………….177 Sooyeon Nikki Lee, New York University Priya Raghubir, New York University Kim P. Corfman, New York University

7.3.3 Form vs. Function: Emotional and Behavioral Consequences of Making Hedonic vs. Functional Tradeoffs………………………………………………………………………………………………………...178

Ravi Chitturi, University of Texas at Austin Raj Raghunathan, University of Texas at Austin Vijay Mahajan, University of Texas at Austin

BREAK

3:15 pm - 3:45 pm

SESSION 8 Saturday, 3:45 pm – 5:00 pm

8.1 Special Session: Product Assortment and Variety in Consumer Decision-Making:

Individual and Situational Factors Location: Telegraph Hill A Chairs: S. (Ratti) Ratneshwar, University of Missouri – Columbia Susan Broniarczyk, University of Texas at Austin and University of Virginia (visiting) Moderators: Susan Broniarczyk, University of Texas at Austin and University of Virginia (visiting) 8.1.1 Retail Assortment Variety: How Focus on “Assortment” versus the “Item” Can Affect Justification Strategies………………………………………………………………………………………………………...179

Barbara E. Kahn, University of Pennsylvania Mary Frances Luce, University of Pennsylvania Seshan Ramaswami, Singapore Management University

xvii

8.1.2 Projecting One’s Own Variety-Seeking Trait on to Decisions Made For Others………………………...180

Tilottama G. Chowdhury, University of Connecticut S. Ratneshwar, University of Missouri – Columbia Kalpesh K. Desai, State University of New York, Buffalo

8.1.3 Product Assortment and Individual Decision Processes…………………………………………………182 Alexander Chernev, Northwestern University

8.2 Competitive Session: Self Construal and Decision Making Location: Union Square Chair: Sandor Czellar, University of Lausanne

8.2.1 Self-monitoring and Attitude Accessibility Revisited……………………………………………………..183

Sandor Czellar, University of Lausanne

8.2.2 That's My Own Creation! Experiencing Attachment to Self-Altered Products…………………………...185 Ruth Mugge, Delft University of Technology, Netherlands Hendrik N.J. Schifferstein, Delft University of Technology, Netherlands Jan P.L. Schoormans, Delft University of Technology, Netherlands 8.2.3 Self Esteem and Social Influence: The Moderating Role of Gender……………………………………...187 Mehdi Mourali, Concordia University Michel Laroche, Concordia University Frank Pons, University of San Diego 8.3 Special Session: Online and Speed Dating: Experiments, Field Data, and Subjective Recollections Location: North Beach Chair: Dan Ariely, Massachusetts Institute of Technology 8.3.1 Revealed Preference Determinants of Mate Selection: Evidence from an Experimental Dating Market..189

Raymond Fisman, Columbia University Sheena Sethi-Iyengar, Columbia University Itamar Simonson, Stanford University

8.3.2 Dating Preference and Dating Decisions: The Effect of Repeated Choices on Consumer Preferences…190 Sheena Sethi-Iyengar, Columbia University Itamar Simonson, Stanford University Raymond Fisman, Columbia University

8.3.3 Learning and Juggling in Online Dating…………………………………………………………………192 Jeana Frost, Massachusetts Institute of Technology Dan Ariely, Massachusetts Institute of Technology

xviii

Paper Presented At The 2003 SCP Summer Conference Identifying Prospects and Reaching Targets: Neglected Distinction within Marketing………………………...193

Geraldine Fennell, Consultant, Columbus, OH Joel Saegert, The University of Texas at San Antonio

List of Contributors (Indexed by session)..……………………………………………………………………...197

1

P.01. ANTHROPOMORPHIZED ELECTRONIC RECOMMENDATION AGENTS: THEIR EFFECTS ON PERSUASION AND ATTRIUTION OF RESPONSIBILITY

Yang Dai, University of California, Los Angeles

Prabu David, The Ohio State University

ABSTRACT

The main purpose of the study is to shed light on the effects of the anthropomorphized recommendation agent on the two core processes of online consumer behavior: persuasion and attribution of responsibility.

Anthropomorphism is manipulated via the presence versus absence of social cues (e.g, thumbnail photo and personal nouns) in the electronic recommendation agent’s interface. Results of our experiment indicate that the addition of social cues augmented the inclination of consumers to claim more credits for themselves in the face of positive outcomes; the general pattern of results also suggests that the addition of social cues compels the consumers to blame the agent more harshly in the face of negative outcomes. This is of interesting implication for the design of effective electronic agents, i.e., integrating more anthropomorphizing features into the electronic agent may backfire by triggering stronger self-serving bias in the consumers.

On the basis of extant literature on social presence and nonverbal communication, we expected to find the anthropomorphized agent to be more persuasive. However, our data suggests the opposite trend: The addition of social cues did not enhance persuasiveness but rather tend to undermine persuasion. Further research can be conducted to identify conditions under which anthropomorphism is detrimental or beneficial for persuasion.

For further information contact: Yang Dai, 110 Westwood Plaza, Suite B401, Los Angeles, CA 90066. Email: ydai@anderson.ucla.edu

2

P.02. THE ANTECEDENTS, CONSEQUENCES AND MODERATORS OF STEREOTYPE ACTIVATION WITHIN AND ADVERTISING CONTEXT

Edith F. Davidson, University of Tennessee

David W. Schumann, University of Tennessee

ABSTRACT

The impact of stereotypes has been examined extensively in the social psychology literature. This research reveals that stereotypes may be activated upon the processing of a stereotype cue (e.g., descriptive words). One stimulus rich environment where the triggering of stereotype activation appears likely is within the practice of advertising. The topic of employing minority actors in advertising has drawn attention from both advertising and consumer behavior researchers (e.g., Elliott 1995, Greenberg 1988). There is empirical support that suggests when minority actors are included in the advertisements, minority consumers are better able to recall the advertising content and have more positive affect toward the advertisement and the actors (Qualls and Moore 1999; Williams, Qualls and Grier 1995) and that minority and majority viewers differ in their interpretation of ad meaning when the advertisements are targeted towards the minority group (Grier and Brumbaugh 1999). This paper synthesizes the social psychology, advertising and consumer behavior literatures and proposes a theoretical model reflecting what are believed to be key antecedents and consequences of stereotype activation that may be found within an advertising context. For further information contact: Edith F. Davidson, Email: edavidso@utk.edu

3

. P.03. HOW INDIVIDUALS COPE WITH COMPLEXITY IN ONLINE ENVIRONMENTS: AN

EMPIRICAL INVESTIGATION USING A GOAL THEORETIC FRAMEWORK

Reetika Gupta, Baruch College Sucheta Nadkarni – University of Nebraska - Lincoln

ABSTRACT

Introduction With the growth of the Internet, businesses are spending billions of dollars to add a wide range of features such as animated pictures, sophisticated icons and links at a web site. However, marketers know very little about the effects of this wide variety of features at a web site (i.e. perceived web site complexity), on individuals. The notion of perceived web site complexity is important because complexity affects not only the communication effectiveness of a web site but also the attitudes towards the web site and purchase intentions.

Despite the increasing importance of

perceived website complexity, substantive research employing this construct has been constrained by two major factors. First, most empirical studies have defined complexity as a uni-dimensional construct focusing only on the structural features (e.g. length of text, background colors and animation) of individual web pages (Stevenson, Bruner and Kumar, 2000; Geissler, Zinkhan and Watson, 2001). Although this conceptualization of complexity maybe appropriate for an individual web page, it may not fully capture the complexity of a website. It ignores the inherent interactive dimension of a website that distinguishes it from other media (Hoffman and Novak, 1996). In contrast to static representations of print advertising or broadcast television, the process of navigating websites is dynamic and interactive wherein users need to continuously interact with and adapt to the information cues at the website. Complexity of a website arises from not only the physical arrangement of pictures and text, but also from the arrangement of pages configuring a website, and the navigational ease in traversing the web pages. A broad and multi-dimensional conceptualization of complexity is required to capture the structural, interactive and dynamic aspects of a website. Second, the absence of a comprehensive conceptualization capturing the dimensions underlying website complexity has seriously impeded research on mechanisms that individuals employ to cope with complexity in online environments.

Given these concerns and potential, this paper attempts to: 1] develop a broad and multi-dimensional conceptualization of the construct, perceived website complexity and 2] outline the mechanisms that individuals use to cope with complexity. To this end, we conducted two empirical studies.

Study 1

In Study 1, we propose that there are three underlying dimensions of perceived web site complexity – structural, coordinative and interactive. Structural complexity at a web site represents the degree to which users perceive the images, graphics, text, and information categories dissimilar. Coordinative complexity refers to the range of different information clusters that can be manipulated at a web site and the degree of inter-relationship between the different information clusters. Finally, interactive complexity is reflected in the level of sophistication of the hypertextual format (connecting different web pages) that individuals use to navigate the web site. Study 1 was divided into two phases: discovery and testing. The purpose of the discovery phase was to develop and perform preliminary testing of measurement items to be used in the testing phase of the study. Through a series of steps, 16 measures of perceived web site complexity, drawn from existing theoretical measures and free association task were retained in the discovery phase. In the testing phase, data was collected from 332 undergraduate students (57% male; 67% belonging to the 19-24 age group) across a wide array of websites to validate the model of website complexity. Confirmatory factor analysis was used in the testing phase to confirm the convergent, discriminant, and nomological validity of the three dimensions of complexity. Study 2 In Study 2, we use the theory of implementation intentions (Gollwitzer, 1999; Gollwitzer and Schaal, 1998); to theorize that in online goal striving situations (e.g. online information search), individuals cope with complexity using one of the two mechanisms: task facilitating and complexity resolving. More specifically, we propose that when individuals

4

encounter complexity in their search activity, they can either bypass the complexity, and focus on the search task by continuing to navigate (task facilitating strategy) or develop systematic execution strategies through the use of support tools to resolve the complexity (complexity resolving strategy). We argue that these strategies will lead to differences in performance expectations. Complexity resolving strategy involves a stronger execution orientation facilitated by the use of support tools, in comparison to the task facilitation strategy. This stronger execution focus leads to higher performance expectations by invoking a higher optimism in attaining the search goal. The experiment for Study 2 includes two basic conditions: task facilitating and complexity resolving. Subjects are given a specific search task where they encounter complexity (in the form of a misleading hyperlink). Subjects in the complexity resolving condition are provided support tools to retrieve the search results while no support tools are available to the subjects in the task facilitating condition. On task completion, performance expectation scores are elicited from the subjects. Conclusion

The current study fills a key gap in existing theory on complexity by uncovering interactive facets unique to online environments, not captured by prior research. The comprehensive conceptualization developed in this paper, is especially useful for investigating the unique ensuing mechanisms that occur when individuals encounter complexity in an online environment. Study 2 demonstrates that the nature of coping mechanisms (task facilitating vs. complexity resolving) that individuals use, leads to differences in performance expectations of online goal directed tasks. Resolving complexity, in an online environment entails not only planning strategies to overcome the complexity, but it also involves implementing those strategies (Chandran and Morwitz, 2003). The availability of support tools in an online environment facilitates this joint process of planning and action; leading to higher performance expectations.

Our theorization is based on a goal

directed search representing an involving situation, where individuals are committed to goal attainment. Future research can investigate if the proposed effect of the coping mechanisms outlined in this study extends to non involving situations. From a practical perspective, our study highlights the role of support tools in helping individuals to

systematically overcome complexity. This implies that web marketers should direct their attention to develop effective support tools that aid individuals in the navigation process.

References available on request

For further information contact: Reetika Gupta, Zicklin School of Business, Marketing Department, Baruch College, 1, Bernard Baruch Way, Box: B12-240, NY, NY 10010 Phone: 646.312.3309 Fax: 646.312.3271 Email: reetika_gupta@baruch.cuny.edu

5

P.04. CONSUMERS’ PERCEPTIONS OF SELLERS’ QUALITY CLAIMS IN ONLINE AUCTIONS

WITH QUALITY UNCERTAINTY

Ying Lan, University of Mississippi Jianwei Hou, University of Mississippi

ABSTRACT Introduction

Quality uncertainty exists between transaction parties in various markets. To solve the problem of quality uncertainty, researchers have proposed signaling theory (e.g., Spence 1974). The signaling theory assumes that there are two types of sellers, sellers of high quality products and sellers of low quality products. Most often, the sellers of high quality products are willing to accurately inform buyers about the quality of their products so as to be compensated by offering such products, whereas the sellers of low quality products tend to hide or misinform buyers about the low quality of their products so as to avoid the loss by offering such products. A variety of quality signals has been identified in the literature, including brand names (e.g., Akerlof 1970), price (e.g., Milgrom and Roberts 1986), advertising (e.g., Milgrom and Roberts 1986), and warranties (e.g., Boulding and Kirmani 1993). Research Hypotheses

Though there are various types of signals in the marketplace, the credibility of those signals varies as perceived by buyers. In some online auctions (e.g., coins and sports trading cards), the quality of items can be determined by the specific grade placed by sellers or third parties (i.e., professional grading companies). If sellers assign a specific grade to their offerings, they attempt to claim the quality level of their items in order to reduce quality uncertainty. Whether such claim is viewed as a quality signal is determined by its credibility.

We suspect that such quality claim has a low credibility thus consumers would not perceive it as an indication of quality level. In additional, we also expect that a higher claimed quality is less trustful than a lower claimed quality. Therefore, the present study tests following hypotheses under quality uncertainty:

1H : Auctions with a quality claim will not (1) increase their likelihood of a sale and (2) attract more bidders than those without a quality claim.

2H : Auctions with a higher claimed quality (1) are less likely to result in a sale and (2) will attract

fewer bidders than those with a lower claimed quality.

3H : Items with a higher claimed quality will sell at lower prices relative to their market prices than those with a lower claimed quality. Methodology

eBay was chosen as the data source because of its popularity. Data were manually collected from late-July 2003 to mid-August 2003 and the final data set consisted of 926 auctions of uncertified 1921 Morgan Dollar. Among 926 auctions, there were 522 auctions with a quality claim at various grades and 404 auctions without a quality claim. The data set also included 666 successful auctions and 260 auctions that receive zero bids. In addition, we also collected data on the market price for each coin grade from PCGS price guide (www. pcgs.com) and Numismedia Price Guide (www.numismedia.com). Results

We test our research hypotheses using two logistic regression models and three OLS regression models. Results indicate that a seller’s quality claim has a significantly negative effect on both the auction’s likelihood of a sale ( 01.0,317.1 <−= pβ ) and the number of bids

( 01.0,091.0 <−= pβ ). Thus, 1H is supported. The standardized coefficients for coin grades generally become smaller and less significant as the grade rises, suggesting that coins with a higher claimed grade are less likely to sell than those with a lower claimed grade, whereas the level of the claimed quality does not have a clear effect on the number of bids. Therefore, 2H is partially supported. A regression analysis also indicates that a lower claimed quality has a stronger positive effect on the final price than a higher claimed quality thereby 3H is supported. Conclusions

Quality uncertainty exists in many online auctions. In order to reduce such uncertainty, some sellers attempt to make a specific quality claim on their items. We investigate how such claim would

6

influence consumers’ bidding behaviors. Specifically, we find that online bidders do not perceive a quality claim as an indication of quality level. First, a quality claim decreases the likelihood of a sale. Second, items with a quality claim do not attract more bidders than those without a quality claim. Among auctions with a quality claim, those with a higher claimed quality reduce the odds of a sale and attract fewer bidders than those with a lower claimed quality. This result is consistent with Akerlof’s (1970) suggestion that low quality products tend to drive out high quality products in a marketplace with quality uncertainty. We also find that items with a lower claimed quality are likely to sell at a higher price relative to their market values than those with a higher claimed quality. For further information contact: Jianwei Hou, Department of Marketing, University of Mississippi, University, MS 38677. Email: jwhou@olemiss.edu.

7

P.05. UNDERSTANDING SPANISH-SPEAKING CITIZENS’ EXPERIENCES WITH MEDIA AND ADVERTISING IN A NON-SPANISH-SPEAKING GEOGRAPHIC AREA

Annalisa Libermann, Miami University Sabrina M. Neeley, Miami University

ABSTRACT

The “Latinization” of the United States

was much like the breaking of a piñata. A culture filled with excitement, mystery, and passion was kept all wrapped up but ready to break free. When the right time struck, the culture-filled piñata exploded onto mainstream America. Latin influences were sprinkled all across American pop culture in the shape of mariachi bands, a talking Chihuahua, and Jennifer Lopez, to name a few. It is not only American enthusiasm for the Latino culture that has caused this phenomenon, but much is owed to the growing population of Latinos/Latinas in the United States. As this rapidly growing population continues to influence the current pop culture, its increasing spending power is causing marketers to take notice.

Companies such as Anheuser-Bush, Procter and Gamble, Sears and McDonalds have spent millions of dollars to better their understanding of the growing Latino population. Marketers everywhere continue to explore the behaviors, thoughts and actions of this diverse segment of the population. Unfortunately, the challenges faced in this communication battle are two-sided. While marketers struggle to understand this segment of our community, Latinos seek fulfillment in a society, which at times ignores their needs and challenges in the marketplace. This problem is particularly prevalent in geographic areas that traditionally do not have Spanish-speaking populations.

This research examines the experiences of Latino/Latina immigrants’ use of different media sources and advertising, in a geographic area without a concentration of Spanish speakers or local Spanish language media. First, it examines the experiences that these immigrants have with media and advertising in general. In particular, it looks at potential language and comprehension barriers, as well as interpretation and information challenges. Secondly, the research studies which type of media are most effective in distributing information to the Spanish-speaking consumer. For further information contact: Sabrina M. Neeley, Department of Marketing, Richard T. Farmer School of Business, 200 Upham

Hall, Miami University, Oxford, OH 45056. (513) 529-1200, (513) 529-1290 (fax), Email: neeleysm@muohio.edu

8

P.06. ULTERIOR MOTIVES VS. IRRATIONAL SUSPICION: FLATTERY AS A CUE IN SALES TRANSACTIONS

Kelley J. Main, University of British Columbia

Darren W. Dahl, University of British Columbia Peter R. Darke, University of British Columbia

ABSTRACT The aim of the current research is to examine whether consumers have positive or negative reactions to flattery received in a sales encounter. This research draws on the persuasion literature in order to examine the factors that influence consumer reactions to flattery.

The results of study 1 demonstrated that consumers who were flattered prior to purchase were more likely to perceive a salesclerk as less trustworthy, to report suspicious thoughts and to report perceptions of manipulative intent than consumers who were flattered after their purchase. These results are consistent with expectations and previous research on flattery in a marketing context (Campbell and Kirmani 2000). Results seem to suggest that flattery is a cue that triggers suspicion and a search for ulterior motives. Flattery elevated levels of suspicion regarding the clerk’s motives that had negative consequences for perceived trustworthiness of the clerk.

The second study, a field study, demonstrated that consumers make less of a distinction between flattery before purchase and flattery after purchase in a real shopping experience as compared to an imagined shopping experience. This distinction of when the flattery occurred became subtler and consumers responded more based on the presence or absence of flattery as opposed to its timing. It was flattery that became a salient cue in the situation to consumers, and their reactions to that flattery were primarily negative in response. Results demonstrated that consumers who were flattered were more suspicious and perceived that the clerk’s behavior was guided by an ulterior motive. This, in turn, led consumers who were flattered to perceive the clerk as less trustworthy, to be less satisfied with their purchase, and to have more negative WOM intentions than consumers who were not flattered.

The current research makes a number of contributions to the research literature. The first is to examine flattery within the context of research on accuracy and defense motives thereby extending previous research. The second contribution is the demonstration of target-observer differences in response to flattery within this theoretical

framework and in a field setting. Third, the current research demonstrates that there are additional consequences to flattery beyond the suspicion of ulterior motives such as lower satisfaction and negative WOM intentions.

For further information contact: Kelley J. Main, Marketing Division, Sauder School of Business, University of British Columbia, 2053 Main Mall, Vancouver, British Columbia, Canada, V6T 1Z2, Email: kelley.main@saider.ubc.ca

9

P.07. THE IMPACT OF REGULATORY FOCUS ON HOW CONSUMERS REPLACE POSSESSIONS

Kelley J. Main, University of British Columbia Lisa Sinclair. University of Winnipeg

ABSTRACT

The current research is interested in determining the choices that individual’s make in replacing possessions that they have had taken from them. Results of the first study demonstrated that when deciding to replace possessions, consumers preferred to purchase a stereo with different features than the one that they previously owned. Given that consumers seem to desire a change in the features of the stereo, one wonders if it may be possible to alter this preference for change. The goal of the second study was to determine if consumer preferences for change in replacing possessions (i.e. changing the features) can be reversed.

The second study illustrates a reversal of findings from the first study. Results suggested that when participants are primed to consider replacing their possessions in a promotion focus, consumers prefer to change the features of the DVD player or stereo that they are replacing. However, for consumers who are primed with a preventions focus, the preference to upgrade features is reversed and consumers do not exhibit a preference to keep the features of the stolen item the same.

The results of this research have suggested that an individual’s regulatory focus influences the choices that they make in replacing possessions that they have had taken from them. The first study demonstrated that consumers who consider replacing possessions that they have lost prefer to change the features of the item. The second study demonstrates this effect can be reversed; when consumers were primed with a prevention focus they were less likely to want to change the options of the product than consumers who were promotion focused.

In terms of future research, it has been suggested in work on irreplaceable possessions that if an item is replaced, the replacement may not be as valued as the original (Grayson & Shulman 2000). In the current research, our focus was on the choices made in replacing a possession. However, a question remains as to the value associated with the replacement as compared to the original. If consumers stay with the same features and replace the item that was stolen or lost, how do they feel about the replacement item? In contrast, if consumers changed the feature or options that were

available to them when they decided to replace the item, how do they value the replacement? For further information contact: Kelley J. Main, Marketing Division, Sauder School of Business, University of British Columbia, 2053 Main Mall, Vancouver, British Columbia, Canada, V6T 1Z1. Email: kelley.main@saider.ubc.ca

10

P.08. NONVERBAL DECODING AND EFFECTIVE CUSTOMER SERVICE

Ereni Markos Emerson College Nancy Upton (Puccinelli) Suffolk University

ABSTRACT

It has long been thought that a

positive attitude among service employees reigns supreme in effective delivery of customer service. The present investigation counters this belief and suggests that an employee’s ability to detect nonverbal cues is as important, if not more so, as a positive attitude. The service provider’s ability to recognize and monitor the customers’ emotions is seemingly important in creating a positive customer experience as well as a good first impression. The focus of the study is employee expressiveness and decoding skills as determinants of a positive customer experience. A mid-sized restaurant served as the research site along with its employees and customers. Surveys were administered to customers so as to capture opinions regarding employee and restaurant service. A 90% response rate was reported. Measures of employee effectiveness and positive attitude were included in the survey and used for statistical analysis.13 employees, both male and female, with a mean age of 32, completed a test measuring nonverbal expressiveness, ACT, and a test measuring nonverbal sensitivity, PONS. Results from both employee tests were weighted, separately, against the “effective server” rating and “positive attitude” rating from the survey.

Furthermore, the study examined the following two hypotheses: H1: Customers are more satisfied with employees characterized by greater ability to detect non-verbal cues. H2: Customers are not more satisfied with employees characterized by greater positive expressivity.

Consistent with the hypotheses, the study found a strong relationship between an employee’s ability to detect nonverbal body cues and the customer’s perception of the employee’s effectiveness as a server. It is suggested here that the ability to pick-up on nonverbal cues may be critical in understanding customers’ expectations and creating positive service experiences. Specifically, the results indicate a strong positive relationship between an employee’s

Body PONS score and his/her rating as an effective server (r =. 563, p = .057). Interestingly, this effect did not hold for the relationship between an employee’s Face PONS score and Server Effectiveness (r = .131, p =. 36). This finding is a noteworthy one as it is congruent with other studies on nonverbal communication. Research suggests that the face is the most controllable nonverbal channel that as such reveals censored messages, whereas, the body exposes involuntarily emotions that yields more honest meaning (Rosenthal & Snodgrass, 1985). Thus, an employee who is good at reading body cues may actually be better equipped to identify what a customer is truly thinking or feeling as opposed to what the customer intends to communicate (e.g. the employee may be able to detect from the customer’s tone of voice in “no fries”, that the customer would indeed like some). In summary, the data put forward suggests that nonverbal sensitivity may enable a more positive service experience for customers as employees are better able to respond to customer needs. These findings illuminate the importance of nonverbal sensitivity not only in service settings but also in organizational settings, as a whole, as it facilitates more effective communication. For further information contact: Ereni Markos, 71 Richardson Rd., Lynn MA. 01904

11

P.09. “WILL I LIKE IT? WILL I USE IT?”:THE ROLE OF DESIRABILITY AND FEASIBILITY IN ADVANCE BUYING DECISIONS

Vanessa M. Patrick, University of Southern California

Paul A. Jackson, University of Southern California

ABSTRACT With the advent of new technologies and infrastructure, the phenomenon of advance selling, which occurs when sellers allow buyers to purchase at a time preceding consumption, has gained considerable attention from marketers (Shugan and Xie, 2000; Xie and Shugan, 2001). This paper looks at the flip-side of this phenomenon – advance buying. Advance buying is defined as a consumer’s commitment, in the form of payment, for the right to consume a product or service in the future. In this paper we aim to develop an understanding of the phenomenon of advance buying and propose a conceptual framework that seeks to explain how consumers arrive at a decision to advance buy. Extending temporal construal theory (TCT; Liberman and Trope, 1998), our model aims to understand the role of contextual and transactional factors in driving advance buying decisions. The conceptual framework proposes the influence of the nature of the product (hedonic vs. utilitarian), the length of time between purchase and consumption (short vs. long) and the number of consumption episodes (single vs. multiple) on the intention to advance buy. Moreover, the proportion of desirability to feasibility concerns is proposed to mediate this influence. Finally, a series of studies are outlined that aim to test these propositions. Results from an initial study that manipulates the length of time between purchase and possible consumption and the number of consumption episodes that can occur, will be presented. For further information contact: Vanessa M. Patrick, Department of Marketing, Marshall School of Business, University of Southern California, Accounting Building, 3rd Floor, Los Angeles, CA 90089-0443. Phone: 213 748 0697, Fax: 213 740 7828, Email: vanessa.patrick@marshall.usc.edu

12

P.10. IMPULSE BUYING, DEPLETION AND STRATEGIES FOR RESTRAINT

Aparna Rao, University of Michigan Zeynep Gurhan-Canli, University of Michigan

ABSTRACT This research is an attempt to explain impulse buying taking into account its emotional as well as cognitive contents. The present idea tries to understand the relationship between impulse buying, self-control failure, depletion and the various strategies for restraint. Almost all attempts in explaining impulse buying have involved exploring the relationship between impulsiveness and self-control and regarded impulse buying as a form of self–control failure (Baumeister, 2002). Self-control involves overriding or inhibiting competing urges, behaviors or desires (Barkley, 1997; Baumeister, Heatherton, & Tice, 1994; Shallice & Burgess, 1993).

The ‘self-regulatory depletion’ concept tends to view self-control as a limited but constant capacity and not as a skill (Muraven, Tice and Baumeister, 1998). The core idea of depletion is that individuals have a limited capacity for self control and prior exertion depletes this strength. In addition, self-control is defined as a psychological conflict between desire and willpower (Hoch & Loewenstein, 1991). Two distinct categories of self-control strategies are developed and employed by individuals to exercise self-control.

Desire reduction strategies include tactics like avoidance, postponement, distraction, and substitution. They are aimed at reducing desire. Alternatively, willpower strategies are aimed at overpowering rather than reducing desire. They include all attempts by consumers to enumerate and make salient the costs of satisfying time-inconsistent preferences. They include pre-commitment, time binding, economic cost assessment, bundling of costs, invoking higher authority and anticipating future guilt or regret.

We conducted an initial pretest to determine the most frequently used strategies from both categories. Next, we designed four studies to test our main hypothesis on the relationship between impulse buying and depletion and the effectiveness of the short-listed strategies from both categories under experimentally manipulated conditions of high and low depletion.

The first study is a main effects study aimed at demonstrating increased impulse buying under higher levels of depletion. The next two studies have been designed to study the different

self control strategies. They involve the following three steps: priming the subjects with the strategy being tested, followed by the depletion manipulation, and then asking them to browse through an online shopping website where they can purchase low cost items if they wish to. The fourth study is an extensive and qualitative survey on recent impulse purchases and attempts at resistance by subjects.

We expect to show increased impulse buying under higher depletion. Also, we expect that the depleted group will indulge in more purchasing measured in ‘number of items’ as well as their ‘worth in dollars’ when they are using willpower based strategies as compared to the non-depleted group. In contrast, we expect that both the depleted and non-depleted groups will indulge in similar patterns of purchasing when they use desire reduction strategies.

In the survey or the fourth study, we expect that individuals who report resisting their impulses though highly depleted, will report lesser use of willpower based strategies but more use of desire reduction strategies. In contrast, individuals who resisted their impulses and were not highly depleted should report more or less equal usage of willpower based strategies as well as desire reduction strategies. The present research is expected to provide some of the reasons behind why individuals indulge in impulse buying and why they fail to use their self-control strategies at times. By providing an explanation for self-control failure that goes beyond dispositional factors and explores situational and other factors, we will be providing important insights on consumer self-control. Consumers can benefit by identifying the right strategy for the right occasion depending on their depletion level and the purchase situation. For further information contact: Aparna Rao, PhD Student in Marketing, D2259, University of Michigan Business School, 701 Tappan Street, Ann Arbor, MI 48109, Phone: 734-647-7667, Email: aparnar@umich.edu

13

P.11. THE EFFECT OF BRAND KNOWLEDGE ON CONSUMER INFORMATION SEARCH

Mototaka Sakashita, Kobe University, Japan

ABSTRACT Overview

Technological advancement has dramatically changed the way consumers search for information when making product evaluations. However, it has also created a problem; with it, they have to process much more information to make decisions. Faced with this difficulty, consumers may utilize the knowledge stored in their memory, which typically leads to applying many choice heuristics. In particular, brand knowledge serves as an information chunk that organizes the various kinds of information stored, thus supporting their information processing.

The focus of this paper is to investigate the role of brand knowledge in guiding consumers to a more efficient information search. Specifically, it investigates how the utilization of brand knowledge affects task effect in the phase of consumers’ product evaluation, especially their information search. It is hypothesized that the utilization of brand knowledge occurs when consumers are facing some well-known brand names and they hold commitment toward those brands. Because of their limited capacity for information processing, consumers tend to acquire a smaller ratio of information through external search when they are facing more alternatives. This task effect, however, is altered when consumers utilize their brand knowledge; consumers with high level of brand commitment (hereafter referred as BC) search for less external information when they are exposed to well-known brand names, while those with low level of BC search for relatively more. When not exposed to any brand names, however, consumers tend to search for less external information regardless of the level of BC when task complexity increases. This is because specific brand associations in memory substitute for the information needed to evaluate alternatives. Brand Commitment

BC refers to preference or purchase intention toward a particular brand, and measures how strongly consumers hold feelings toward a particular brand in a product category. It is measured based on three questions: (1) There is a favorite brand in this product category, (2) There is a particular brand I would purchase in this product category, and (3) If there were no products of my

favorite brand at the retailer, I would go to another retailer to seek a product of that brand. Task Effect

Task variables characterize the complexity of the decision making, and the complexity increases as the number of alternatives increases. Increase in the number of alternatives available to consumers decreases the level of performance with compensatory choice heuristics, thus consumers tend to apply non-compensatory choice heuristics when faced to more complex tasks. This indicates that non-compensatory choice heuristics are used to simplify the information processing and tend to be applied when the information load is high. The Effect of Brand Knowledge

As previously studied, brand names affect consumers’ perceived quality. This implies that consumers evaluate product qualities using brand knowledge stored in their memory. This means that brand knowledge contains relatively more information compared to other attribute information such as price. Numerous researchers have identified the role of brand knowledge as a chunk in consumers’ memory, which organizes information such as purchase experiences. Also, Keller (1998) identifies three types of brand associations: attributes, benefits, and attitudes. These associations contain various kind of information and affect consumers’ product evaluation. Provided all the evidences, it can be concluded that brand knowledge systematically contains great amount of information consumers need to make purchase decisions. Therefore, they can simplify their external information search by using brand knowledge stored in their memory.

Some studies show that the exposure of actual brand names decreases the degree of consumers’ external search. For example, the experiment of choice task conducted by Jacoby et al. (1977) found that consumers searched for a smaller number of attribute dimensions when they were allowed to use brand names than when they were not. Another experiment by Jacoby et al. (1978) showed that even when exposed to brand names, consumers’ external search varied according to their individual differences; increased external search was related negatively to their attitudinal brand loyalty. In essence, previous studies have revealed two points related to brand knowledge and

14

information search: (1) utilizing brand knowledge enables consumers to simplify their external search, and (2) this effect is more salient for those who have strong positive attitudes towards particular brands (i.e. high brand commitment) than those without such attitudes. Experiment

Using a computer-oriented IDB method, a laboratory experiment was conducted to gather empirical evidence. Task complexity and brand-name exposure were independently varied in a 2 x 2 between-subject design. The first variable was manipulated by differentiating the number of alternatives shown to the participants, 6 alternatives in the easy task and 12 in the difficult one. The second independent variable was manipulated by showing / not showing the brand names to them (“Sony, Panasonic and Sharp” or “A, B and C”). The level of brand commitment was measured and treated as an additional experimental factor (high or low, divided by average). 114 undergraduate students were recruited to participate in an experiment where they were asked to perform three evaluation tasks, and each student received a bottled-drink for his/her participation. Results

The results indicate that when exposed to brand names, consumers with a high level of BC utilized their brand knowledge thus lowering the ratio of information acquired, while those with a low level of BC did not. Further analysis revealed that those with high BC and relatively high levels of brand knowledge, preference, positive image and purchase intention particularly reduced the amount of information searched using brand knowledge in memory. Conclusion

This study shows empirical evidence of the effect of brand knowledge on consumers’ external search at the stage of product evaluation. In particular, it shows how consumers deal with the increase in task complexity. Key findings drawn from this study are: (1) consumers’ ways of dealing with the increased task complexity vary according to the utilization of brand knowledge, and (2) the level of BC plays an important role in deciding how consumers deal with more alternatives. For further information contact: Mototaka Sakashita, 7-27-6, Morikitamachi, Higashinada, Kobe, Japan, 658-0001.

15

P.12. THE INFLUENCE OF MATERIALISM AND INDIVIDUALISM ON PRICE PERCEPTION AND SHOPPING BEHAVIORS

Kittichai Watchravesringkan, University of Arizona

Jennifer Yurchisin, University of Arizona

ABSTRACT

Cultural values have received considerable attention as factors influencing consumer decisions. In the US, the values of materialism and individualism are reflected in consumers’ attitudes and subsequent behaviors. One particular consumption-related attitude that may be affected by such values is price perception. The concept of price possesses two distinct roles: positive, in which an object’s price serves as an indicator of prestige and quality, and negative, in which an object’s price is viewed as a monetary sacrifice by consumers (Lichtenstein et al., 1993). Although price perceptions in terms of these positive and negative roles have been shown to influence consumption behavior, little research has been performed to determine the direct effect of values on the perceived roles of price.

Furthermore, while several researchers have examined different market outcomes related to perceived price, this research has mainly focused on search related to specific consumption problems (Punj and Staelin, 1983). However, researchers have argued that some search activity that occurs without consumption-need recognition (e.g., browsing) may actually be a more accurate predictor of shopping preference than search related to a specific need (Bloch and Richins, 1983). Therefore, researchers need to investigate both ongoing search, or search activity that occurs without a need to solve a recognized and immediate consumption problem, and purchase behavior. This study examines how American consumers’ values and attitudes affect their consumption behavior. College students were chosen because they represent a large consumer segment with a high degree of spending power (Kessler, 1998). The objectives were to: (1) examine the impact of values on different aspects of consumers’ price perceptions, (2) assess the influence of aspects of price perception on consumers’ ongoing search behavior, and (3) investigate the relationship between ongoing search activity and shopping behavior.

Method

Data were collected through a self-administered convenience approach using students

at a major Southwestern university. The questionnaire contained five major sections concerning materialism, individualism, price perception, ongoing search and purchase behaviors, respectively. Apparel was chosen as a subject for this study because this product category can be associated with both status and value (Richins, 1994).

Results

A principal component factor analysis using varimax rotation was executed on materialism and price perception. For materialism, results revealed five underlying factors with an eigenvalue exceeding 1.0. The first factor consisted of five items with the highest eigenvalue (4.11) and the most variance explained (29.35%). Thus, this dimension was used in the final analysis for assessing materialism. Regarding price perception, the results yielded eight factors with an eigenvalue greater than 1.0. The first 7-item factor and the second 6-item factor were selected for the subsequent analysis because they yielded the highest eigenvalues (10.83 & 7.32, respectively) and proportion of variance extracted (27.76% & 18.78%, respectively). Moreover, the first factor captured the value consciousness and sale proneness constructs associated with the negative role of price. Similarly, the second factor captured the prestige sensitivity construct associated with the positive role of price. Subsequently, a structural equation analysis was performed using a two-stage technique via LISREL 8.3. No significant relationship was found between materialism and individualism. However, all of the other hypothesized paths were significant in the model. That is, materialistic consumers were likely to perceive price as both a signal of prestige and value-conscious sale searching. Likewise, individualism positively influenced consumers’ price perception cues in terms of both the prestige and value-conscious sale searching aspects. Participants’ favorable attitudes towards the positive and negative roles of price positively influenced ongoing search behavior, which subsequently positively influenced consumers’ shopping behavior.

16

Discussion and Implications Materialistic values were not correlated

with individualistic values in this study. This result is similar to the American Family Values Survey conducted in early 1990, which reported that, although America is an individualistic culture, Americans tend to focus less on materialistic values than on earning a good living and being financially secure (Schwartz, 1991). Highly materialistic and highly individualistic participants were found to have favorable attitudes toward both psychological aspects of price. While highly materialistic people, who tended to be concerned with self-image and status, responded favorably toward prices that signaled a sense of prestige, they also responded favorably to price cues associated with value-conscious sale searching. Similarly, highly individualistic people were also likely to base their price perceptions on both the positive and negative roles of price. These findings indicate that, while Americans still associate high price with prestige, the recent economic downturn seems to have created concern about the value of products for all American consumers.

The two major cues of price as perceived by consumers had an effect on marketplace behavior. Regardless of whether consumers had a favorable attitude toward the positive or negative role of price, ongoing search activity was a predictor of actual shopping behavior in both cases.

Overall, the findings suggest that companies need to focus not only on promoting the psycho-sociological aspects (i.e., prestige) associated with their merchandise, but also focus on developing effective promotional campaigns (e.g., sale events) that emphasize the value associated with their products. Regarding the effect of ongoing search and actual shopping behavior, retailers need to provide excitement (e.g., retail-tainment) to attract and retain consumers because ongoing search behavior leads to actual shopping behavior.

References Bloch, P. H., & Richins, M. L. (1983), “A

Theoretical Model for the Study of Product Importance Perceptions,” Journal of Marketing, 47 (Summer), 69-81.

Kessler, A. (1998), “College Students: A Market Worth Pursuing,” Bank Marketing, 30 (7), 14-17.

Lichtenstein, D. R., Ridgway, N. M., & Netemeyer, R. G. (1993), “Price Perceptions and Consumer Shopping Behavior: A Field Study,” Journal of Consumer Research, 30 (May), 234-245.

Punj, G. N., & Staelin, R. (1983), “A Model of Consumer Information Search for New Automobiles,” Journal of Consumer Research, 9 (March), 366-380.

Richins, M. (1994), “Special Possessions and the Expression of Material Values,” Journal of Consumer Research, 21 (December), 522-533.

Schwartz, S. (1991), “Materialism Declining in U.S. Survey,” National Underwriter, 95 (51), 14.

For further information contact: Kittichai Watchravesringkan, Division of Retailing and Consumer Sciences, The University of Arizona, Tucson, AZ 85721-0033. Phone: 520-621-1295, Fax: 520-621-3209, E-mail: kitticha@ag.arizona.edu

17

P.13. HOW PERCEPTION OF PACKAGE SIZE CHANGES INFLUENCES CONSUMER CHOICE

Michael Wiles, Indiana University

ABSTRACT

The issue of consumer perception of package size changes was recently brought to my attention by a friend. She complained that a packaged food company had reduced its product by 25 percent, but the price had remained the same. I had also purchased the product, but I had not noticed the reduced size.

This unannounced package size decrease does not appear to be a rare event. Researchers identified over 25 product categories that had downsized their products over a five year period by reducing package size and the level of fill. Firms also change package sizes by adding air or excessive packaging to make their products appear larger. Andy Rooney, among others, has pointed out how bags of potato chips are puffed full of air and how the cardboard troughs of candy bars stick beyond the candy. Thus, manipulating package sizes appear to be a common managerial practice.

As long as firms disclose the actual package content amount, firms can legally reduce package size and content without lowering price. By the same token, firms can introduce packages that appear to contain more at a higher price. Because of the disclosure requirements, package size changes are not considered deceptive. But since consumers rarely read the quantity information on the package, consumer detection of package size changes is an important public policy issue.

Consumer response to packaging is an under researched area of consumer behavior. One research stream has focused on package size and shape effects on consumption. The other stream has examined how changing package appearance affects consumer attention, evaluation and choice. Changing package form and increasing novelty of the package’s colors lead to increased attention and consideration, but package color and form changes have inconclusive effects on sales response. Our goal, therefore, is to better understand the effect of package size changes on brand choice.

The objective of this paper is to answer two questions about package size changes. First, does consumer perception of package size changes follow Weber’s Law? Second, how do package size changes influence brand consideration and choice? I propose that consumer detection of package size changes are governed by Weber’s

Law and that the difference threshold does not decay over time. I also expect that that package size increases increase brand choice, and package size reductions decrease brand choice due to a size-value heuristic. Additional moderating effects are proposed for the package size change-brand choice relationship. These propositions will be tested using a computerized shopping simulation. References available upon request. For further information contact: Michael Wiles, Indiana University, Kelley School of Business, 1309 East Tenth Street, Bloomington, IN 47405, Phone: 812-855-8878, Fax: 812-855-6440, Email: mwiles@indiana.edu

18

1.1.1 UP CLOSE AND PERSONAL: AN INDIVIDUAL LEVEL ANALYSIS OF THE DISPOSITION EFFECT

Ravi Dhar, Yale University

Ning Zhu, University of California, Davis ABSTRACT In this paper, we focus on one of the most widely documented bias in investor behavior, the disposition effect. The disposition effect refers to the tendency to sell previously purchased stocks that have appreciated in price (“winners”) and the reluctance to sell those that are trading below their purchase price (“losers”). Starting with Shefrin and Statman (1985), a number of researchers have demonstrated the basic effect using different investor databases (Odean 1998, Venezia and Shapira 2001, Weber and Camerer 1998). This work clearly documents the existence of the disposition effect. Now that the existence question has been asked and answered, it is possible to explore the conditions under which the effect is prevalent, and what investor characteristics are correlated with the bias. These “How” and “When” questions are important for several reasons. If we find that trading heuristics are correlated with specific conditions, this has clear implications for the dynamics of asset prices in bubbles and crashes. Second, if we find that certain type of investor is more susceptible to biases, it will have welfare and regulatory implications. In particular, social security and retirement investment account reform depends in part on the assumption that individual economic choice is by and large rational. If we find that this is true for some but not others, the regulation and reform will clearly have differential effects. Indeed, self investment choice may not be pareto-improving. Another implication of differential biased behavior is that rational agents may profit from the poor heuristic of irrational agents. This motivates theorists to accommodate investor heterogeneity in sophistication and style in their models (Barberis and Shleifer 2002). Our work in this paper takes a step toward determining which investors have advantages over other investors in the equity investor universe. Our paper differs from previous research in that we investigate the disposition effect at the level of each individual. All evidence to date for the disposition effect has been demonstrated by aggregating the data across all trades and traders to arrive at the mean disposition effect for a representative investor

in the market place. Recent research, for example Goetzmann and Massa (2002) and Dhar and Kumar (2002), find significant heterogeneity in investor beliefs and trading styles. Such systematic differences in trading heuristics across individuals imply that the mean value is not the whole story. This paper conducts a cross sectional study of the disposition effect that analyzes investor characteristics contributing to the heterogeneity in the tendency of investors to ride losers and sell winners. Our analysis strongly confirms findings of previous studies (Odean 1998, Grinblatt and Keloharju 2001b) that individual investors, on average, exhibit the disposition effect. However, we find that despite the significant disposition effect on average, about one fifth of investors in our sample exhibit behavior opposite to the disposition effect. Consistent with our main hypotheses, we find that investor characteristics that correspond to greater sophistication such as investor income, profession and trading experience attenuate the magnitude of the disposition effect. Specifically, we find that wealthier and investors in professional occupations show a significantly smaller disposition effect. Furthermore, investors who trade frequently also have a smaller disposition effect, implying that repeated trading experience may help investors trade “out of” the disposition effect. Our findings have important implications for policy makers as well as behavioral financial theorists. First, as certain investors are more susceptible to the disposition effect than others, individual investor organizations should focus on helping early-stage investors become aware of this tendency and adjust their trading accordingly. Second, since a large fraction of individual investors never realize their losses, we suggest that brokerage firms should try to educate their clients of the disposition effect, thereby improving their clients’ after tax portfolio performance and increasing the added value of their services. Finally, certain demographic and socio-economic groups show a greater disposition effect, which may adversely affect their portfolio return. The increase in self-investing highlights the role of

19

government agencies and non-profit organizations in making investors aware of their trading heuristics and biases. We therefore suggest that government agencies should exist to better inform and educate low-income, non-professional and young investors. For further information contact: Ravi Dhar, Yale University, Yale School of Management, 135 Prospect Street, Box 208200, New Haven, CT 06520-8200, Email: ravi.dhar@yale.edu

20

1.1.2 PROMOTION AND PREVENTION ACROSS MENTAL ACCOUNTS: PROMOTION AND PREVENTION ACROSS MENTAL ACCOUNTS: WHEN FINANCIAL PRODUCTS DICTATE

CONSUMERS’ INVESTMENT GOALS

Rongrong Zhou, Hong Kong University of Sciences and Technology Michel Tuan Pham, Columbia University

ABSTRACT We propose that consumers’ investment decisions can be usefully studied from a self-regulation perspective (e.g., Carver and Scheier 1998). In particular, consumers’ investment decisions are largely driven by generic regulatory systems known as promotion and prevention (Higgins 1997, 1998). The promotion system regulates nurturance needs and those motives related to aspiration and accomplishment (i.e., ideals). The prevention system regulates security needs and those motives related to safety and responsibilities (i.e., oughts). Since promotion and prevention-oriented financial self-regulation is difficult to achieve simultaneously, investors tend to perform the two types of self-regulation using separate “mental accounts” (a promotion focus account and a prevention focus account). Once these two accounts have been set up, consumers would post various financial products to these two accounts according to these products’ representativeness to the motives and goals defining these accounts (see Brendl et al. 1998). Financial products such as individual stocks and trading accounts tend to be associated with a promotion focus, which emphasizes growth and aspirations. In contrast, financial products such as mutual funds and retirement accounts tend to be associated with a prevention focus, which emphasizes security and protection. This association in turn leads to systematic differences in the way consumers make decisions about these products. A major correlate of promotion versus prevention is a differential sensitivity to positive versus negative outcomes. The promotion system tends to be more sensitive to positive outcomes (e.g., gains), and the prevention system tends to be more sensitive to negative outcomes (e.g., losses). Therefore, we predict that investment decisions involving financial products tagged with a prevention focus will exhibit a relatively greater sensitivity to potential losses than to potential gains. However, investment decisions involving products tagged with a promotion focus will exhibit a relatively greater sensitivity to potential gains than to potential losses. This prediction is tested in

Experiment 1. Results from Experiment 1 reveal that consumer investors evaluating an individual stock held in a trading account were found to be very sensitive to upside potential, but curiously oblivious to downside potential. In contrast, investors evaluating a mutual fund held in a retirement account were found to be very sensitive to downside potential, but oddly insensitive to upside potential. Another major correlate of promotion versus prevention self-regulation is a different propensity toward risk. In most situations, promotion-focused regulation entails greater risk-taking, whereas prevention-focused regulation entails greater risk aversion. Therefore, we argue that in decisions involving financial products tagged with a promotion focus, consumers should exhibit greater risk-seeking than in decisions involving financial products tagged with a prevention focus. We predict that this difference in risk propensity will hold even for money that bears only a mental association with promotion versus prevention-focused financial products. In particular, we predict that money to be withdrawn from a prevention-tagged account will be invested more conservatively than money to be withdrawn from a promotion-tagged account. Furthermore, this effect will be amplified when consumers’ self-regulation concerns are heightened. These predictions are tested in Experiment 2. Results reveal that respondents whose money was to be withdrawn from a trading account were much more willing to invest in a risky business venture than respondents whose money was to be withdrawn from a retirement account. These differences in risk propensities were accentuated when respondents were encouraged to reflect on their goals, that is, when their self-regulation systems were actively engaged. Experiments 1 and 2 provided input-output evidence of a relation between certain financial products and investment behavior (in terms of sensitivity to gains and losses and risk propensity). Experiments 3 and 4 provide more process-level evidence that this relation could be due a mental association between certain financial products and

21

distinct regulatory focuses. In experiment 3, it was found that making decisions involving different types of financial products resulted in subsequent unrelated decisions being carried out with distinctive promotion versus prevention orientations. Specifically, respondents who had made decisions about individual stocks in trading accounts were later found to prefer consumer products with promotion-related benefits and favor approach strategies in friendship. In contrast, respondents who had made decisions about mutual funds in retirement accounts were found to prefer products with prevention-related benefits and favor avoidance strategies in friendship. In experiment 4, it was found that the mere priming of a promotion versus prevention focus through tasks that have little to do with financial decision making produced distinct patterns of allocation of money across types of accounts and across types of assets. Specifically, when the unrelated tasks (proofreading and anagrams) were framed in terms of avoiding misses, thus priming a prevention focus, respondents’ allocations tended to shift (a) toward the IRA (as opposed to the online trading account) and (b) toward the mutual fund (as opposed to the individual stock). In contrast, when the unrelated tasks were framed in terms of maximizing hits, thus priming a promotion focus, respondents’ allocations tended to shift (a) toward the online trading account (as opposed to the IRA) and (b) toward the individual stock (as opposed to the mutual fund). Taken together, the results show that (a) investors are asymmetrically sensitive to potential gains versus potential losses across financial products; (b) investors are differentially risk-seeking with money mentally associated with different financial accounts; and more importantly (3) these effects occur because investors associate different financial products with distinct self-regulatory focuses. Overall, the findings are consistent with the view that promotion and prevention self-regulation are pursued through separate mental accounts and that consumers post different financial products to these mental accounts set up around different regulatory focuses. As a result, decisions involving these products are carried out with a distinct promotion or prevention orientation. For further information contact: Rongrong Zhou, HKUST, Email: mkrrzhou@ust.hk

22

1.1.3 ONCE-BURNED-TWICE-SHY, COUNTER-FACTUAL REASONING, AND THE REPURCHASING OF STOCKS BY INDIVIDUAL INVESTORS

Terrance Odean, University of California, Berkeley

Michal Strahilevitz, University of Arizona Brad Barber, University of California, Davis

ABSTRACT We test two hypotheses regarding investor decision making by analyzing trading records for 66,465 households at a large discount broker (LDB) between January 1991 and November 1996 and 665,533 investors at a large retail broker (LRB) between January 1997 and June 1999. The first hypothesis is that, when deciding whether to repurchase a stock that they have previously sold, investors are strongly influenced by potential counterfactuals. When an investor repurchases a stock that was previously sold, this highlights the counterfactual that the investor could have achieved the same position by never selling in the first place. If the stock has declined since the investor sold it, focusing on that counterfactual generates good feelings (e.g., "it could have been worse"). If the stock has gained since the investor sold it, focusing on the counterfactual generates regret (e.g., "if only"). We find that investors are far less likely to repurchase stocks when the counterfactual generates regret, that is, investors are less likely to repurchase stocks that have had positive returns since the sale. Investors who already own a stock are also more likely to purchase additional shares of that stock if its price has declined since they bought it rather than increased. We confirm this earlier finding of Odean (1998) with our two much larger and more recent datasets. This behavior is consistent with regret or pleasure associated with the counterfactual of having made the entire purchase at the initial price. However, this behavior can also be understood in terms of Kahneman and Tversky’s (1979) Prospect theory: investors facing a loss take on more risk (i.e., additional shares). In contrast, investors who have not previously owned a stock (during our sample period) are more, not less, likely to buy stocks with positive recent returns. Our second hypothesis is that repurchase decisions are influenced by a simple form of reinforcement learning whereby investors are more likely to repurchase stocks that were previously sold for a gain (e.g., "good to me so far") than stocks

previously sold for a loss (e.g., "once burned twice shy"). We find that investors are twice as likely to repurchase stocks previously sold for a gain, but this effect is much stronger for stocks that have had negative returns since they were previously sold. In the case of all our findings, investors do not appear to benefit financially from the behavior, which may contribute to underdiversification. We examine alternative explanations for our findings. For further information contact: Terrance Odean, Haas School of Business, 545 Student Services #1900, University of California at Berkeley, Berkeley, CA 94720-1900, Email: odean@haas.berkeley.edu

23

1.2.1 CONSUMER RESPONSE TO PERFORMANCE-BASED PERSUASION ATTEMPTS TO REBUILD WEAK BRANDS

Julie A. Ruth, Rutgers University

Anne S. York, University of North Carolina

ABSTRACT

Studies of persuasion have investigated the development of attitudes towards new objects and issues as well as objects and issues that are familiar to consumers. Moreover, recent research has focused on how various communication strategies affect strong and weak brands. Yet, little research has been directed toward the development of prescriptive measures that weak brands could undertake in order to regain their stature in the eyes of consumers. We address this gap in research by focusing on performance improvement messages and how they can be structured to yield favorable attitude change. In our case, the brand is highly familiar but has been weakened by poor performance on important attributes.

Our research addresses message characteristics hypothesized to influence consumer attitudes toward such brands. Investigated in two experiments, these message characteristics include information form (numeric or verbal information), comparison of current performance to the brand’s past performance (trend) or the industry’s performance (competitor analysis), and size of performance improvement (large or small). We investigate these message characteristics in order to understand how evidence of improvement should be configured to elicit the largest degree of favorable attitude change. We also couple these message characteristics with a well-known aspect of message processing – namely, the message source – to investigate interaction effects.

Based on consistency and correspondence theories, our results demonstrate that message source exhibits robust moderating effects with the other three message characteristics on attitude change. Implications for theory and practice are discussed. For further information contact: Julie A. Ruth, School of Business, 227 Penn Street, Rutgers University, Camden, NJ 08102, Email: jaruth@camden.rutgers.edu

24

1.2.2 THE EFFECT OF BRAND COMMITMENT ON THE SIZE OF THE CONSIDERATION SET

Sekar Raju, State University of New York at Buffalo H. Rao Unnava, Ohio State University

ABSTRACT

This research focuses on the effects of brand commitment on consumers’ consideration sets. In particular, we investigate the differences between high and low commitment consumers on the size of their consideration set, and the possible reasons for the differences in size. The results of three studies show that brand commitment is instrumental in proactively eliminating competing brands from an individual’s consideration set.

Multiple brands exist in most product categories but consumers rarely consider all the available brands when making a purchase decision. Consumers limit the number of brands they would consider purchasing to a smaller, manageable number. This smaller set of brands is called a consideration set (Kardes 2001).

While discussing consideration sets, researchers have speculated that the size of the consideration set depends on the commitment (or brand loyalty) that consumers have toward the brand. Specifically, high commitment consumers are expected to have a smaller set size than low commitment consumers. We tested this hypothesis in the first study by having subjects recall all the brands they would consider purchasing if they were in the market to buy the product category. As expected, high brand commitment consumers were found to have a smaller consideration set size compared to less committed consumers.

We next turn our attention to understanding the process by which committed consumers were restricting the number of brands in the consideration set. A review of the past literature suggests that the consideration set size is affected by memory factors (e.g. inhibition effects) or through one’s attitude toward the brand.

It is known that memory plays an important role in determining the size of one’s consideration set when it is elicited by a recall measure (e.g., Lynch and Srull 1982; Nedungadi 1990). If a brand does not come to one’s mind, it cannot be part of the consideration set in a memory-based task. On the other hand, when a brand is recalled, it has been shown that activation of the brand name in memory can serve to inhibit the recall of other brands (e.g., Alba and Chattopadhyay 1985). In the first study, subjects with higher levels of brand commitment listed

fewer brands than subjects with lower levels of commitment. This effect may be explained in terms of a memory effect. It is possible that high commitment subjects have their preferred brand at the top of their mind that may then inhibit the recall of other brand names. For low commitment subjects the preferred brand may not be as strongly positioned at the top of their minds, thereby limiting the inhibition effect.

In study two we addressed this explanation

by presenting a stimulus-based task where subjects circled the brands that they would consider buying from a list of brands provided to them. If the set size difference between high and low brand commitment subjects disappears, it indicates that memory plays a role in determining the set size. However, if the difference persists even in a stimulus task, the memory explanation (inhibition effects) can be ruled out. The results revealed that the set size difference between the high and low brand commitment consumers continued to exist even with a stimulus task, indicating that the set size difference was not attributable to memory effects.

We now turn to the attitude explanation. Brands that are evaluated positively are more likely to be included in the consideration set compared to those that are not evaluated positively. Thus, a high commitment individual may limit the size of the consideration set in one of two ways. First, because of the attachment to the preferred brand, the attitude toward the brand may be polarized (Petty and Krosnick 1995). Therefore, for other brands to be considered as viable alternatives, those evaluations have to match that of the preferred brand. Several brands may fail to reach this high threshold level, and thus not be included in the consideration set.

A second approach that committed consumers may use to control the number of brands in the consideration set is to evaluate the other brands less positively. In this case, the evaluation of the brands in the marketplace are altered or biased to be less favorable. Therefore, given the same brand, high commitment subjects may rate it less positively than do low commitment subjects, thereby restricting the potential candidates for inclusion in the consideration set. A brand that might have just made it into the consideration set originally will now be left out because it is

25

evaluated less favorably. This issue is investigated in study three.

We test these two alternative explanations by identifying and comparing the evaluations that high and low commitment subjects had for the lowest ranked brand within the consideration set (WCS) and the highest ranked brand outside the consideration set (OCS). If high commitment subjects have a higher threshold of acceptance into their consideration set than low commitment subjects, the evaluation of WCS brand for high commitment subjects should be significantly greater than the evaluation of WCS brand for low commitment subjects, while there should be no difference in the evaluations of the OCS brand between high and low commitment subjects. On the other hand, if high commitment subjects have smaller consideration sets because they lower their evaluations of competing brands, we would expect to see a less favorable OCS brand attitude for the high compared to the low commitment subjects and find a non-significant difference in the attitude between high and low commitment subjects for the WCS brand.

The findings from this study support the explanation that high commitment subjects have smaller consideration sets because they lower their evaluations of competing brands.

To summarize, we find that brand

commitment proactively guides purchase behavior by limiting the number of brands committed consumers include in their consideration set compared to less committed consumers. Of the three potential explanations for the difference in the consideration set size – memory, raising the threshold, and lowering the ratings of competing brands, we find that high commitment consumers seem to be restricting the set size by lowering the ratings of the competing brands. For further information contact: Sekar Raju, State University of New York at Buffalo, 231 Jacobs Management Center, Buffalo, NY 14260, Email: sraju@buffalo.edu

26

1.2.3 SPILLOVER OF NEGATIVE INFORMATION ON BRAND ALLIANCES

Nicole L. Votolato, The Ohio State University H. Rao Unnava, The Ohio State University

Sekar Raju, The State University of New York at Buffalo

ABSTRACT As instances of co-branding and alliances

increase in number, the brands involved tend to become associated with one another in the minds of consumers. Keller (1993) suggests that when consumers relate two brands, they transfer information relating to one brand to the other brand in the association. Simonin and Ruth (1998) call this phenomenon the spillover effect. While positive spillover effects have been demonstrated, there has been less attention in the literature to the possibility of a negative spillover. Negative spillover occurs when a negative act of one company in an association affects the other company in an alliance.

Recent research has examined the effects of negative information spillover in associations between brands or between brands and endorsers (Swaminathan and Reddy 2003; Till and Shimp 1998), and suggests that negative spillover from one brand to another in an alliance is likely. Evidence from the marketplace seems to suggest that the parent brand will escape spillover effects when the ingredient brand is the target of negative information.

We attempt to understand the discrepancy

that exists between the literature and the market by investigating whether negative spillover occurs, and the conditions under which it occurs. We explore two types of negative information: moral and competence. Negative moral information results from behavior by the brand that conflicts with a consumer’s established idea of ethical standards (Wojciszke et al. 1993). Negative competence information results from a company’s failure to meet quality standards, as perceived by consumers (Ahluwalia 1996). We also examine two types of alliances because research suggests that consumers react differently to the two types of negative information, depending on whether it is about the partner company (intrinsic relationship) or the endorser partner (extrinsic relationship).

Wojciszke et al. (1993) suggest that consumers are more forgiving of competence failures than moral failures when the target of the negative information is a person. Ahluwalia et al. (2000) suggest that for companies, failures in

executing primary functions are less likely to be forgiven than failures in the moral domain for companies. Therefore, we hypothesize an interaction between information type and alliance type such that if two brands are engaged in an intrinsic relationship, competence failures of the ingredient brand will result in less positive attitudes towards that brand and the parent brand than moral failures, and if two brands are engaged in an extrinsic relationship, moral failures of the ingredient brand will result in less positive attitudes towards that brand and the parent brand than competence failures.

A 2x2 between-subject design was used with

type of negative information (moral or competence) and type of relationship (intrinsic or extrinsic) as the two factors. Subjects read a fictitious newspaper article representing one of the four experimental conditions. All articles involved Saturn as the parent brand and another ingredient brand (Manufacturer XYZ or spokesperson John Smith) as the target of the negative information.

A 2x2 ANOVA was conducted on ingredient brand attitudes, parent brand attitudes, and competence and moral beliefs. This analysis for ingredient brand attitudes yielded a significant interaction with attitudes in the hypothesized direction. The parent brand analysis resulted in a non-significant interaction demonstrating that negative information about one of the partners did not spill over to the other partner. The analysis of competence and moral beliefs revealed significant main effects which suggested that the manipulations worked as intended, but the spillover to the parent brand did not occur as expected. The lack of spillover effects is consistent with what has been observed in the market, but inconsistent with the findings of prior research. Three possible reasons were investigated. First, if the product category used, cars, is of limited interest to the students, then they may not process the information as carefully as they were expected to. This issue was addressed in experiment two, where a popular brand of clothing was used as the parent brand. Second, if people have strongly established attitudes toward a brand that currently exists on the market, it may be difficult to affect those attitudes with one piece of negative

27

information about its partners. This issue was addressed in study three by the use of fictitious brand names. Third, lack of spillover to the parent brand might be due to the low severity of negative information pertaining to the ingredient brand. Study four addressed this issue. Studies 1 through 4 consistently showed that negative information about a partner brand affects that brand. If the information is about competence, the partner brand is affected more if it is a company than if it is a spokesperson, but if the information is about morality, the partner brand is affected more if it is a person than if it is a company. Whether the parent brand is well known or fictitious, or whether the negative information is moderately severe or is very severe, the parent brand appears to escape the wrath of the subjects. An examination of cognitive responses revealed that subjects did not believe the parent brand to be equally culpable. This suggests that if the subjects believe the parent brand is aware of the partner’s negative actions, they may attribute the cause of the negative information to both brands (Jones at al. 1971), resulting in spillover to the parent brand. This was tested in Study 5. A paragraph was added to stimuli for all experimental conditions that indicated the parent brand to be aware of the partner brand’s misdeeds, yet chose to maintain the current arrangement. This information was expected to shift some of the blame for the current situation to the parent brand too. The analysis yielded a significant interaction for both ingredient and parent brand attitudes, suggesting that spillover did occur in the hypothesized direction. A significant main effect of blame was found such that subjects in the experimental groups placed more blame on the parent brand than subjects in the control group. The findings suggest not all types of negative information are similar in their effects on brand evaluations, and brands are generally impervious to negative publicity surrounding its partner brand. The parent brand was affected only when subjects perceived the brand to be blameworthy for the negative occurrence.

For further information contact: Nicole L. Votolato, Fisher College of Business, The Ohio State University, Fisher Hall 530, 2100 Neil Avenue, Columbus, OH 43210, Phone: (614) 292-4710, Email: votolato.2@osu.edu

28

1.3.1 FORGETTING AND FALSE REMEMBERING: THE ROLE OF IMAGERY ON ERRORS OF OMISSION VERSUS COMMISSION

Ann E. Schlosser, University of Washington School of Business

ABSTRACT

“It is singular how soon we lose the impression of what ceases to be constantly before us. A year impairs, a luster obliterates. There is little distinct left without an effort of memory, then indeed the lights are rekindled for a moment - but who can be sure that the imagination is not the torch-bearer?” Lord Byron

As reflected in the above quotation, it is often difficult to separate true from false memories. In marketing, there has been a call “to explore the factors under [marketers’] control that can influence imagery vividness and concreteness and that thus affect consumers’ abilities to remember product-related information” (MacInnis and Price, 1987, p. 477). Though research has examined the effect of marketing communications on true memories, with few exceptions (Braun 1999; Braun, Ellis & Loftus 2002), the effect of marketing communications on false memories has been largely ignored.

One objective of this research is to explore the factors under marketer’s control in interactive media environments that evoke vivid images and thereby improve consumers’ recollections of product information. Specifically, having a virtual product experience (i.e., object-interactivity) may be one such tool. It is further proposed that although object-interactivity may improve certain aspects of what individuals remember, it may also lead to the creation of vivid internally-generated recollections that pose as real memories. Indeed, memories can be of events externally triggered (perceived experiences) and internally generated (those imagined; Arbuthnott et al., 2001). Thus, there are two types of memory errors: errors of omission and commission (Kimball and Bjork 2002). Errors of omission occur when individuals forget information that was externally triggered. Errors of commission are false memories, which occur when individuals incorrectly believe that an internally-generated memory was externally generated. Thus, the second objective is to examine whether object-interactivity reduces one type of error (errors of omission) while increasing another (errors of commission).

Object-interactivity allows the user to directly manipulate objects in a virtual world (Schlosser 2003). Direct manipulation of virtual

objects occurs when there is a continuous change in images as a result of user behaviors that resemble the corresponding physical behavior (Shneiderman 1987). For instance, clicking on the zoom button on a graphic of a camera and then seeing the graphic change to zoom in on the focal object would be an example of object-interactivity. Typing the syntax command “zoom in” or clicking on a hyperlinked text to zoom in on the focal object would not be.

By seeing the result (e.g., the close-up of the focal object) as a direct consequence of the user’s action (e.g., clicking on the zoom button), it is likely that the action and the result will be stored together as a unified image in the user’s memory. Indeed, imagery appears to improve the learning of object pairs because it provides a holistic construction (Bower 1972). Thus, the object-interactive site should improve individuals’ memory of the feature needed to perform a given function (the button located on the top left-hand side of the camera is used to zoom in on an object). Thus, those who visit the object-interactive site should correctly match more features with their corresponding functions than will those who visit the passive site.

Although object-interactivity should improve individuals’ true memory, it may also lead to more false memories. It is proposed that when individuals try to remember whether a product has certain capabilities, individuals will retrieve and search their stored mental image of the product. However, such processing objectives can cause people to construct a new representation of the experience (Wyer et al 2002). In fact, reconstruction is especially likely when individuals attempt to remember an event (Arbuthnott et al., 2001; Heaps and Nash 2001). According to the source-monitoring model, individuals mistake internally-generated representations for externally-generated representations when the internally-generated representation is clear and did not take a great deal of cognitive effort to create (Johnson et al 1993). It is proposed that with an already vivid mental image, those who visited the object-interactive site will easily envision attributes that were absent. Because of the ease and vividness of this image, they may mistake this internally-generated representation for an externally- (marketer-) generated representation, thereby

29

making an error of commission. Those who visited the passive site, however, should have a nonvivid image of the product and thus should have a more difficult time generating an image of an absent feature. Thus, they should correctly identify the absent feature as such.

If the ability to easily generate vivid mental images causes individuals to make more errors of commission, then individuals who are capable of evoking vivid mental images on their own (vivid imagers) should make more errors of commission than those less capable of this (nonvivid imagers). Furthermore, because instructions to imagine cause individuals to use mental images (Heaps and Nash 2001), encouraging those at the passive site to imagine themselves using a product (i.e., to use mental imagery) while encoding information should result in greater source confusion when later attempting to discern whether a remembered feature was real or imagined. Consequently, they should make more errors of commission than will those who visited the passive site without such instructions.

Two experiments were conducted, the results of which are consistent with the hypotheses. Specifically, although object-interactivity enhanced memory of the features used to perform specific functions, it also led to more errors of commission compared to the passive site. The ability to easily generate vivid mental images appears to account for false memories: vivid imagers made more errors of commission than nonvivid imagers did. Neither improvements nor decrements in learning could be attributed to processing elaboration or processing style. Furthermore, it was not imagery ability itself but rather the ability to evoke vivid mental images (not control these images) that affected the amount of errors of commission made. Moreover, as predicted, visitors to the passive site made more errors of commission when they were encouraged to evoke their own images (i.e., use their imagination) while encoding information than when they were given general instructions.

30

1.3.2 THINKING ABOUT MONEY: THE ULTIMATE EQUALIZER?

On Amir, Yale University Dan Ariely, Massachusetts Institute of Technology

Ziv Carmon, INSEAD ABSTRACT

Money has inherent benefits: it can be stored to produce savings, it can smooth consumption over time, it can divide wealth to be used more efficiently, and most importantly, it is a universal language for trade, substitutable for almost any experience. The underlying assumption in the use of money is that the mapping from pleasure/benefit (utility) into monetary sums is clear and unambiguous. We argue that this is not always the case. We claim that money is difficult to map onto utility or happiness and therefore individuals rely on external cues and heuristic processes, even when those should not be used. We argue that the difficulty is in the mapping itself, and this impediment comes from several characteristics of money, which make it particularly hard to map: it is of no (hedonic) experiential nature, it is numerical and linear, thus compatible with irrelevant cues, and it is easily associated with external market values that may be locally immaterial.

In the first two experiments, we show that

individuals overly rely on dimensions that are more compatible with money or numerical values when they need to evaluate an experience in monetary terms, but not in other terms. We find this discrepancy for both maximal willingness to pay (WTP) for a concert, where participants cared more about the cost of the concert than attributes they themselves thought would influence their experience more (room-temperature during the show) when thinking about the right price for a ticket, and for minimal willingness to accept (WTA) as wages for performing labor-intensive tasks. The third experiment extended these results to evaluations of other people’s work, in a different context, and replicates the findings. Moreover, it demonstrates that individuals rely on irrelevant external money-compatible cues more than about their own potential losses.

The fourth experiment demonstrates the

difficulty of the mapping by comparing WTA in contexts that either have an available market reference price or do not. We find that, as predicted, the existence of such reference decreases the

sample variance by two orders of magnitude, supposedly centering the distribution around the reference price. More importantly, absent such reference, we find variances that suggest that individuals do not have a clear mapping between predicted effort and a monetary sum. This means that when asked for a minimal amount one is willing to accept in order to perform a certain task, the difficulty of the question makes people resort to using external cues that may or may not be relevant, or suited for each individual.

The fifth experiment replicated the main

effects of the fourth and specifically demonstrated the role of reference price information vs. that of general knowledge about the task, by controlling the amounts and types of information participants in the different conditions were exposed to. We find that while providing task-related information decreases variance slightly, providing remotely related market-price information (reference) decreases variance dramatically. The results suggest that while providing more information that should improve the hedonic evaluation or prediction only slightly helped participants, providing a distant market-price changed their behavior dramatically. We interpret this pattern as evidence that individuals avoid the “calculation” of the mapping of experience to money when they can, and when they don’t they have a hard time actually performing it. The final experiment added support to the notion that the difficulty lies in the mapping itself, by providing direct evidence that individuals take longer to provide a monetary reaction than a happiness estimation. This was true even after equating the response technology. Moreover, asking participants to think of other, pleasure equivalent, experiences did not influence the elicited pleasure distributions, but greatly changed the means and variances of elicited willingness to pay.

We conclude that while individuals are

well equipped to evaluate experiences at a hedonic or internal level, and thus could make beneficial choices between experiences, they are not as proficient in translating these experiences into monetary amounts. We find that some of the characteristics that make money useful may lead to

31

errors when actually thinking about using it. These findings portray market situations in which individuals may choose the wrong job, hire the wrong contractor, or make an erroneous bid. For further information contact: On Amir, 135 Prospect st., New Haven, CT 06511. Email: on.amir@yale.edu

32

1.3.3 TOWARDS AN UNDERSTANDING OF BENEFIT DEPRECIATION: AN EXPERIMENTAL

APPROACH

Jennifer Christie Siemens, University of Dayton Norma A. Mendoza, University of Arkansas

ABSTRACT

The way that individuals psychologically represent and account for their money has traditionally been studied from an economic perspective. Prelec and Loewenstein (1998) define coupling as the psychological link between the benefits and the payment associated with the consumption of an object. Gourville and Soman (1998) propose that when a payment occurs before the benefit of a transaction, the sunk cost impact of that transaction becomes less with the passage of time (defined as “payment depreciation”). Gourville and Soman (1998) also suggest that this depreciation could occur in the opposite way, whereby an upstream benefit becomes depreciated over time so that the downstream payment feels more like a pure loss (defined as “benefit depreciation”). This study seeks to establish the validity of the benefit depreciation concept as suggested by Gourville and Soman (1998). It is predicted that a transaction will be perceived as less satisfying, less pleasurable, and less fair when benefits temporally precede their associated payment in a transaction (compared to a transaction where benefit is consumed at time of payment). Furthermore, it is predicted that benefit depreciation will increase continuously as the time elapsed increases. Subjects (N = 157) were approached during class and were told that they would receive $5.00 for watching a brief video and completing a 30-minute survey during class time. Subjects in the no-delay condition were given $5.00 and were then shown a five minute video and then given the survey to complete. Subjects in the delay conditions completed these tasks over two class sessions. During the first session, delay conditions were paid the $5.00 compensation and watched the video and were told that they would complete the survey on another day. After a time delay (3 days, 1 week or 2 weeks), they were approached again during class to complete the survey. This resulted in four block-randomized conditions (one no-delay and three time delay conditions). Subjects in all conditions completed the entire study in approximately forty minutes, with the only difference between conditions being the timing of

the transaction cost. Transaction satisfaction, pleasure associated with the transaction, transaction fairness and behavioral intentions were assessed (lower numbers indicate more favorable response).

Relative to subjects in the no-delay condition, subjects in the delay conditions viewed the transaction as significantly less satisfying (F (3, 153) = 8.81, p < .001) and significantly less pleasurable (F (3, 153) = 8.25, p < .001). Analyses showed a trend in increased dissatisfaction as the temporal delay between transaction benefit and transaction cost increased (no-delay M = 3.16, 3-day M = 3.99, 1-week M = 4.37, 2-week M = 4.79, where higher means indicate less satisfaction). The trend for subjects’ pleasure-related reaction to the transaction mimicked the trend for transaction-related satisfaction, showing a steady decrease in pleasure as temporal delay increased (no-delay M = 3.47, 3-day M = 4.14, 1-week M = 4.55, 2-week M = 4.96). Subjects’ intentions of completing the transaction again and recommending the transaction to a friend were lower when a temporal delay separated transaction benefits and transaction costs (F (3, 153) = 5.35, p < .01). Behavioral intentions decreased as the temporal delay between transaction benefit and transaction cost increased (no-delay z = -.4497, 3-day z = -.0287, 1-week z = .1289, 2-week z = .3215). Subjects in the temporal delay conditions also reported lower perceptions of transaction fairness than did subjects in the no-delay condition (F(1, 153) = 6.38, p<.001). While the mean values indicate a trend in decreased perceptions of fairness as the temporal delay increases, a leveling off in the significance of this trend occurred between the 3-day and 1-week delay conditions (no-delay M =2.38, 3-day M = 2.94, 1-week M = 3.06, 2-week M = 3.40). In sum, it was found that the hedonic impact of the upstream benefit becomes depreciated over time, resulting in the downstream payment feeling more like a pure loss. When given an upstream transaction benefit, individuals experienced less satisfaction and less pleasure related to the transaction at the time of payment of a downstream transaction cost. Furthermore, individuals experiencing a temporal delay have lower intentions of performing the transaction again for the same benefit, and have lower perceptions of

33

the fairness of the transaction. The finding of benefit depreciation has important implications for repeat purchases and word of mouth, especially those that follow a “buy now, pay later” format. Managers should attempt to lessen the degree to which a benefit is depreciated within a temporal delay period. For example, they might allow for debit cards or bank transfers (or other mechanisms such as Pay Pal) to be used rather than credit cards so that the payment is perceived as being made immediately at the time of purchase. In fact, an area of future research would be to investigate whether the loss associated with benefit depreciation would still hold true when these other payment mechanisms are employed. Benefit depreciation might also be lessened by reminding consumers of the benefit that they enjoyed from the product or service to reinforce their psychological link.

Future research should further investigate these findings using other operationalizations of consumption benefits and costs. This would also help to explore alternative explanations to these findings, such as the possibility that individuals were discounting over time only due to the use of a monetary benefit. Additionally, future research should investigate the role of mood in the benefit depreciation process. It is possible that after consuming a benefit, subjects are in a better mood and thus are more likely to respond favorably regarding the transaction.

For further information contact: Jennifer Christie Siemens, Management and Marketing Department, University of Dayton, Miriam Hall 810, 300 College Park, Dayton, Ohio 45469, 937-229-1086, Email: Jennifer.siemens@notes.udayton.edu

34

2.1.1 WHEN BELIEF AND EVIDENCE CONFLICT: FACTORS THAT MITIGATE CONSUMERS’

SELECTIVE USE OF PRICE-QUALITY DATA

Tracy Meyer, University of Cincinnati James Kellaris, University of Cincinnati Frank Kardes, University of Cincinnati

Maria L. Cronley, Miami University Steven S. Posavac, University of Rochester

ABSTRACT Access to objective price-quality data should allow consumers to discern the magnitude of covariance relationships. Three experiments show the influence of preexisting price-quality beliefs on the formation of subjective price-quality covariance estimates. Consistent with selective hypothesis testing, respondents demonstrated limited use of available data. Individuals best able to interpret large amounts of data were those with plenty of time and low need for closure. Data formatted in a random (vs. rank) format was also found to improve performance. For further information contact: Tracy Meyer, phone: (513)556-7104, fax: (513)556-0425, Email: meyertr@email.uc.edu

35

2.1.2 POSITIONING HIGH-TECH HYBRIDS: FACTORS THAT FACILITATE AND INHIBIT THE

USE OF A MULTIPLE (VERSUS SINGLE) CATEGORY INFERENCE STRATEGY

Jennifer Gregan-Paxton, University of Delaware Steve Hoeffler, University of North Carolina

Min Zhao, University of North Carolina ABSTRACT Prior research has established that categorization plays a central role in new product learning. Very little is known, however, about category-based learning under conditions of categorization ambiguity. Categorization ambiguity exists when information about a new product makes it difficult or impossible to place the novel offering in a single, existing category. In this research we examine the impact of category familiarity and the nature of the category cue (perceptual versus conceptual) on the incidence of multiple (versus single) category inferencing. Multiple category inferencing occurs when two (or more) categories are used to form inferences about a new product. The results of three studies suggest that consumers rely on a single category inference strategy when a high familiarity category is presented perceptually and a multiple category inference strategy when a low familiarity category is presented perceptually or when a high tech hybrid is formed by combining the functionality of two low familiarity products.

36

2.1.3 THE UNBIASING EFFECT OF BIASING CUES

Wouter Vanhouche, University of Florida Stijn van Osselaer, Erasmus University Rotterdam

ABSTRACT Several authors have stressed the important role of cues such as price and other attributes in biasing consumers' product judgments. For example, Rao and Monroe (1989) reviewed evidence showing that consumers often use price to judge quality, even in categories where the actual price-quality correlations are low. Carpenter, Glazer, and Nakamoto (1994), among others, showed that consumers use irrelevant attributes, which are uncorrelated with quality, to judge product quality. The current research replicates those findings, but more importantly, shows that potentially biasing cues--price, irrelevant attributes, and even attributes that are merely imperfectly correlated with product quality--can actually improve consumers' judgments of product quality during later purchases. We argue that price and irrelevant attributes do not just function as heuristic cues that bias quality judgments but also as memory cues that help consumers to correctly retrieve quality information from previous experiences. When consumers have no direct quality information, they will use price or attribute information as a heuristic cue to judge quality. However, after receiving quality information, e.g., by experiencing the product, consumers will not only update the general heuristic relationship between the cue and quality, but also store exemplar information about the experienced product that includes a specific level of the cue (e.g., price or attribute), quality level, and other information such as brand name. Over time, this specific memory of a particular price or attribute can help consumers retrieve the specific quality level that accompanied the specific price or attribute, leading to more accurate quality judgment than if no price or attribute information was ever given. Thus, contrary to what is suggested by previous research, extrinsic and potentially biasing cues can actually make consumers more accurate and thereby increase consumer welfare. In Experiment 1, participants received information about three brands of orange juice (learning phase). For each juice, participants in the price condition were first shown a brand name and a price, then asked to judge the quality level of the juice on a five-star scale, and then given feedback about the actual quality level of the juice. The

actual price-quality correlation was zero. After a 20-minute delay, participants were then presented with the brand names and prices again and asked to recall the actual quality levels of all three juices (test phase). In the no-price condition, no price information was ever given. As expected, participants in the price condition used price as a heuristic cue in the learning phase, as evidenced by high correlations between price and quality judgments. However, at test, quality judgments were more accurate in the price condition than in the no price condition, as evidenced by higher correlations between quality judgments and actual quality in the price condition. Thus, whereas initial, pre-experience quality judgments were biased by price information (that was actually uncorrelated with real quality), post-experience quality judgments were made more accurate by the same price information. In Experiment 2, price was replaced by an attribute, the location (US state) where the juice was bottled. The location attribute was irrelevant in the sense that it was not highly correlated with product quality. (Unlike many other experiments involving irrelevant attributes, participants were not explicitly told that the location attribute was irrelevant.) Again, we found that this attribute biased initial quality judgments, but made post-experience judgments more accurate. Experiment 3 shows that the accuracy-enhancing effect of adding an extrinsic cue is not limited to situations in which there is no overall relationship between the extrinsic cue and product quality. In the learning phase, participants in the price condition received the same types of information for each brand as they did in Experiment 1 (brand, price, quality level). This time, higher-priced products had higher quality levels, with one exception. The last product had a high price but low quality. Participants in the no price condition received the same information, except that no price information was given. During a test phase, participants in the price condition were provided with brand names and prices and were asked to recall the quality level of each of the products. Participants in the no price condition did the same but were not given any price information. As expected, participants in the price condition judged the quality level of the "exceptional"

37

product more accurately than participants in the no price condition. In Experiment 4, we tested the hypothesis that potentially-biasing extrinsic cues (i.e., cues that participants expect to be related to an outcome such as quality) can lead to an even stronger accuracy-enhancing effect than cues that participants do not expect to be related to the outcome. This might occur because a product experience that disconfirms an expectation may draw attention to that experience and exemplar information will be stored better than if participants had no expectation about the outcome. The high-expectation condition was identical to the price condition in Experiment 3. The low-expectation condition replaced price by package size. As expected, participants in the price condition judged the quality level of the "exceptional" product more accurately than participants in the low-expectation condition. Thus, the cue with the strongest potential to bias (price) ended up having the strongest unbiasing effect over time. These findings have clear implications for marketers and consumers. They indicate, for example, that adding attribute information that violates participants expected attribute-quality relationships (e.g., putting a high price on a low quality product or introducing irrelevant attribute information) can make low quality products more attractive initially, but over time may actually help consumers recognize those products' mediocrity by reducing confusion in memory, leaving that product worse off than if the attribute information had never been added. Thus, instead of hurting consumer welfare, potentially biasing and irrelevant attributes can improve consumer welfare. From a theoretical perspective, our results suggest that consumers use cues such as price and irrelevant attributes not just as heuristic cues that have general relationships with benefits such as quality (e.g., "higher price usually means higher quality") but that individual levels of price or other attributes are also stored as part of exemplars that can be used as memory cues to retrieve specific experiences. It is argued that these two functions rely on different memory and learning systems. Further experiments will asses moderators that determine when cues are used as general heuristic cues versus specific memory cues. Future research may also explore indirect effects of adding misleading attribute information on the quality implications of other product features such as brand names. For further information contact:

Wouter Vanhouche, University of Florida, Department of Marketing, 212 Bryan Hall, Gainesville, FL 32611, +1-352-392 ext. 1333#, wouter.vanhouche@cba.ufl.edu

38

.2.2.1 EFFECTS OF NONCONSCIOUS GOAL PRIMING ON CONSUMER CHOICE BEHAVIOR

Baba Shiv, University of Iowa Joel Huber, Duke University

Tanya Chartrand, Duke University ABSTRACT What sets goal pursuit into motion? Perhaps the most intuitively appealing and compelling answer is that we do. We deliberate among our various desires and decide to pursue a particular goal in a particular social situation. We determine which strategies will best serve us in attaining the goal, engage in goal-directed behavior and plans of action, and evaluate our progress made toward the goal (Gollwitzer, 1990; Heckhausen, 1991). Thus, intuition tells us that goal pursuit is started in motion by our conscious will, and that it is a deliberate, intentional process. Reflecting this, most models of self-regulation posit continuous, conscious choice and guidance as a central feature, if not the core foundation, of goal pursuit (Bandura, 1986, 1997; Cantor & Kihlstrom, 1987; Carver & Scheier, 1981; Deci & Ryan, 1985; Dweck, 1996; Locke & Latham, 1990; Mischel, Cantor, & Feldman, 1996). However, goal pursuit does not always involve deliberate direction of goal-driven behavior. Sometimes goal pursuit occurs outside of one's awareness, intent, and even control. Social environments automatically activate goals frequently associated with them in the past, and these goals then operate to guide information processing and behavior without conscious intervention (Chartrand & Bargh, 2002). Recently, attention has turned to the mechanisms by which goal pursuit can occur without conscious guidance. There is now ample evidence that nonconscious goals, once activated, guide subsequent cognition and behavior just like their conscious counterparts. Individuals are neither aware of the goal activation itself, nor of the goal's subsequent guiding role. In a direct investigation of nonconscious goal pursuit, Chartrand and Bargh (1996) replicated a study in which participants were given conscious goals to either memorize information or form an impression of a target, and were then exposed to sentence predicates that fell into four trait categories (Hamilton, Katz, & Leirer, 1980). In the original study, participants were given a surprise free recall test, and results revealed that those given the goal to form an impression recalled more of the predicates and clustered them more by trait category. Chartrand and Bargh

replicated this, but gave participants no conscious goals. Instead, they primed participants with either an impression formation goal or a memorization goal, asked them to read the series of behavioral predicates, and then gave them a surprise free recall test. Replicating the results of the original study, Chartrand and Bargh (1996) found that participants who were primed with an impression formation goal were more likely to (a) recall more behaviors, and (b) organize the behaviors according to trait category, compared to those primed with a memorization goal. Importantly, none of the participants were aware during the study of having a goal to memorize information or to form an impression. Thus, they pursued these goals nonconsciously (see also Séquin & Pelletier, 2001). Bargh, Gollwitzer, Lee-Chai, Barndollar, and Trötschel (2002) recently conducted a series of studies that tested for the presence of motivational states during nonconscious goal pursuit. Specifically, there are unique motivational qualities that have previously been reserved to describe conscious goal pursuit (e.g., Atkinson & Birch, 1970; Bandura, 1986; Gollwitzer, 1990; Lewin, 1951). Bargh et al. (2002) found that, like conscious goals, nonconscious goals increase in strength over time, lead to persistence when participants encounter obstacles during their goal pursuit, and lead to resumption of goal-directed behaviors following interruptions. In fact, there is now neurophysiological evidence that the pursuit of conscious and nonconscious evaluative goals invoke extremely similar brain patterns (Gardner, Bargh, Shellman, & Bessenoff, 2001), suggesting that once activated, people can pursue nonconscious goals as they do conscious goals. In the current set of studies, we applied the model of nonconscious goal pursuit to a consumer choice setting. Is it possible that consumers can have goals related to purchasing behavior nonconsciously activated, which then automatically guide subsequent consumer cognition and behavior? To test this question, we devised a supraliminal priming manipulation that would unobtrusively activate one of two goals in participants: a goal for prestige, or a goal for value. In many contexts these two goals in lead to

39

opposing choice patterns. We argue that when given a choice between a product that is somewhat expensive but prestigious and a product that is a good value but not as prestigious, the goal that is nonconsciously guiding thoughts and behaviors will determine which choice the individual makes. Individuals were given one of two scrambled sentence tasks (see Bargh & Chartrand, 2000). Words related to either a goal for prestige or a goal for value were embedded in the two versions. Thus, half of the participants had a nonconscious goal for prestige, and the other half a nonconscious goal for value. At this point, participants were given scenarios that described various products on the market. It was positioned as a choice between two products: one that was clearly higher on the prestige dimension, and the other that was higher on the value dimension. Participants were asked to respond which item they would be more likely to purchase. We expected that participants who were earlier primed with a prestige goal would be more likely to choose the high-prestige item, and those primed with a value goal would be more likely to choose the high-value item. This is in fact exactly what we found. A followup study currently underway seeks to more clearly identify the motivational aspect of the nonconscious process. The implications of nonconscious goal pursuit for consumer choice are provocative and compelling. To the extent that consumers have goals activated by the environment outside their awareness or intent, their choice may be determined by factors outside their control. We have taken the first step in exploring this by identifying two goals that can nonconsciously guide consumer choice behavior. Other goals remain to be identified, as well as the factors that can attenuate this influence. For further information contact: Tanya Chartrand, Fuqua School of Business, Duke University, Room A304, 134 Towerview Drive, Durham, NC 27708, email:tanya.chartrand@duke.edu

40

2.2.2 NONCONSCIOUS PRIMING OF DECISION GOALS

Patricia W. Linville, Duke University Jesper Nielsen, University of Arizona Steve E. Hoeffler, UNC – Chapel Hill

ABSTRACT

Dual-process models of information processing have shown us that environmental and personal factors influence how carefully we process information and our reliance on decision heuristics. To give a few examples, motivation (Petty and Cacioppo 1981; Maheswaran and Chaiken 1991), time pressure (Payne, Bettman, and Johnson 1988), and circadian rhythm (Bodenhausen 1990) have all been shown to influence the extent to which consumers engage in fast, low effort, efficient, heuristic processing versus the slower, higher effort, more accurate, systematic processing. In our research we investigate whether decision goals (e.g. to be accurate versus efficient) can be activated outside our awareness to nonconsciously guide our information processing and decision making behavior. We examine the extent to which such nonconsciously activated goals of being either accurate or efficient influence cognitive behavior as evident by more or less thorough information processing. Additionally, we investigate how these goals influence judgments and decision outcomes. Each study primes decision processing goals followed by a judgment or decision task. The goal of our first experiment was to determine the role of nonconscious primes on the decision process. Participants engaged in a series of information search and decision tasks after being subliminally primed with a decision goal (accuracy-related, efficiency-related, or neutral words). We monitored participant information acquisition behavior and decision outcomes using a process-tracing technique (i.e., Mouselab) used previously in related work (e.g. Payne, Bettman, and Johnson 1988). Our findings confirm that the primed decision goals influenced the decision process. As predicted, participants primed with an accuracy related goal spent more time processing each piece of information than did the control group while participants primed with an efficiency related goal spent less time processing each piece of information. Additionally, participants who were primed with efficiency goals were more likely to

return to information already visited, indicating that they did not initially process the information carefully. Additional studies focus on the role of primed decision goals on decision outcomes and the use of heuristics in reaching these outcomes. In these studies we use a supraliminal priming method (i.e., word search puzzle) to unobtrusively prime participants with decision goals of being accurate, neutral, or efficient. We find that these primed goals influence the use of judgmental heuristics such as representativeness and the use of stereotypes as heuristics. In all of our studies, we examine both the temporary priming of information processing goals as well as participants’ chronic use of decision goals as measured by a decision making questionnaire we developed. In general, we find that it is possible to nonconsciously prime decision goals, and these primed goals guide decision behavior without awareness or intent. We address whether automatically activated decision goals follow the same process and thus produce the same decision outcomes as goals set in motion by conscious intentions. We close by discussing the role of nonconscious priming by external, environmental events on our decisions in everyday life. For further information contact: Jesper H. Nielsen, Eller College of Business, University of Arizona, 1130, E. Helen, P.O. Box 210108, Tucson, AZ 85721, Email: jesper@eller.arizona.edu

41

2.2.3 SELECTING VERSUS REJECTING IN BINARY CHOICE: QUEST FOR COMPATIBILITY

Margaret G. Meloy, Penn State University J. Edward Russo, Cornell University

ABSTRACT Choosing the best alternative from a large set of options may be approached from either of two directions. Decision makers can either select (or retain) some alternatives, implicitly eliminating all others; or they can reject some options, implicitly retaining the rest. When successive narrowing is used to create the consideration set, systematic differences have been observed between selecting versus rejecting (usually termed inclusion and exclusion). Most notably, the size of this set is typically larger after one round of rejection than after one round of selection (Huber, Neale, and Northcraft 1987). The reason seems to be that because rejection is so final, decision makers tend to retain any alternative whose evaluation is uncertain when working under rejection (exclusion) instructions (Levin, Jasper, and Forbes, 1998). Although select-reject differences may seem plausible in multi-alternative choice, why should any such differences exist in binary choice, when only two alternatives remain? Shafir (1993) argued that the key difference between the goals of selecting versus rejecting rested on importance weights and compatibility. Under reject instructions, an alternative’s negative features might receive greater weight, whereas select instructions might lead to heavier weighting of the alternative’s positive features. Ganzach (1995) hypothesized that “an accept choice is likely to entail more commitment than a reject choice”. Finally, Wedell (1997) proposed a more specific alternative explanation for Shafir’s result, the “accentuation” of attribute differences in the select task. The current work takes as a starting point the three papers described above and extends this work to positively and negatively valenced options (i.e., “benefits and burdens”; Mannix, Neale and Northcraft 1995). Might there be select-reject differences when the alternatives themselves are uniformly described in very positive (negative) terms? If Shafir’s basic claim is correct, there might still be manifestations of a compatibility effect, but that compatibility would be between the goal of the task and the valence of the alternatives (i.e., selecting the better of two positive options and

rejecting the worse of two negative ones). Conversely, selecting the better of two bad alternatives and rejecting the worse of two good ones would comprise the two incompatible cases. If such a compatibility effect were operating, the decision processes might be systematically different, with the incompatible cases resulting in greater conflict and a more disrupted flow of the choice process. Study 1 examined the choice process with an experimental method that has been used to measure distortion of information to favor a tentatively preferred option (Russo, Meloy and Medvec 1998). Decision task (select, reject) and valence of the options (positive, neutral, negative) formed the two factors in the experimental design. Subjects made decisions between two desirable or undesirable required university courses, and between two excellent or horrible employees. The employee cover story relied on the plausibility of downsizing in the reject condition and nepotism in the select condition to motivate the decision task. Measures of confidence in the final choice, as well as measures of intermediate confidence, accentuation of attribute differences, and distortion of information during the choice process were the core dependent measures. If the incompatibility between the goal of the task and the stimulus valence resulted in greater non-conscious internal conflict and angst, confidence (final and intermediate) would be reduced, attribute evaluations would be less extreme, and information distortion would be reduced. The results of Study 1 confirmed all of these predictions. Selecting from one of two unattractive options and rejecting one of two desirable options resulted in greater decision conflict. The manifestations of this conflict were reduced confidence in the final choice, less confidence during the choice process, less accentuation of attribute differences, and a reduction of information distortion (all p< .05). The flow of the decision process was badly disrupted when the decision task and the valence of the options were incompatible. Study 2 followed the 2X3 design of Study 1 and added the collection of verbal protocols to examine decision makers’ thoughts during the choice

42

process. The findings of Study 1 were replicated (all p < .05). Further, the verbal protocols revealed that for subjects in the incompatible cells, there was an increased probability of task reframing to ease the incompatibility. Finally Study 3 examined compatibility in the context of multi-alternative choice, including negatively valenced alternatives. To date, positive alternatives have been used in the phased narrowing tasks characterizing these studies. Further, subjects were instructed to follow one of three strategies in creating consideration sets: inclusion (i.e., “tell us which of the options you want to consider further”), exclusion (i.e., “tell us which options you would not consider further”), and pure rejection (i.e., “tell us which of these options you want to get rid of”). Although prior work on selecting and rejecting in multi-alternative choice has focused on the first two strategies, we included the pure rejection strategy to better examine the limits of compatibility. Employees were again used as the choice domain, with the first two strategies involving job applicants for a new position. The third strategy was framed as a downsizing task. The results of Study 3 were consistent with a compatibility effect. In multi-alternative choice, when the valence of the alternatives and the task were compatible (i.e., selecting positive options or rejecting negative alternatives), the sizes of the consideration sets were significantly smaller than when an incompatibility existed. We note that both forms of exclusion (i.e., would not consider further and pure rejection) yielded parallel results. The three studies, taken together, provide evidence that in binary and multi-alternative choice, select-reject differences are moderated by the valence of the options. When there is any incompatibility between the task and valence, decision conflict is exacerbated. Resolution of the conflict occurs through reframing of the decision problem and the generation of larger consideration sets. REFERENCES Ganzach, Y. (1995), “Attribute scatter and decision outcome: Judgment versus choice,” Organizational Behavior and Human Decision Processes, 62, 113-122. Huber, J., Neale, M., & Northcraft, G. (1987), “Decision bias and personnel selection strategies,” Organizational Behavior and Human Decision Processes, 40, 136-47.

Levin, I.P., Jasper, J.D., & Forbes, W.S. (1998), “Choosing versus rejecting options at different stages of decision making,” Journal of Behavioral Decision Making, 11, 193-210. Mannix, E.A., Neale, M. & Northcraft, G.B. (1995), “Equity, equality or need: The effects of organizational culture on the allocation of benefits and burdens,” Organizational Behavior and Human Decision Processes, 63, 276-286. Meloy, M.G. (2000), “Mood-driven distortion of product information,” Journal of Consumer Research, 27, 345-359. Russo, J.E., Meloy, M.G., & Medvec, V.H. (1998), “The distortion of product information during brand choice,” Journal of Marketing Research, 35, 438-452. Shafir, E. (1993), “Choosing versus rejecting: Why some options are both better and worse than others,” Memory and Cognition, 21, 546-556. Wedell, D.H. (1997), “Another look at reasons for choosing and rejecting,” Memory and Cognition, 25, 873-887. For further information contact: Margaret G. Meloy, Smeal College of Business, Penn State, 701-N BAB, University Park, PA 16802, email: mmeloy@psu.edu

43

2.2.4 IN SEARCH OF CONSUMER GOALS

Kurt A. Carlson, Duke University

ABSTRACT Decision process goals are desires over actions, cognitions, and feelings that relate to the decision process. This paper introduces a method for tracking the activation levels of decision process goals during consumer choice processes. The method relies on playback of an audio recorded or videotaped concurrent verbal protocol to aide retrospective goal recall. The method captures activation levels of multiple goals at multiple points during the choice process. Data from several consumer choice studies validate and demonstrate the method. In Study 1 the method was used to measure activation levels of consumers’ decision process goals in the two conditions of Huber, Payne, and Puto’s (1982) asymmetric dominance effect. Results reveal that only those goals hypothesized to drive the phenomenon (finding reasons and dominance structuring) were differentially active across conditions, and that these differences were most pronounced for good introspectors. A group of observers, who were given the actors’ memory cues (including their verbal protocols) did not report theory-consistent goal activation differences, regardless of whether they were good introspectors or not. Study 2 replicated Study 1, but without verbal protocols. Results revealed that even though the choice effect obtained, the goal activation data did not. Together, the data from Study 1 and Study 2 suggest that decision process goals can be recalled so long as consumers are provided with a sufficiently rich set of cues to constrain and aide recall. Study 3 applied the explicit goal assessment method to a real choice between 12 bottles of wine. Participants included experts and novices who were assigned to either a low or a high structure display condition. Goal activation levels for 15 goals were measured for five intervals of the choice process. Results revealed numerous theoretically-consistent goal activation patterns. For example, the early stages of the choice process were characterized by activation of learning goals (develop an impression of the wines), whereas the later stages were driven by choice goals (e.g., be confident with one’s choice). Moreover, there were dramatic differences between novices and experts, with the experts being

more interested in learning for learning’s sake, when the choice environment was conducive to such learning. The data also provided insights into selection of processing strategy and the impact of time pressure on consumers' decisions. The article concludes with a discussion of the various uses of the method. These include improving our understanding what impact experimental manipulations have on consumers’ decision process goals, explaining anomalies in consumer decision making processes, recovering consumers' goals in realistic choice settings, and identifying how environmental and individual differences can lead to differences in goal activation. For further information contact: Kurt A. Carlson, Fuqua School of Business, Duke University, Room A342134, Towerview Drive, Durham, NC 27708-0120, email: kurt.carlson@duke.edu

44

2.3.1 PROCESSING ABSTRACT VERSUS CONCRETE BRAND INFORMATION: THE INFLUENCE OF RELATIONSHIP NORMS IN CONSUMERS’ CHOICE OF PROCESSING

STRATEGIES

Pankaj Aggarwal, University of Toronto Sharmistha Law, University of Toronto

ABSTRACT

Recent work has shown that sometimes consumers cross the threshold of commercial transactions and bring the brand ‘alive’ by giving it quasi-human qualities (Fournier 1998; Aaker 1997; McGill 1998). More recently, Aggarwal (2001) finds that consumers’ evaluation of a brand is influenced by the norms of relationship that are salient at the time of brand evaluation. In the present research, we extend prior work on consumer-brand relationships to investigate differences in information processing strategies adopted by consumers depending on the type of relationship that they perceive with the brand.

We study two types of relationships, adopting a distinction developed by Clark and Mills (1993) between communal relationships, in which concern for a partner’s need is paramount (for example, relationships with friends and family members) and exchange relationships, in which a matched benefit is expected back from the partner (for example, relationship between business partners). There is substantial evidence that people’s information processing strategies are influenced by a variety of situational and contextual factors in which the decisions are made, for example, experimental instructions (Meyers-Levy 1991), competitive ads (Malaviya, Kisielius and Sternthal 1996), level of involvement (Petty, Cacioppo and Schumann 1983), motivation to achieve an outcome (Crowe and Higgins 1997), age-differences (Koutstaal and Schacter 1997) and cultural orientation (Nisbett, Peng, Choi and Norenzayan 2001).

Our thesis that different relationship types will be associated with alternative information processing strategies is based on the premise that an exchange orientation will be associated with greater item-specific encoding of brand information whereas a communal orientation will result in more holistic processing. Prior work (Clark 1984, Clark, Mills and Corcoran 1989) has found that people working on a task jointly with their partner were more likely to keep track of their own inputs separately from their partner’s when in an exchange relationship but attend more to information about

others’ needs when in a communal relationship. Presumably, keeping track of inputs in exchange relationships is done in order to allocate the final benefits in proportion to the inputs but such record keeping is unnecessary in communal relationships where benefits are distributed according to partners’ needs. As a result, we expect exchange oriented consumers to keep close track of specific and concrete attributes of the brand since only that can enable them to track the balance between inputs and outcomes from brand consumption. Additionally, we expect communal consumers not to do such a close scrutiny of specific brand attributes but instead to attend more to information about the abstract brand benefits to assess if the brand satisfies their needs.

Across three studies we test the overall hypothesis that consumers with a communal orientation towards a brand are more likely to attend to abstract brand information relative to exchange consumers who attend to every nitty-gritty detail. All three studies are scenario-based laboratory experiments.

In Study 1, relationship norms were first manipulated using an inter-personal interaction situation. Participants then read a short passage about a fictitious clothing store, described using both concrete as well as abstract information. Later, the participants took a four-alternative multiple-choice recognition test. There were six questions in all each with four choices consisting of a correct concrete, correct abstract, plausible inference, and incorrect response. Since we expect the exchange-oriented participants to attend more to concrete brand details, we hypothesized such participants to show higher rates of acceptance for correct concrete responses, and higher rejection of incorrect responses relative to communal participants. Results supported this prediction. In addition, we hypothesized that since communal participants attend to abstract brand information, they would show higher rates of acceptance of correct abstract responses relative to exchange participants. This hypothesis was not supported. There are two alternative explanations for these results. First, it was possible that contrary to our

45

prediction, an exchange orientation does not make people differentiate between concrete and abstract brand information. Instead, this orientation prompts individuals to attend to all brand information equally deeply. Alternatively, it is possible that exchange participants do indeed attend primarily to concrete brand information at encoding, but use their specific brand knowledge to correctly construct responses about abstract brand information.

Study 2 explored these two alternative hypotheses with the help of an additional measure, namely, recognition response latencies. We reasoned that since recognition requires search and retrieval from memory, if respondents are actively constructing the abstract brand information, the time required to complete such a task would be long, and should reflect, to a large extent, memory construction. However, if the abstract brand information had been encoded as such, and not based on construction, then participants should respond quickly. Thus, we expected the response time measure to be particularly revealing of information processing strategies in that communally oriented individuals relative to their exchange counterparts, were expected to take less time to identify the correct abstract information and plausible inferences. Recognition performance replicated the results of Study 1 and the response latency measure confirmed our predictions: exchange orientation resulted in significantly slower responses to the abstract information and plausible inferences compared to a communal orientation. Together, these results suggest that exchange orientation prompt people to attend primarily to brand specific information whereas a communal orientation makes people attend mainly to abstract information about a brand. The first two studies manipulated the relationship norms using a scenario independent of the brand context. Even though such a manipulation provides a conservative test of the theory, we wanted to ratify the results with a brand-specific manipulation. In addition to this, in Study 3 we used a conjoint like experiment to evaluate people’s relative weights for abstract and concrete brand attributes. Results were consistent with the findings of the first two studies, that is, communal participants put greater weight (part-worth) on abstract but not on concrete brand attributes. In summary, results of the three studies offer insights about how the type of relationship, that consumers have with a brand, influences the way in which they process information about that brand. For further information, contact:

Prof. Sharmistha Law, 1265, Military Trail, Division of Management, University of Toronto, Scarborough, Ontario, Canada M1C 1A4. Phone: 416 287 7320, Email: law@rotman.utoronto.ca

46

2.3.2 CLASSICAL CONDITIONING OF IMPLICIT PRODUCT ATTITUDES

Bryan Gibson, Central Michigan University ABSTRACT Research within consumer psychology has historically focused on how conscious motivational states, attitudes and beliefs influence consumer behavior. Recent work in social psychology, however, has demonstrated that much social cognitive work is done at the implicit or nonconscious level. In other words, a good deal of thought is carried out outside of conscious awareness, and that this implicit cognition can subsequently affect behavior in important ways. These nonconscious effects can be internalized in the form of implicit attitudes. Implicit attitudes are unavailable to conscious awareness and consideration, but can influence the actions of the individuals possessing them. In the present study I integrate the concept of implicit attitudes into the consumer domain by developing an implicit measure of a product attitude (using standard Implicit Associations Test (IAT) methodology), and then comparing and contrasting how both this implicit attitude and a more traditionally measured explicit attitude respond to a classical conditioning procedure. I hypothesized that 1) explicit attitudes would be unaffected by the conditioning procedure; 2) implicit attitudes would shift toward the positively conditioned product, but only for those participants whose attitude was initially neutral; and 3) contingency awareness would not be necessary for conditioning to occur. A 2 (Conditioning: Favor Coke vs. Favor Pepsi) x 3 (Initial Attitude: Favor Coke vs. Favor Pepsi vs. Neutral) design was used to examine how implicit and explicit product attitudes are affected differently by classical conditioning procedures. Participants (N = 47) were pretested regarding their preferences for Coke and Pepsi, and then completed a computer based conditioning procedure that either paired positive images (e.g., a field of flowers, a smiling couple) or negative images (e.g., exhaust from a car, a pile of trash) with images of either Coke or Pepsi. Following the conditioning procedure, participants completed the Coke-Pepsi IAT, explicit measures of attitudes towards Coke and Pepsi, and a measure of contingency awareness. A 3 (prefer Coke, prefer Pepsi, neutral) x 2 (Coke positive, Pepsi positive conditioning) ANOVA was carried out on both the IAT and explicit attitude measures. For the IAT variable,

there was a significant main effect for prior attitude toward the products. Those having more favorable prior attitudes toward Pepsi had an IAT score more favorable toward Pepsi (M = .515) and those having more favorable prior attitudes toward coke had an IAT score more favorable toward Coke (M = -.324). Participants who were neutral had an IAT score falling between these two scores (M = .026). Of primary interest, however, was how the conditioning procedure affected participants in each of the pretest conditions. For those with an initial preference for Pepsi, their IAT was not affected by the conditioning (M = .496 vs. M = .533). For those with an initial preference for Coke, their IAT was not affected by the conditioning (M = -.410 vs. M = -.237). Participants who were initially neutral, however, were affected by the conditioning procedure (M = .303 vs. M = -.251, t(14) = 1.92, p < .05, one-tailed). Those conditioned to prefer Coke had a stronger preference for Coke, those conditioned to prefer Pepsi had a stronger preference for Pepsi. A similar analysis was carried out on the explicit measure with the pretest explicit attitude serving as a covariate. This analysis showed no significant effects, all F’s < 2.50, all p’s > .12. Results indicate that there was a significant effect of the conditioning variable, F(1,41) = 59.56, p < .001. This indicates that participants did have some degree of accuracy in their memory of the contingency between the CS and US. Follow up tests show that within the Coke positive condition, those who were initially neutral were significantly worse at identifying the CS-US contingency (M = -.48) than were those who preferred Coke (M = -.97) or Pepsi (M = -.73). Similar trends emerged for the Pepsi positive condition, but did not reach significance. The current research shows that for well known products, implicit attitudes are more easily conditioned than explicit attitudes, but only for people who were neutral about the products initially. A number of important follow up questions arise from this research. First, the findings highlight mechanisms that could lead to discrepant implicit and explicit attitudes, but does not address how these attitudes may conflict with each other, or how they may eventually align with each other. Second, the present research does not address how implicit attitudes influence behavior

47

toward the product. Some research in social psychology suggests that implicit attitudes have more direct influence on behavior that is relatively quick, automatic, and not well thought out. This suggests that implicit product attitudes may play a greater role in product choice when the consumer is distracted, or is making an impulse purchase. These are situations in which there is less chance for cognitive variables to intervene, leading the consumer to be guided by an implicit attitude. Of course, in many, if not most cases, the implicit and explicit attitudes may be the same. As this study demonstrates, however, implicit and explicit product attitudes need not always match. For further information contact: Bryan Gibson, Department of Psychology, Central Michigan University, Sloan Hall, 101, Mt. Pleasant, MI 48859.

48

2.3.3 THE INFLUENCE OF NEED FOR COGNITION ON TEMPORAL CONSTRUAL

Kiwan Park, University of Michigan Joseph R. Priester, University of Michigan

ABSTRACT

The Temporal Construal Theory (Trope &

Liberman, 2003) posits that future events are differentially construed, and thus evaluated, as a function of whether they are to be experienced in the near or distant future. In the distant future, events are construed at a general, abstract level whereas in the near future, events are construed at a specific, concrete level. The Temporal Construal Theory, however, is silent regarding the boundary conditions under which its explanation works. In this paper, we explore the question of when temporal distance influences evaluation, comparing three alternative hypotheses: default linkage hypothesis, non-thoughtful hypothesis, and thoughtful hypothesis.

By introducing the construct of elaboration as a moderator, we test which of the three competing hypotheses emerge. The default linkage hypothesis suggests that elaboration should not moderate the findings predicted by the Temporal Construal Theory, the non-thoughtful hypothesis suggests that the findings predicted by the Temporal Construal Theory should hold for the low rather than the high elaboration conditions, and the thoughtful hypothesis suggests that the findings predicted by the Temporal Construal Theory should hold for the high rather than the low elaboration conditions.

We conduct an experiment to examine the effect of temporal distance on the preference of work-study options (Trope & Liberman, 2000, study 2). The stimuli concern work-study options that students may take in the future. One option offers an interesting main task with an uninteresting training, whereas the other offers an uninteresting main task with an interesting training. A main task is presumed to be more central, primary, goal-relevant than a training. Trope and Liberman (2000) found that the preference for the former were evaluated more favorably in the distant future than in the near future, whereas the latter were evaluated more favorably in the near future than in the distant future. We test how temporal construal effect emerges differentially under conditions of high (vs. low) elaboration. In the present study, we measure the degree of elaboration with the Need for Cognition scale. Prior research has demonstrated that individuals high in need for cognition are more

likely to elaborate information than individuals low in need for cognition. (e.g., Cacioppo et al., 1996).

As predicted by the thoughtful hypothesis,

a three-way interaction of Elaboration X Time X Option emerged as significant, F(1, 154) = 3.38, p = .05. Follow-up analyses showed that, under conditions of high elaboration, two-way interaction of Time X Option was significant, F(1, 75) = 11.5, p = .01, indicating that the preference for the interesting main task over the uninteresting main task is more pronounced in the distant future, compared to the near future. This result is compatible with the explanation of the Temporal Construal Theory. In contrast, under conditions of low elaboration, the two-way interaction of Time X Option is not significant, F <1, NS.

This study provides a first moderator to the Temporal Construal Theory. The results demonstrate that temporal distance only influences evaluations under conditions of high elaboration. Specifically, although the results predicted by the Temporal Construal Theory emerge for individuals high in need for cognition, they do not emerge for individuals low in need for cognition. The present research also suggests that although the individuals high in need for cognition are likely to elaborate regardless of temporal distance, the nature, and thus outcome, of this elaboration differs as a function of whether the future event is described as being near or far.

For further information contact: Kiwan Park, PhD Candidate in Marketing, University of Michigan Business School, 701 Tappan Street, Ann Arbor, MI, 48109-1234. Phone: (734) 647-7667, Email: kiwanp@umich.edu

49

2.4.1 THE MODERATING IMPACT OF AGING ON RESPONSES TO EMOTIONAL ADVERTISING APPEALS

Patti Williams, University of Pennsylvania

Aimee Drolet, UCLA ABSTRACT Increasingly, researchers recognize the growing size and economic importance of the older (age 65 and over) consumer segment (e.g., Yoon 1997; Cole and Gaeth 1990). Accordingly, interest in and research on the psychology of older consumers has steadily increased. Much of this research has focused on detriments in information processing due to aging (for a review, see Roedder-John and Cole 1986). Past research has explained changes in how older consumers process information primarily in terms of the deleterious effects of aging on cognitive ability. However, in addition to cognitive effects, aging has motivational effects that can also change how individuals process information. Specifically, aging increases the motivation to attend to emotional versus non-emotional (i.e., factual) information (e.g., LaBouvie-Vief and Blanchard-Fields 1982). One explanation for older adults’ increased focus on emotional information has to do with their chronic perceptions of time as limited (Carstensen 1992). That is, according to socioemotional selectivity theory, individuals tend to assess time as either limited or expansive (Carstensen 1992), and such assessments produce motivational changes in information processing goals. In particular, when individuals view time as limited they tend to be more present-oriented, which is associated with finding satisfaction in the moment, and thereby devote greater attention to social connectedness, feeling states, and deriving emotional meaning, all of which can be experienced and enjoyed now rather than in the long-term (Carstensen, Isaacowitz, and Charles 1999). In contrast, when individuals view time as expansive, they tend to be more future-oriented, which is associated with more attention to planning, being analytical, and pursuing knowledge so as to be prepared for the future. We conducted four experiments to examine the impact which aging has upon the processing of

emotional advertising appeals. In three experiments we compare preferences among older versus younger adults for emotional versus rational appeals. In a fourth experiment, we look at responses to different types of emotional appeals. Our first experiment demonstrates the influence of aging and associated time horizon perceptions (e.g., as limited vs. expansive) among both young and older adults on their attitudes toward emotional and rational appeals for a real brand. This experiment shows results similar to those found in Fung and Carstensen (2003), namely that older adults prefer emotional appeals whereas younger adults prefer rational appeals. However, in addition, we show the moderating effects of aging for real ads and also on explicit recall of emotional versus factual information. In experiment 2, we examine a potential boundary condition of the results in experiment 1 by looking at appeals for both emotional (e.g., hedonic) and rational (e.g., utilitarian) products. Researchers have considered product category types according to their emotional (e.g. hedonic) versus rational (e.g. utilitarian) properties (e.g., Ratchford 1987; Holbrook and Hirschman 1982), and suggest that this distinction may make emotions more or less relevant in evaluations (e.g., Pham 1998). Results show that older adults have more positive attitudes toward and better recall for the emotional appeals relative to the rational appeals, regardless of the underlying product category type (emotional or rational). In experiment 3, we seek evidence that age-related differences in time horizon perceptions indeed drives the results observed in experiments 1 and 2. We manipulate perceptions of time horizon within emotional and rational appeals (rather than before appeals are presented) for both older and younger adults. We also include a no time horizon control. Last, we collected recall measures. Results show that older adults in both the control and limited time horizon conditions prefer and recall more from emotional relative to rational ads, replicating the results in experiments 1 and 2. Similarly, young adults in both the control and expansive time horizon conditions prefer and recall more from

50

rational (vs. emotional) appeals. However, among older adults in the expansive time horizon condition, we see a reversal of the previous results. Older adults in an expansive time horizon condition prefer and recall more from rational, relative to emotional appeals. We find a similar reversal among young adults in the limited time horizon condition, in that they now prefer and have better recall for emotional, relative to rational appeals. This experiment offers important insight into motivational differences among young and older consumers that are brought on at least in part by differing perceptions of time horizon and that impact the processing of and then preferences for emotional versus rational messages. Finally, in experiment 4, we examine the degree to which aging moderates preferences for emotional appeals that focus on achieving purely positive emotions through consumption of the advertised product, versus those that focus on avoiding negative emotions through consumption of the product. Results show that in the control and limited time horizon conditions, older adults prefer appeals that focus on avoiding negative emotions (relative to purely positive emotions), consistent with an antecedent-focused emotional regulatory strategy. In contrast, in the control and expansive time horizon conditions, younger adults prefer appeals that focus on achieving positive emotions (relative to those that focus on avoiding negative emotions). Again, we find evidence of the moderating impact of aging on responses to emotional appeals and also show that these differences are related to motivational tendencies arising from differing perceptions of time. For further information contact: Patti Williams, The Wharton School, University of Pennsylvania, 3730 Walnut Street, Philadelphia, PA 19104-6371, e-mail: pattiw@wharton.upenn.edu.

51

2.4.2 TRYING TO DISCREDIT INFORMATION CAN MAKE IT SEEM TRUE TO OLDER

ADULTS

Ian Skurnik, University of Toronto Carolyn Yoon, University of Michigan

Norbert Schwarz, University of Michigan

ABSTRACT A crucial task for consumers is to distinguish true and false information, both when they initially encode it and when they subsequently try to remember it. The problem of distinguishing true from false information is especially important because often people show a bias to think that information that they have seen before is true. A single presentation can enhance later perceived truth of a claim, even when the claim was previously identified as false. Although people have many motivational reasons to ignore or counterargue warnings about false information, this bias toward truth arises when people are trying to remember claims as accurately as possible. Drawing on psychological research in source and recognition memory, we suggest that this “illusion of truth” effect is greatest when (a) the claim seems familiar due to the prior exposure, but (b) details of the original truth-specifying context cannot be recalled. This reasoning suggests that we may become increasingly susceptible to the illusion of truth as we age. Normal human aging is associated with declines in memory for context or source, whereas experienced familiarity is largely unaffected by age. In daily life people often hear information many times before finding out definitively whether the information is true or false. Advertising claims and rumors can spread quickly before independent reviews of products are available. Through repeated exposure, people can become highly familiar with consumer claims before the claims can be verified. On one hand, repeated exposure to a claim may give people greater opportunity to elaborate and remember it more accurately, making eventual warnings about falseness more effective. On the other hand, our conception of the illusion of truth suggests that repetition may enhance people’s sense that they have seen the claim before, without also enhancing their memory for any truth-specifying context. Our prediction is that trying to discredit already familiar information can backfire and increase belief in truth. Given the age-related pattern of memory deficits mentioned above, this

tendency for warnings to backfire should be greater for older than for younger adults. We investigated the illusion of truth for claims about health and medicine. Adults over 65 years of age are the greatest users of medical services and the fastest growing segment of the North American population. Medical misinformation is widely available, and acting on such information can have dire consequences for the well-being and financial resources of older adults and their families. In addition, there are well-documented trends for patients to take complementary and alternative treatments, such as herbal supplements, despite warnings against such treatments from their physicians. Older and younger adults participated in a study of “medical information processing” by studying 36 individual claims (e.g., “DHEA supplements may lead to liver damage, even when taken briefly”). Half of these claims were presented once, and half were presented three times. All claims were identified as true or false (claims were extensively pretested to equate knowledge across age groups and to minimize certainly about truth value; for ethical reasons all claims were objectively true according to the US NIH). The timing of this true/false identification was manipulated between subjects. For half the participants, the truth value of each claim was identified immediately with each presentation of each claim. For the other half, the truth value was identified only on the last presentation of each claim – in other words, in this condition participants saw repeated claims twice before finding out whether they were true or false. After 10 minutes, all participants saw the list of claims again with new claims mixed in, and indicated whether each statement was true, false, or new. The experiment had a 2 (age: older or younger adults) x 2 (truth revealed at study on every or on last presentation) x 2 (1 or 3 presentations at study) factorial design, with the last factor within subjects. The experiment is complete and all data have been analyzed. For this procedure, the illusion of truth occurs when participants are more likely to call

52

originally false sentences “true” than to call originally true sentences “false.” A 2 (response type: “true” to false or “false” to true) X 2 (age) X 2 (number of presentations) X 2 (truth each or last presentation) mixed ANOVA on the responses revealed a number of significant effects. When truth of claims was identified on each presentation, repetition of claims helped people avoid errors in remembering truth (relative to once-presented claims), with older adults making more errors overall. When truth of claims was identified on the last presentation only, repetition actually increased the likelihood of misremembering false claims as true (relative to once-presented claims), and this backfire effect was greater for older than for younger adults. Older adults mistakenly thought that 38% of the repeated false claims were true when the claims were familiar before being discredited, compared with 12% when the claims were identified as false on every presentation. There was no parallel pattern in the tendency to misremember true information as false. We highlight the risks of communication strategies that merely identify a given claim as unsubstantiated or false. For example, attempts to warn people about an erroneous, outdated, or deceptive claim may increase belief in the claim by raising its familiarity. Older adults may be particularly susceptible to this unintended effect. We discuss potential implications for public policy. For further information contact: Ian Skurnik, Rotman School of Management, University of Toronto, 105 St. George Street, Toronto, Ontario M5S 3E6, Canada, e-mail: iskurnik@rotman.utoronto.ca.

53

2.4.3 THE SHRINKING DECISION PROCESS OF OLDER CONSUMERS: ANOTHER AVENUE TO REPEAT PURCHASE

Raphaëlle Lambert-Pandraud, Negocia

Gilles Laurent, HEC ABSTRACT Marketing practitioners often report more frequent repeat purchase by older consumers (e.g., Secodip, 1998). However, few research studies have tried to examine this phenomenon. In the present study, we employ a survey methodology to investigate the following three questions: 1) to what extent do consumers have a greater tendency, as they age, to re-purchase the previous brand; 2) are there characteristics related to their purchase processes, such as the tendency to re-purchase from the same store, or to consider fewer alternative brands; and 3) what are the possible reasons underlying such behaviors? In order to address these questions, we analyze survey data from 31,497 recent buyers of a new car (a representative sample of the population of new car buyers in France over a one-year period). We find, among older purchasers, elevated rates of repeat-buying. We also find, more broadly, evidence of a general “shrinkage” of older purchasers’ decision processes. In particular, we find support for the view that an older consumer’s decision process is more focused on the previously purchased brand. Older buyers (60 to 74 years), and particularly those who are very old (over 75 years), are more likely to consider the previously purchase brand and to repurchase it. This reflects an extreme tendency by older groups, compared to their younger counterparts, to limit their consideration to only the previous brand; and the propensity holds even when satisfaction with the previous car is controlled for. Furthermore, an older buyer is more likely than a younger buyer, to purchase the car from the same dealer from whom the previous car was purchased. Additional analysis suggests that consideration sets are relatively narrow for older purchasers. Older buyers (and especially very old buyers), compared to younger buyers, not only consider fewer brands and models, but visit fewer dealers. This includes a higher frequency among older buyers to consider a

single brand and a single model, and to visit only one dealer. Finally, we find that when older buyers (and especially very old buyers) switch away from their previous brand, they are likely to consider and switch their purchase to well-established national brands rather than newer import brands. We consider a number of possible explanations to account for our findings. We examine three major explanations in greater detail. One explanation is based on cognitive views of aging, that repeat purchase is a consequence of reduced cognitive abilities with age (Chasseigne et al., 1997). We suggest that a consumer repurchases the same brand by default, because his memory limits the consideration set to the previous brand, or to the previous brand plus another known brand. An elderly consumer may also no longer be able to evaluate several options in great detail. Declines in memory and fluid intelligence are likely to decrease the size of the evoked set, and to increase the tendency to consider familiar brands, which in turn induces repeat purchase (Lesser and Kunkel, 1991). Older adults tend to rely more on decision heuristics, which orient brand choice at the consideration stage without intensive search and evaluation. Another explanation for repeat purchases by older consumers relies on notions of risk aversion among the elderly. Aversion to risks that are associated with changes, even if the present situation is far from ideal, is a well-documented phenomenon in gerontology (Botwinick, 1978). Older persons are thus viewed as more cautious in decision-making, which is likely to be reflected in preference for a risk-reducing strategy such as repeat purchasing. A third possible explanation involves the view that individuals gain increased consumer expertise with age (Alba and Hutchinson, 1987). Based on this perspective, we suggest that accumulation of purchase and usage experience with a product category should lead to greater expertise about the products in the category, and to faster decision processes. However, automobile purchases tend to occur at long intervals, typically several years, during which the models available in the market change. As such, the added experience that older

54

consumers bring to bear on automobile purchases may not be useful in their decision processes. We thus analyze data from a separate sample to examine the effects of expertise on repeat purchases. Finally, we discuss some implications of our research findings for public policy. Given the older consumers’ reduced cognitive abilities and increased risk aversion, we discuss specific supports that can be provided to help them in the purchase decision process. For further information contact: Raphaëlle Lambert-Pandraud, Negocia, 8 avenue de la Porte de Champerret, 75838 Paris, France, e-mail: rpandrau@negocia.fr

55

3.1.1 REGULATORY FEELINGS AND THEIR RELATIONSHIPS TO BEHAVIOR: THE CASE OF HUNGER

Laurette Dubé, McGill University

Catherine Paquet, McGill University

ABSTRACT In the old psychological terminology, the bodily need detection systems of the brain were thought to generate “drives,” which were viewed as powerful affective responses motivating decision and action tendencies necessary to solve regulatory imbalances. Progressively, it was recognized that, at a logical level, the notion of drive and behavior may be redundant for a coherent explanation of behavior and that the operation of specific regulatory feelings that could be linked to specific unitary brain processes could be better candidates for intervening variables between the bodily need and its satisfaction (see Panksepp 1998). Regulatory feelings are bodily states subjectively experienced with intense regulatory imbalances and the specific sensations that accompany the satisfaction of bodily needs -- the hungers, thirsts, cravings, and the various sensory delights that arise from interacting with needed resources. In a way, these may be seen as key intermediaries between physiological and other bodily states and the behavior they necessitate. Feelings of hunger are regulatory feelings tied to the energy-balance regulation. Models of hunger, derived from animal and human research, show that feelings of hunger arise in response to a diversity of bodily signals, as part of complex short-term feeding control mechanisms. These signals include: (1) oral factors; (2) a large array of stomach and gastrointestinal factors that act upon brain mechanisms; (3) metabolic processes that originate from various body compartments, especially the liver, which are reflected in circulating nutrients, some of which affect the brain; (4) diverse neurological factors; and (5) a portfolio of non-specific influences that have little to do with energy regulation, such as feelings of sickness or malaise, as well as emotional and mood changes, or variation in social, functional or environmental contexts in which feeding behaviors are performed (Panksepp 1998). In other words, hunger is a regulatory feeling associated with, but not always determined by, the bodily state of nutrient depletion or repletion (deCastro and Elmore 1988). However, comprehensive empirical

evidence in humans still remains to be obtained, in particular for signals 4 and 5. In mapping the psychological reality of hunger and its behavioral impact in the consumer domain, different issues are of both of theoretical and practical relevance. First, it is necessary to understand better the mechanisms by which hunger influence food behavior. Thus far, research in humans and animals have shown that there is a causal relationship between feelings of hunger and food behavior. In parallel, it is also well established that hunger influences various facets of sensory perceptions, cognitions, and evaluations related to sensory qualities of food. Since there is also robust evidence to support that sensory qualities of food contribute to food behavior, it is intuitively compelling to suggest that hunger influence food behavior via its coloring of sensory qualities of food. However, to our knowledge, a strict test of such mediating effects is yet to be provided. Moreover, because regulatory feelings like hunger are intimately associated with physiological and neurological systems, it is likely that residual direct effect on behavior persists, beyond the mediation of sensory perceptions. Second, if hunger is influenced by variations in mood, emotions and other internal psychological states, as well as by contextual variations in external conditions in which feeding occurs, it would be important to estimate these influences and delineate their boundaries in order to provide marketers with insights to help design more effective product/service and communication strategies. In the present paper, we first review the existing literature on hunger and then report two studies designed to explore these issues. Study 1 consisted of a longitudinal field study in which elderly consumers (21 women; 11 men; average year of 80), clients of institutional foodservices, were observed on a large number of meals (average of 46.2 ± 14.6 meals/participant), self-reporting the intensity of pre-meal feelings of hunger and contemporaneous emotional states, as well as post-meal evaluation of sensory food qualities. Food behavior was derived from observation of plate leftovers in relation to standardized recipes and portion size (energy and

56

protein content). The objective was to test the direct and indirect (via its impact on perceptions of sensory qualities of food) effects of hunger on behavior and explore its emotional correlates. The direct and mediated effects of hunger on food intake were assessed through mediation analyses, where hunger intensity is considered as independent variable, food behavior as the dependent variable, and sensory quality evaluations as potential mediators. This requires the estimation of a series of regression analyses. Given the multi-level nature of our data, we used a random coefficient approach with random intercepts and slopes using the SAS Proc Mixed procedure. Following estimation of relationships between hunger and food behavior and between sensory qualities and food behavior, subsequent analysis revealed that hunger remained a positive and significant (beta=.006, p<.0001) predictor of food behavior when palatability of food was added as predictor. The sobel test was significant (2.81; P=.005), which suggests that the mediated effect is not null. However, the effect of hunger remained significant, indicating that the mediation is only partial. Hence, hunger seems to influence food intake directly, but also directly via its effect on palatability. Further analyses revealed that both positive (direct relationship) and negative (reverse relationship) were significant correlates of hunger intensity. The second study consisted of a laboratory experiment designed to test research hypotheses bearing on the effects of variations in the consumption context (i.e., sensation-focused vs. distraction) on changes in the intensity of feelings of hunger resulting from the same objective consumptive behavior. Building on research demonstrating the “appetizer” effect tied to a conditioned response to the taste of food (LeMagnen, 1992; Powley, 1977; Panksepp, 1998), we anticipated that the changes in feelings of hunger after the consumption of a given amount of food will be lesser in sensory-focused contexts than under distraction conditions. Participants were given to eat the same quantity of chocolate in one of two consumption context conditions, namely, sensation focus (asked to eat while pay attention to taste, color, texture, mouth-feeling, etc.) and distraction away from sensation (asked to eat while performing competing task). Pre- and post-consumption feelings of hunger were measured. ANCOVA with repeated analysis measurements was performed to test for the effect of contextual manipulation in consumption context on changes in feelings of hunger as well as to explore the

moderating effect of contemporaneous mood (positive emotions), as well as individual differences in physical (BMI) and psychological (predisposition toward eating, restraint, emotional, situational). Results were consistent with expectations for the effect of contextual manipulation on changes in hunger intensity. BMI and situational eating predisposition emerged as individual-level moderator. Theoretical and managerial consequences of the results of both studies are discussed. For further information contact: Laurette Dubé, Faculty of Management, McGill University, Montreal, Quebec, Canada, H3A 1G5, email: laurette.Dubé@mcgill.ca

57

3.1.2 EFFECTS OF DISTRACTIONS WHILE TASTING A FOOD SAMPLE: THE INTERPLAY OF INFORMATIONAL AND AFFECTIVE COMPONENTS IN SUBSEQUENT CHOICE

Baba Shiv, University of Iowa

Stephen M. Nowlis, Arizona State University

ABSTRACT Research in the domain of pain suggests that attending to the experience of pain relative to being distracted from it decreases its judged unpleasantness. In this research we ask if a parallel process occurs with pleasant experiences. Specifically, will distraction (versus paying attention) while tasting a food item like chocolate or ice-cream increase choice of the sampled item as suggested by research on pain? If findings in the domain of pain apply to the domain of pleasure as well, an implication would be that marketers ought to distract consumers rather than have them pay attention during the tasting experience. This is because the basic findings in research on pain would suggest that the intensity of pleasure and, thus, the strength of subsequent preferences for the sampled food item, is likely to be higher when consumers are distracted than when they are paying attention. In this research, we first present the underlying psychological mechanisms that have been proposed in the literature to account for findings in the domain of pain. We then develop a two-component model (TCM) for pleasurable somatosensory experiences such as tasting a food sample, should the predictions in the domain of pain apply to the domain of pleasure as well. According to the TCM, somatosensory experiences of tasting food items like chocolate (generally a pleasurable experience) is composed of two components. One component is affective in nature (i.e., involves emotional reactions arising from the sensory experience) and is associated with automatic processes. The second component is informational in nature (for instance, its health-consequences, the quality of the chocolate, its sweetness and texture, the spatio-temporal characteristics of the tasting experience, etc.). This informational component is associated with controlled processes and can be assumed to be less positively valenced than the affective component. If individuals are distracted while tasting chocolate, the affective component is likely to serve as the predominant input to the intensity of pleasure that is ultimately experienced, and, therefore, to the subsequent preferences for the

sampled chocolate. On the other hand, if individuals are not distracted while tasting chocolate, both the less positively valenced informational component and the more positively valenced affective component are likely to serve as inputs to the intensity of pleasure that is ultimately experienced. These inputs, in combination, are likely to be less positive than the affective component that is likely to predominate when the level of distraction is high. Paradoxically, therefore, if the predictions arising from research on pain are applicable to the domain of pleasure as well, higher levels of distraction ought to lead to more favorable preferences for the sampled chocolate than lower levels of distraction. We carried out several experiments that served to demonstrate the basic effect of distraction while tasting a food sample on subsequent choice of the tasted option. The experiments also served to examine whether the TCM, which has been used to account for the findings in research on pain, would apply to the domain of pleasure as well. To meet this goal, we delineated different moderators consistent with the TCM in the various experiments. For example, in one experiment, we tested the TCM by varying levels of distraction and by varying the valence of the informational and the affective components. In some conditions, participants tasted a milk-chocolate with a positively valenced affective component and a less positively valenced informational component. If the TCM is valid, then the basic pattern of results predicted to arise from research on pain ought to obtain in these conditions—higher choice of the tasted sample when the level of distraction is high than when it is low. In another set of conditions, participants tasted a soy-chocolate with a less positively valenced affective component and a more positively valenced informational component. If the TCM is valid, then the pattern of results ought to reverse in these conditions—lower choice of the tasted sample when the level of distraction is high than when it is low. In a second experiment, we tested the TCM by varying the intensity of the affective component by having participants taste a milk chocolate after tasting a salty or a sweet cracker. Prior research suggests that our taste-buds

58

become more (less) sensitized to sweetness following experiences of salty (sweet) food items. Therefore, if the TCM is valid, then the basic pattern of results predicted to arise from research on pain ought to obtain in the salty-cracker conditions—higher choice of the tasted chocolate when the level of distraction is high than when it is low. This pattern ought to attenuate in the sweet-cracker conditions. In a third experiment, we tested the TCM by measuring participants’ emotional reactions to the tasting experience and then examining whether these measured emotions mediate the effects of distraction on subsequent choices. Across the various experiments, we demonstrate that the basic effect of distraction while tasting a food sample on subsequent choice is consistent with predictions arising from research on pain. We also provide support to the TCM and rule out various alternative accounts. We conclude by discussing the implications of our findings and presenting several directions for future research. For further information contact: Baba Shiv, Tippie College of Business, University of Iowa, S336 Pappajohn Bldg, Iowa City, IA 52242-1000, email: baba-shiv@uiowa.edu

59

3.1.3 ON THE LIMITATIONS OF PLEASURE INTENSITY IN SHAPING BEHAVIOR OVER

TIME

Jordan L. LeBel, Concordia University ABSTRACT Previous research holds that pleasure serves as a decision aid, underlying adaptive and efficient decision making. Between two options, humans, it seems, are hard wired to select the one procuring greater pleasure. In other words, the more intense the pleasure, the more likely is the behavior. This may hold when pleasure and approach tendencies are considered broadly between two stimuli but some evidence suggests that, as a consumption experience unfolds over time, pleasure intensity or reward value can change quickly and influence consumption decisions and behaviors in directions unforeseen by traditional approach-avoidance and utility theories. In this paper, I examine the influence of changes in pleasure intensity on consumption behavior within episodes where consumers have repeated opportunities to stop or continue consumption. Research on intertemporal choice and decision making has focused on the relationship between specific moments and patterns within a consumption episode and its overall attractiveness (anticipated or retrospective). Moreover, research has typically relied on unpleasant stimuli (e.g., movie clips of amputation; high pitched noise), assuming that findings would transpose in the domain of pleasurable experiences. As a result, we know far less about the evolution of initial moments and how within-episode changes in hedonic intensity influence behavioral responses. Scattered evidence suggests that within-episode changes in hedonic reactions may be inversely related to the intensity of the initial hedonic response at the onset of consumption. For instance, opponent process theory supports the notion that repeated exposures will produce an appetizing effect on an initially low hedonic response leading to increases over time but will cause decreases to initially high responses. If such a temporal pattern were supported within a single consumption episode, it is conceivable that a stimulus of initially high intensity may not be associated with more consumption than a stimulus of initially lesser pleasure intensity. Further, at any given moment, individual and situational factors may bear on one’s subjective experience, influencing both the

intensity of hedonic reactions and behavioral responses. In this presentation, I report the results of two studies designed to explore the influence of initial pleasure intensity, its evolution, and its relationship to behavioral responses associated with the consumption of chocolate within experiences that unfold over time. Study 1 was designed to test, in a naturalistic context, the possibility that initial hedonic reaction to a stimulus can change dramatically within a relatively short time span and impact behavior in unexpected directions. Pleasure intensity was manipulated on a within-subject idiosyncratic basis and participants were presented with a chocolate of different intensity level on each of three separate 90-minute episodes where moment-to-moment hedonic reactions and consumption were monitored. Analyses accounted for individual-level factors including Body Mass Index (weight/height2), sensation-seeking tendency, hunger, mood, and restraint in eating behaviour. Results revealed three important findings. First, even while accounting for psychological (mood) and physiological (hunger, BMI, total intake) factors, different levels of pleasure intensity experienced at the onset of each consumption episode were associated with different magnitude of change in hedonic reactions over the entire episode. Specifically, higher initial pleasure intensity (as measured immediately after the first unit of consumption) was associated with a greater decrease in hedonic response from first to last unit. Lower pleasure at the onset was associated with very minimal change and moreover, case-by-case analyses revealed that for nine participants low intensity pleasure actually increased from the first to the last unit of consumption, in one case by as much as 6.1 out of 10. Secondly, results also showed that successive pleasure ratings possessed a first-order Markov property such that pleasure experienced at any given moment was anchored and adequately predicted by the immediately preceding hedonic response and not dependent on every preceding moments in the episode. Finally, analysis of covariance revealed that total consumption was not influenced by the level of pleasure experienced early on in the consumption episode. Consistent with previous research on ingestive behavior, significant covariates revealed

60

that higher levels of BMI were associated with greater consumption and pleasant pre-consumption mood was associated with lower consumption. Further evidence of satiating power of high pleasure intensity is the fact that lower consumption was associated with lower higher pleasure ratings at the last unit of consumption. Case-by-case analyses revealed that for eight participants, the amount consumed in the low intensity condition was greater or equal to consumption in the high intensity condition. Study 2 was conducted in the laboratory and the interval of time between consumption units was fixed to three minutes. Eighty-five participants were asked to consume six five-gram squares of cinnamon-favored milk chocolate while their attention was directed either toward sensory experience or on an unrelated task. Pleasure, desire to consume another unit immediately, and willingness to pay for that next unit were measured after each square was eaten. Participants whose attention was directed on the unrelated task experienced a greater decrease in hedonic reaction from the first to the sixth unit of consumption, compared to those who remained focused on sensory experience. In the task-focus condition, the average ratio of the pleasure response to the first unit divided by the pleasure response to the last unit was 1.38, while this ratio was 1.03 in the sensation-focus condition. Thus, hedonic reactions were more stable and more intense throughout the consumption episode when participants’ attention was focused on their sensations. Further, the relationship between pleasure intensity and the two decision making indicators (i.e., desire to eat and willingness to pay for the next piece) was also sensitive to the manipulation of attentional focus. Results show that the strength of the relationship between sensory pleasure intensity and decision-making (desire to eat and willingness to pay for next chocolate piece) did not vary over time in contexts where subjects focused on an unrelated task (correlations remained in the sixties for desire to eat and remained NS for willingness to pay). In contrast, the strength of these relationships increased over time in contexts that focused on sensory experiences (correlations went from .62 to .84 for desire to eat; from NS to .31 for willingness to pay). For further information contact: Jordan L. LeBel, John Molson School of Business, 1455 de Maisonneuve Blvd. West, Montreal, Quebec, H3G 1M8, email: jlebel@jmsb.concordia.ca

61

3.2.1 COMPARATIVE ADVERTISING: EFFECTS OF ATTRIBUTE TYPICALITY AND STRUCTURAL ALIGNABILITY ON BRAND EVALUATION AND SIMILARITY JUDGMENT

Xiaoli Nan, University of Minnesota, Minneapolis, MN

Sela Sar, University of Minnesota, Minneapolis, MN ABSTRACT

This study investigates the differential effects of different types of comparative ads on consumers’ evaluation of the sponsoring brand and perception of the similarity between the sponsoring and comparison brands. The format of comparative ads may differ along two basic dimensions: 1) characteristics of the featured attribute and 2) the way in which comparison is made. With regard to the first dimension, the featured attribute may be typical vs. atypical of the associated product category. With regard to the second dimension, comparison can be made by either referring to an attribute both the sponsoring and comparison brands possess (e.g., brand A pain reliever relieves pain for 24 hours while brand B pain reliever relieves pain for only 12 hours) or by referring to an attribute that is unique to the sponsoring brand (e.g., brand A pain reliever contains vitamin C while brand B pain reliever does not). These two types of comparative ads differ along what’s called structural alignability, namely the ease with which the attributes of one object can be aligned or mapped onto another object (Zhang, Kardes, and Cronley 2002). Here, the first type may be called alignable comparative advertising while the second type may be called non-alignable comparative advertising.

This study found that when the featured attribute was typical of the associated product category, non-alignable comparative advertising was no more effective in improving evaluation of the sponsoring brand and in decreasing perceived similarity between the sponsoring and comparison brands than was alignable comparative advertising. On the other hand, when the featured attribute was atypical, non-alignable comparative advertising was significantly more effective in improving evaluation of the sponsoring brand and in decreasing perceived similarity between the sponsoring and comparison brands. For further information contact: Xiaoli Nan, School of Journalism and Mass Communication, University of Minnesota, 111 Murphy Hall, 206 Church Street S.E., Minneapolis, MN 55455, E-mail: nanx0005@umn.edu, Phone: (612) 626.0221

62

3.2.2 WHEN TWO BRANDS ATTACK EACH OTHER: A COMPARATIVE ADVERTISING

EXAMINATION

Nicole L. Votolato, The Ohio State University H. Rao Unnava, The Ohio State University

Payal Karumbiah, The Ohio State University

ABSTRACT

A comparative ad compares specifically named brands from the same generic product or service class on specific attributes (Wilkie and Farris 1975). Past research in comparative advertising has almost exclusively employed experimental designs in which subjects viewed either a comparative or a noncomparative ad, with accompanying filler ads. This research suggests that comparative ads are more effective than noncomparative ads. Interestingly, there is not much research that examines the relative effectiveness of comparative vs. noncomparative ads when two brands attack each other through comparative ads, which is the case most often in media advertising. For example, if Colgate claims they are better than Crest on certain attributes, and then Crest advertises that they are better than Colgate on certain other attributes, what is the net result of this type of a campaign? Are these comparative ads any better or worse than noncomparative ads with the same brand information? Also, it is not clear whether the effects of comparative advertising are similar for well-known vs. new brands. The objective of this research is to examine the relative effectiveness of the comparative vs. noncomparative advertising formats when two well-known brands or two new brands attack each other.

Pechmann and Ratneshwar (1991) have shown that a comparative ad can have a positive impact on nonfeatured attributes of an advertised brand and a negative impact on featured attributes of a comparison brand. Thus, a brand that compares itself to another brand should be successful in undermining the comparison brand on those attributes. For instance, if Crest compares itself to Colgate on certain attributes, consumers’ ratings of Crest on those attributes should be higher than the ratings they would give Colgate. This effect should not occur if Crest used a noncomparative ad. Consumers should rate Crest high on those attributes, but they have no reason to undermine Colgate on the same attributes. Therefore, consumers’ beliefs about featured attributes of a brand in a comparative ad would be more positive for the advertised brand and less

positive for the comparison brand. For a noncomparative ad, consumers beliefs about the featured attributes of the advertised brand will be positive, but no effect on the beliefs about the comparison brand are expected. Importantly, the lack of an effect on nonfeatured attributes should be more pronounced for known brands than unknown brands because consumers’ knowledge structures about known brands are better established. A 2 x 2 between-subjects design with a control group was used to test these hypotheses. The two factors examined were ad type (comparative or noncomparative) and brand type (known or unknown). Respondents in each condition were exposed to two advertisements from different brands in the toothpaste category. Both ads involved either well-known brands (i.e., Crest and Colgate) or unknown brands (i.e., Neem and Meswak). Both ads were either comparative or noncomparative. In the comparative ad conditions, one of the brands from each pair was the advertised brand and the other brand was the comparison brand. In the second comparative ad, subjects saw the same two brands, but their roles were reversed.

A pretest identified the three most frequently listed attributes for Crest and Colgate, and these were used in the corresponding advertisements. There were no overlapping attributes in the two advertisements. The Neem and Meswak ads were identical to the Crest and Colgate ads with the exception of the brand name. Attitudes towards each of the two brands of toothpaste and each of the two target ads were assessed. Beliefs about the six attributes seen across the two ads were obtained for each of the two brands. Cognitive responses were also collected. Finally, because of the use of two well-known brands, in the control condition attitudes and beliefs towards Colgate and Crest were obtained using the same scales as in the experimental conditions. We examined the effects of comparative vs. noncomparative advertising for known vs. unknown brands and for featured vs. nonfeatured attributes. When known brands were used, advertising type made no difference in subject’s beliefs for the featured and nonfeatured attributes

63

when known brands were used. The attribute ratings for Crest and Colgate were not different from each whether a comparative ad or a noncomparative ad was used. In contrast, for unknown brands, comparative advertising resulted in lower beliefs for featured and nonfeatured attributes than noncomparative advertising, for both brands. Thus, contrary to previous findings in the comparative advertising literature, subjects were less convinced by brand advertising when two new brands attacked each other through comparative advertising. The attitudes toward Crest and Colgate in the control condition support our initial thinking that Crest and Colgate were equally liked among the student population. In experimental conditions, there were no differences in brand attitudes whether the ads were comparative or not. While the absence of attitudinal differences between Crest and Colgate is understandable (because subjects may have relatively stable attitudes toward them), the absence of attitudinal differences for Neem and Meswak between comparative and noncomparative advertising are unexpected. Given the lower belief ratings for comparative ads for unknown brands, one would have expected them to affect attitudes toward the brand as well. While the attitudes toward Neem and Meswak in the comparative ad condition were lower than the attitudes in the noncomparative ad condition, this difference was not significant. Similar nonsignificant, but directionally consistent results were obtained for subjects attitudes toward the advertisements.

Our research addresses an important void in this literature – how consumers respond when faced with two comparative ads for brands attacking each other. The findings indicate that when two established brands attack each other, no additional benefit was gained by using the comparative vs. noncomparative advertising format. However, when two new brands attacked each other, use of the comparative advertising format penalized both brands. Compared to the noncomparative ads, the comparative ads resulted in less positive attribute beliefs for both brands. These results are contrary to predictions made by comparative advertising literature and suggest that the findings from past research may not be generalizable to practical situations in which two brands attack each other.

For further information contact: Nicole L. Votolato, Fisher College of Business, The Ohio State University, Fisher Hall 530, 2100 Neil Avenue

Columbus, OH 43210, votolato.2@osu.edu, (614) 292-4710

64

3.2.3 PURCHASING NOW, OR LATER: INTERACTIVE EFFECT OF TIME AND ATTRIBUTE ALIGNABILITY

Selin A. Malkoc, University of North Carolina at Chapel Hill

Canan Ulu, Duke University Gal Zauberman, University of North Carolina at Chapel Hill

ABSTRACT

In many consumer settings, purchase decision and the product experience are temporally separated. One of such situations takes place when the consumption of the product is delayed. Consumers often face situations where the purchase decision is immediate but the consumption of the product is in the future. For instance, summer outfits become available for purchase long before they can actually be worn, thus consumers purchasing these products, although making an immediate decision, do not consume them up until later. Does the consumer decision making process differ depending on the consumption immediacy? Will the changes in time horizon for consumption systematically affect the way consumers represent and evaluate the options?

There has been ample amount of research on how consumers make choices and form preferences, most of which concerning immediate consumption scenarios. A well-accepted answer to this question suggests that consumers represent alternatives by their attributes and then compare these attributes in order to determine their preference (Bettman, Johnson and Payne 1991). In evaluating their options, consumers mostly rely on attributes that are shared by alternatives (alignable differences) and mostly ignore attributes that are identical (commonalities) or that do not have a correspondence in other alternatives (non-alignable differences) (Gertner and Markman 1994).

However, characteristics that are unique to representation of future events can lead these results to be inapplicable. Temporal construal theory (Liberman and Trope 1998, Trope and Liberman 2000) suggests that distant future events have more abstract, general and de-contextualized representations (high-level construal), whereas near future events have more concrete, specific and contextual representations (low-level construal). Therefore, product options that are considered for future consumption will have more abstract representations. This abstraction in the representation would enable the use of holistic (alternative-based) strategies and creation of abstract level attributes in the evaluation process (Johnson 1984). Consequently, in constructing

abstract features or creating a holistic evaluation for each option, consumers are as likely to use non-alignable differences as they use alignable differences.

We hypothesize that the relative importance of non-alignable differences, as compared to alignable differences, will be higher when the consumption is expected to take place in the distant future as opposed to near future. We tested this hypothesis with an experiment, by manipulating time horizon (tomorrow vs. six months from now) between subjects with the following scenario.

“Imagine that you have been chosen to be the part of the group who is going to prepare the end of classes party this year, which going to take place tomorrow night (six months from now)You have been given the responsibility of choosing and buying the microwave popcorns for the party.”

Following this, descriptions of two popcorn brands were presented (in a counterbalanced order), which had equal overall attractiveness ratings (Zhang and Markman 2001). Brand P is designed to be better in alignable attributes, while Brand Q is designed to be better in non-alignable attributes. Hence, a shift in preference between these brands in different time horizons will reflect the different attention given to alignable and non-alignable differences. Participants were asked to indicate their preference (11-point scale anchored Prefer P and Prefer Q, with a neutral point of indifference), allocate 100 points between the brands and provide thought listings.

We expected to find a preference toward the alignable-better brand in near future, and a shift toward the non-alignable better brand in distant future, since consideration of the non-alignable attributes would make non-alignable better brand as attractive as the alignable-better brand. Thus, finding participants to be either indifferent or have a preference toward the non-alignable better brand in distant future, would be an evidence for increased attention to non-alignable attributes.

65

As expected, alignable-better brand was preferred by 77% (68 out of 88) of the participants in near future, but only by 65% (58 out of 89) in distant future (Z = -1.78, p <.05). In contrast, 23% (20 out of 88) of the participants were either indifferent1 or preferred non-alignable better brand in near future, as opposed to 35% (31 out of 89) in distant future (Z = 1.78, p <.05). Similarly, mean number of points allocated to the non alignable-better brand was greater in distant future (M = 43.38) than in near future (M = 37.80); (F(1, 177) = 4.816, p < 0.05).

Lastly in thought listings, which were coded by an independent rater who was blind to the hypotheses, we expected to find a greater mention of the alignable attributes in near future, but not in distant future. As expected participants mentioned alignable attributes (M = 1.40) more than the non-alignable attributes (M = 0.77) in near future (F(1, 176) = 3.799; p = .053), but not in distant future (MA = 1.08, MNA = 0.90; F(1, 176) < 1 , p >.1),suggesting more emphasis on alignable difference in near future, but a more balanced consideration to both types of differences in distant future.

The results of the experiment provide support for the hypothesis that the representation and evaluation of alternatives are systematically different for situations with varying consumption immediacy. Specifically, we find that consumers’ disproportional emphasis on the alignable attributes are decreased when they expect the consumption take place in distant future as opposed to being immediate. This result is consistent with our prediction that people’s representations of the options in future time framing does lead to a more abstract and holistic processing.

Taken together, current findings add to the understanding of how consumers evaluate and chose among alternatives when the consumption is in future, by demonstrating a systematic shift in attention paid to different types of attributes for immediate versus future consumption scenarios. In addition, the findings in this paper also extends the structural alignment theory (Medin, Goldstone, & Markman 1995), by addressing the issue of when and how the contextual factors promote the processing of the non-alignable attributes, which have been found to be effortful and difficult. Specifically, we have showed that the differences in time horizon can moderate the effect of alignability on preference formation. For further information contact:

Selin A. Malkoc, Kenan-Flagler Business School, University of North Carolina at Chapel Hill, McColl Building, #3490, Chapel Hill, NC 27599 – 3490, Phone: (919) 843-8386, FAX: (919) 962-7186, E-mail: Selin_Malkoc@unc.edu References Bettman, R. J., Johnson, E., and Payne, J. (1991).

Consumer decision making. In T. S. Robertson & H. H. Kassarjian (Eds.), Handbook of consumer behavior (pp. 50-84). Englewood Cliffs, NJ: Pretince Hall.

Gertner, D., and Markman, A. B. (1994). Structural alignment in comparison: No difference without similarity. Psychological Science, 5(3), 152-158.

Johnson, M. D. (1984). Consumer choice strategies for comparing noncomparable Alternatives. Journal of Consumer Research, 11, 741-753.

Liberman, N., and Trope, Y. (1998). The role of feasibility and desirability considerations in near and distant future decisions: A test of temporal construal theory. Journal of Personality and Social Psychology, 75(1), 5-18.

Trope, Y., and Liberman, N. (2000). Temporal construal and time-dependent changes in preference. Journal of Personality and Social Psychology, 79(6), 876-889.

Zhang, S., and Markman, A. B. (2001). Processing product unique features: Alignability and involvement in preference construction. Journal of Consumer Psychology, 11(1), 13-27.

66

3.3.1 SWITCHING BETWEEN PRICING SCHEDULES: A DYNAMIC ADVANTAGE FOR FLAT RATES

Tom Meyvis, New York University Jinhong Xie, University of Florida

ABSTRACT

Many services offer both linear rate (per-unit fee) and flat rate (fixed fee) pricing options. Examples include cell phone calling plans, public transportation, and health clubs. Several studies have documented a flat rate bias: consumers often choose flat rates when they would have been better off by paying per unit (Kridel, Lehman, and Weisman 1993, Nunes 2000). A number of factors may contribute to this exaggerated flat rate preference, including increased convenience, increased enjoyment of the service, reduced uncertainty, self-control goals, and overestimation of usage. The objective of this project is to examine whether, aside from the documented static preference for flat rates over linear rates, there may also be a dynamic preference for flat rates. More specifically, even when flat and linear rates are equally attractive, consumers may find it easier to switch from the linear rate to the flat rate than to switch from the flat rate to the linear rate.

In our first experiment, we created three different between-subjects conditions. In the control condition, subjects were indifferent between an attractive flat rate and a comparably attractive linear rate. In the switch-from-linear condition, subjects were first presented with the attractive linear rate from the control condition and an overpriced flat rate. Subjects naturally preferred the linear rate. When this flat rate was then gradually reduced to the attractive rate from the control condition, the majority of the subjects switched to the flat rate. Finally, in the switch-from-flat condition, subjects were first presented with the attractive flat rate from the control condition and an overpriced linear rate. Subjects naturally preferred the flat rate. When the linear rate was then gradually reduced to the attractive linear rate from the control condition, only a minority of subjects switched.

There are a number of mechanisms that can explain the observed disproportionate “stickiness” of flat rates. First, consumers who currently have a flat rate may forget to adjust their usage downward when considering switching to a linear rate. This would make the linear rate appear overly expensive and thus discourage the switch. In contrast, when consumers consider switching to a flat rate, the increase in usage is very salient, as the flat rate in

essence implies “use as much as you want.” Second, consumers who consider switching to a linear rate may adjust their usage, but anchor on their current flat rate usage and adjust insufficiently (e.g., Mussweiler and Strack 2001), again making the linear rate appear overly expensive. Finally, consumers may sufficiently adjust their estimates of service usage, but feel worse about reducing their usage of the service (when switching to a linear rate) than they do about giving up money (when switching to a flat rate). This last explanation assumes that consumers exhibit a greater loss aversion bias for service usage than they do for money. This would be consistent with prior findings from the promotion literature that consumers are more reluctant to trade down (sacrifice quality for price) than to trade up (sacrifice price for quality) (e.g., Hardie, Johnson, and Fader 1993).

Whereas the first experiment demonstrated the disproportionate stickiness of flat versus linear rates, the subsequent experiments tried to distinguish between the three alternative accounts of the effect. The results of the second experiment demonstrated that the effect was magnified when participants were asked to explicitly estimate how much their usage of the service would change if they would switch to the new rate. This is inconsistent with the first explanation (i.e., consumers forget to adjust their usage downward when considering to switch to a flat rate), but consistent with both the second explanation (i.e., anchoring and insufficient adjustment) and the third explanation (asymmetric loss aversion for money vs. usage). In a third experiment, we observed that the effect persisted when the usage estimates for the different (randomly ordered) pricing plans were elicited before the experiment. Since this manipulation removed any systematic anchoring effects, these results ruled out the anchoring and insufficient adjustment account, but were consistent with the asymmetric loss aversion account. Finally, in the fourth experiment, we observed that the disproportionate "stickiness" of flat rates persisted even when the increase in spending when switching from a linear rate to a flat rate was made salient.

Together, these studies indicate that consumers are more likely to switch from a linear rate to an equivalent flat rate than they are from a flat rate to an equivalent linear rate. This dynamic advantage

67

for flat rates seems to be driven by a greater loss aversion for service usage than money. Furthermore, this asymmetric loss aversion is not the result of a greater salience for reducing usage than for increasing expenses, but instead seems to result from fundamental differences in the valuation of decreasing benefits versus increasing expenses. References Hardie, G. S. Bruce, Eric Johnson, and Peter S.

Fader (1993), "Modeling Loss Aversion and Reference Dependence Effects on Brand Choice," Marketing Science, 12 (Fall), 397-409.

Kridel, Donald J., Dale E. Lehman, and Dennis L. Weisman (1993), “Option Value, Telecommunications Demand, and Policy,” Information Economics and Policy, 5, 125-144.

Mussweiler, Thomas and Fritz Strack (2001), “Considering the impossible: Explaining the Effects of Implausible Anchors,” Social Cognition, 19(April), 145-160.

Nunes, Joseph C. (2000), "A Cognitive Model of People’s Usage Estimations," Journal of Marketing Research, 37 (November), 397-409.

For further information contact: Tom Meyvis, New York University, 44 West 4th Street KMC 9-75, New York, NY 10012 (tmeyvis@stern.nyu.edu).

68

3.3.2 PENNY WISE AND POUND FOOLISH: THE LEFT DIGIT EFFECT IN PRICE ENCODING

Manoj Thomas, New York University Vicki Morwitz, New York University

ABSTRACT The popularity of nine-ending price points amongst retailers has attracted researchers’ attention as early as 1930. Past research suggests that using a nine ending price (e.g., $1.99 instead of $2.00) can have a significant impact on firms’ profits (Schindler and Kibarian 1996, Blattberg and Neslin 1990, Stiving and Winer 1997). However, the cognitive process that leads to price ending effects remains enigmatic (Monroe and Lee 1999). We conducted four experiments to examine how price endings affect numerical cognition processes. Results of study 1 suggest that nine endings in a product price reduce the perceived magnitude of the price relative to a reference price only when the left-most dollar digit changes. Study 2 shows that effects of nine endings in a price are moderated by the difference between the nine-ending price and reference price; nine-endings effects are more likely to manifest when the prices being compared are close to each other. Study 3 measures response times to confirm that the moderating effect of distance was mediated by cognitive load. Study 4 replicates this effect in a non-price domain, namely consumers’ perception of quality rating. Based on these four studies, we propose three properties of nine ending price cognition: the left digit effect, the distance effect and domain invariance. The Left Digit Effect. The left digit effect refers to the property that underestimation of a nine ending price is more likely when the left-most digit of the nine ending price changes. Using a nine ending instead of a zero ending, for example $2.99 instead of $3.00, changes the dollar digit (from three to two) and it is this change in left digit rather than the one cent drop that affects magnitude perception of the price. The analog model of numerical cognition (Dehaene 1997) states that multi-digit number comparison entails two processes; first the different digits in the numbers are converted into holistic mental quantities represented on an internal analog scale and then these analog representations are compared. However, despite the fact that perceivers take into account all the digits in encoding the magnitude of a multi-digit number, due to left-to-right processing, at least in some situations, the left-

most digit is likely to exert a greater influence on the encoded magnitude.

Distance Effect. The left-digit effect does not

manifest always. Perceivers tend to anchor magnitudes of multi-digit numbers on the left digit only when the internal discriminability between the two numbers being compared is poor. Research in marketing suggests that price evaluation often, if not always, involves a comparison of two prices, a target price and a comparison standard (Adaval and Monroe 2002, Janiszewski and Lichtenstein 1999, Niedrich, Sharma and Wedell 2001, Winer 1988). However, before two numbers can be compared the numerical symbols have to be mapped on to the internal analog scale. This process of mapping from numerical symbol to mental magnitudes imposes a cost on the speed of mental calculations when the numbers being compared are close to each other. The closer the numbers being compared, the greater is the difficulty in discriminating the numbers on the internal analog scale. Consequently, the closer the numbers being compared, the greater is the likelihood of the left digit effect. Domain Invariance. Domain invariance refers to the property that the underestimation of nine-ending numbers is not restricted to the domain of prices; it manifests with other multi-digit numbers also. Past research has often attributed the popularity of nine ending prices to perpetuated retailing practices (Gabor 1977, Gabor and Granger 1964, Nagle and Holden 1987, Schindler 1991). Based on a survey of published material and informal conversations with consumers and retailers, Schindler (1991) proposed a list of fourteen meanings that price endings are likely to communicate to consumers. These meanings can be broadly classified into two groups: price-related meanings (such as “low price,” “discount price”), or meanings concerning non-price attributes of the product or retailer (such as “low quality”). However, if consumers’ favorable response to nine ending prices is solely on account of the images of ‘discount’, ‘low price’ etc. evoked by these prices, then these effects should be confined to the domain of prices. On the contrary, if these effects are, at least partly, due to the influence of left-to-right processing during the magnitude encoding of multi-digit numbers, then these effects should remain invariant to changes in domain.

69

Drawing on the premise that left to right processing is a fundamental characteristic of multi-digit encoding, we predict that left digit effect and its interaction with distance effect will manifest in all nine ending numbers and not just prices. (It may be noted that we are not ruling out the image effects phenomenon in price cognition; rather, our suggestion is that nine ending numbers can be underestimated even when these numbers are not associated with images of low magnitudes.) The question whether consumers respond differently to a price that is a cent lower than the nearest round figure is interesting from a theoretical as well as managerial perspective. From a theoretical perspective, the answer to this question will contribute to the literature on cognitive responses to numerical stimuli. The answer will be germane to researchers debating whether numbers are processed analogically or digitally (Dehaene, Dupoux and Mehler, 1990; Hinrichs, Yurko and Hu, 1981). For further information contact: Manoj Thomas, Phone: (212) 998-0524, Email: mthomas@stern.nyu.edu

70

3.3.3 REFERENCE OTHER, TRUST AND PERCEIVED PRICE FAIRNESS

Lan Xia, Bentley College Kent B. Monroe, University of Illinois at Urbana-Champaign

ABSTRACT

The issue of price fairness is becoming increasingly relevant as more companies apply dynamic pricing using new technologies. Such pricing practices enable companies to offer differential prices to different consumer segments or different prices to the same customer on different occasions. However, consumers may not find these practices to be acceptable leading to perceptions that some pricing practices or prices are unfair. For example, Coca-Cola’s idea of installing “smart vending machines” that changes prices by temperature and Amazon.com’s practice of offering different prices for the same product to the same consumer at different occasions were not perceived favorably by consumers.

Existing research on price fairness has focused mainly on the processes of setting a particular price including the firm’s inferred motive for a pricing practice (Campbell 1999), and attributions of causes of a price change (Bolton, Warlop, and Alba 2003; Vaidyanathan and Aggarwal 2003). While research suggests that judgments of price fairness are based on comparisons with a reference other, this comparative aspect of price fairness has been relatively under researched. In this research, we examine the influence of different comparative others and suggest that comparing with other customers leads to the most unfairness perceptions relative to consumers’ self-comparison or comparing between two sellers. Further, previous research has shown that a good reputation may serve as a buffer to decrease the severity of unfairness perceptions (Campbell 1999). However, trust with a retailer developed from personal transactional experiences may enhance the perception of unfairness when a consumer was charged a higher price compared to a reference other. In addition, we also explore the consequences resulting from when a price is perceived as unfair and examine potential mediating factors for different types of consumer reactions to such a perception. We propose that perception of unfairness leads to both perceived lower value and negative emotions. The effect of unfairness perception on purchase intentions is mediated primarily by perceived lower value while the effect of unfairness perception on the desire to

complain and disseminate negative word-of-mouth is mediated mainly by negative emotions.

To test the propositions, we conducted a 3 (comparative other) x 3 (price outcome) x 2 (trust) factorial experiment. Subjects first read an online shopping scenario, followed by a set of questions measuring their price fairness perceptions, perceived value, negative emotions, purchase intention, and desire to complain. Comparative other was manipulated as subject’s previous experience with the price, another customer’s price, or another seller’s price on the same product. Price outcome was manipulated as the same, higher, or lower than the subject’s own price. Trust was manipulated as either familiar with the retailer and having done business with the retailer (with inter-personal trust) or a new encounter with the retailer (without inter-personal trust).

The results largely supported our predictions. Customer-customer comparisons had the largest effect among the three types of comparative others in the equality and disadvantaged inequality conditions. This result suggests that there is the potential danger that perceptions of price unfairness will occur when a seller offers different prices to consumers for the same product. We found that a good buyer-seller relationship (i.e., inter-personal trust) enhances buyers’ perceptions of price fairness when the price outcome is equality or advantaged inequality. However, when the price outcome is a disadvantaged inequality a good prior buyer-seller relationship enhances perceptions of price unfairness. Finally, our study showed that consumers may take different actions when a price is perceived to be unfair and these actions were driven by different factors. A decrease in purchase intentions is driven primarily by perceived value while the desire to spread negative word-of-mouth is driven mainly by a negative emotion. This result has important managerial implications. While the seller may lose potential sales when consumers forgo potential transactions when facing an unfair price, the damage from negative word-of-mouth may be more severe. Hence, to reduce the potential larger scale damage, it is crucial to offer consumers an avenue to “vent” their anger in additional to offering financial compensation.

71

For further information, contact Dr. Lan Xia, Marketing Department, Bentley College, 175 Forest St., Waltham, MA 02452, Email: lxia@bentley.edu

72

3.4.1 (WINNER): IN NEED OF A FAVORABLE CONCLUSION: THE ROLE OF GOAL-BIASED REASONING IN CONSUMER JUDGMENTS AND EVALUATIONS

Gustavo E. de Mello, University of Southern California

Advisor: Professors Deborah MacInnis and David W. Stewart ABSTRACT

Millions of educated consumers routinely purchase and use products for which the available information is dubious or even contradictory of the products’ claims. The alternative medicine industry, for example, is a multi-billion dollar business that thrives in marketing “miracle” cures and solutions to virtually anything a consumer may desire: health, beauty, even spiritual harmony. This is no trivial matter: consumers may use these clinically untested products, and they may not only be wasting their money, but also putting their health at risk. This dissertation posits that, through a process we term goal-biased reasoning, consumers engage in strategic information gathering and processing to arrive at a conclusion that is in agreement with the attainability of a personally-relevant goal. In the proposed framework, consumers for whom the attainability of a personally-relevant goal has been challenged select cognitive processing strategies and generate beliefs that result in positive attitudes toward the purchase and consumption of products that are presented as means to goal attainment.

We propose that consumers’ vulnerability to misleading promotions and unsubstantiated claims is not a question of “smarts,” education, or expertise: Instead, consumers fall victim to their own motivation to see their end-states as attainable, which leads them to seek and process information in a self-serving way so as to conclude that the end-state they are striving for—health, wealth, beauty, etc.—is within their reach. At the current state of research, we present preliminary evidence from two studies that suggest that, indeed, consumers engage in cognitive processes and behaviors that result in an increased vulnerability to certain product claims, and an amplified proneness to forming favorable evaluations and judgments of some products and services.

In two studies, subjects who were led to doubt their ability to attain a personally-relevant end-state (achieving a high academic performance by the end of the semester), exhibited cognitive processing styles that differed from those of subjects who were led to have more confidence in achieving the same end-state. This difference in cognitive processing style was observed in the evaluation of a product that presented itself as a

means to attaining the desired end-state (i.e., a memory and mental focus booster), but not in the evaluation of other end-state unrelated products (i.e., a stain remover, and a joint pain reliever).

In the first study, we analyzed how consumers differing in their perceived ability to attain a desired end-state (high academic performance) evaluated both a goal-relevant product (a memory and mental focus booster) and a goal-irrelevant product (a stain remover) when the only information available about it is that the product claims to either boost brain performance or remove stains. Additionally, subjects were told either that the product was “the only one on the market” (extraordinary claims condition) vs. “one of many on the market” (ordinary claims condition) making such claims. We predicted that subjects in the low perceived ability condition would find the product claims as more believable, and expect the product to be more effective, than those in the high perceived ability condition. Second, we expected high perceived ability subjects (but not low perceived ability subjects) to be more distrustful of “extraordinary” product claims than of “ordinary” product claims. Finally, we expected to observe these results only for the end-state relevant product, and not for the irrelevant product.

As hypothesized, results showed that subjects in the low perceived ability condition found the “memory and mental focus booster” product claims more credible than did subjects in the high perceived ability condition. Also, and based on the minimum information provided, low perceived ability subjects thought the product was better and more effective than did high perceived ability subjects. There were no significant differences in the evaluation of the “stain remover” product between conditions.

The interaction between perceived ability and how extraordinary the claims were was marginally significant, but also in the predicted direction. While low perceived ability subjects did not care how ordinary or extraordinary the claims were, those in the high perceived ability condition found the extraordinary claims less credible than the ordinary claims. Similarly, while high perceived ability subjects discriminated between common vs. extraordinary product claims, tending to believe the product was better and more effective when there

73

were many products on the market than when the product making the claims was only one, low perceived ability subjects did not discount the “extraordinary” product claims quality or effectiveness.

The design of the second study closely resembled that of Study 1: the same manipulation of perceived ability was employed, and the target product was the same (memory and mental focus booster). The goal-unrelated product was in this case a “joint pain reliever.” Additionally, subjects were able to look up product information at will from two different sources: a “partial” source (i.e., the product’s brochure), and an “impartial” source (i.e., a newspaper article on the product). The study was designed to uncover some of the cognitive processes favored by goal-biased consumers, in an experiment that tracked the amount and type of product information sought by the subjects.

While subjects evaluating the “joint pain relief” product didn’t exhibit any difference in information search patterns based on perceived ability to attain their desired end-state, those evaluating the “memory and mental focus booster” displayed a more biased information search strategy: specifically, subjects evaluating the “memory booster” looked up more information from the “partial” source (product brochure) than did those evaluating the “joint pain reliever.” This difference was even more pronounced in the in the low perceived ability condition. The search for information from the “objective” source (newspaper article) didn’t present significantly different patterns across conditions.

In sum, we propose that consumers cope with the uncertainty of attaining a personally-relevant end-state by selecting cognitive processing strategies and generating beliefs that result in positive attitudes toward the purchase and consumption of products that are presented as means to the attainment of such end-state. The consequences of this phenomenon are important: in addition to the expenditure of money in potentially useless products, consumers may engage in riskier practices that endanger their health and that of others. This dissertation proposal aims at making several contributions: (a) gaining a better understanding of goal-biased reasoning processes in a consumer context; (b) identifying some of the cognitive strategies used by goal-biased consumers when searching for and evaluating products that present themselves as means to goal attainment, and boundary conditions for the occurrence of goal-biased processes; and (c) outlining future studies to generate further knowledge that can be implemented by marketers and policy makers to

promote a more socially responsible use of marketing communications for certain products and services. For further information, contact: Gustavo E. de Mello, Department of Marketing, University of Southern California, Leventhal Bldg., 306F, Los Angeles, CA 90098-0443, Email: demello@marshall.usc.edu.

74

3.4.2 (RUNNER-UP) UNINTENDED PURCHASE OPPORTUNITIES: CONFLICT, CHOICE, AND CONSEQUENCE

Anirban Mukhopadhyay, Columbia University

Advisor: Professor Gita Johar ABSTRACT Consumers spend substantial proportions of their expenditures on products they had not intended to buy. Correspondingly, marketers spend billions of dollars trying to influence purchase incidence. Yet little is known about how consumers respond to unintended purchase opportunities, i.e., tempting offers that they had not intended to buy. This dissertation addresses the question of how consumers respond to such unintended purchase opportunities, and how a decision to either buy or not buy can affect responses to subsequent tempting offers. Distinct from the majority of research on self-control, it builds on the insight that purchase consists of the two related but independent activities of spending and acquisition, and proposes a conceptual model specific to situations of unintended purchase.

The insight that purchase and consumption are conceptually different is key to this model, because most approaches to the study of self-control treat it as a problem of over-riding an immediate temptation in service of preserving a long-term goal (e.g., Ainslie 1975; Kivetz and Simonson 2002; Trope and Fishbach 2000). Such consumption-related situations are different from a typical purchase context, where a tempting offer consists of an attractive product-price combination, and while the attractive price may spur consumers to spend where they had not intended, thereby threatening the goal of not spending, the product in question may serve certain other goals independently. It is argued that unintended purchase opportunities simultaneously cue both—the goal of not spending (Fishbach, Friedman, and Kruglanski 2003) as well as the goal served by acquiring and using the product (Stern 1960). These goals may be in conflict, and under a high need to justify, i.e. when in conflict, consumers will look for reasons to support their decision (Shafir, Simonson, and Tversky 1993). Reasons in such situations could be of two types—those supporting the attainment of the goal served by acquisition at the cost of violating the goal of not spending, and those supporting the preservation of the goal of not spending at the cost of forfeiting the goal served by acquisition. Hence the conflict may be resolved through the relative justifiability of the two options

available to the consumer, namely, buy or not buy. (If however these goals are not in conflict, the decision would be made on a simple economic cost-benefit trade-off.)

Further, conflict between the goals of acquisition and not spending cued at unintended purchase opportunities should also drive complex affective responses (Rook and Gardner 1988). Specifically, the choice to give in should lead to the affective consequences of guilt (Lascu 1991) from violating the goal of not spending and happiness from the attractive product acquired. On the other hand, the choice to hold back should lead to pride from having protected the goal of not spending and regret (Carmon and Ariely 2000) from having forfeited the goal served by acquisition.

Six experiments are planned to test this

model. At the time of writing, four experiments have been completed. The first three are scenario-based studies involving stimuli that were selected after pretests screening on several criteria. Products on usage may differentially serve long-term and short-term goals. Products that deliver higher value in the short term than in the long term tend to offer instant gratification by serving a short-term goal. Wertenbroch (1998) identified such products as “those that people are tempted to over-consume,” namely, “vices.” Conversely, products that offer lower value in the short than in the long term tend to be those that “people are tempted to under-consume,” namely, “virtues.” It is proposed that either type of product, even within the same category, may function as a temptation. However, since virtues deliver greater value in the long term, it is predicted that they will be easier to justify buying. Products were identified (in two categories—books and software) such that they were equally attractive (and not too attractive) and equated on price within category, but some delivered greater value in the short than in the long term and were characterized as more indulgent than prudent (i.e. “vices”), while others delivered greater value in the long than in the short term, and were more prudent than indulgent (i.e. “virtues”).

Results from the first three experiments demonstrate: (1) the justifiability of buying at an unintended purchase opportunity depends both on the decision at a prior opportunity, as well as the

75

nature of the product currently on offer, “virtue” or “vice”; (2) accessibility of reasons for prior decisions to buy or not buy drives responses to current offers, with high accessibility of past restraint inducing self-rewarding behavior; and (3) complex affective outcomes combining happiness, pride, guilt, and remorse result from decisions to either buy or not buy. Experiment 4 extends the investigation from scenario-based studies to the domain of advertising response, and builds on the finding that restraint (i.e. not buying) at one point in time simultaneously induces feelings of pride and unhappiness. It is found that that under such conditions, affect-consistent advertising appeals may be more effective, even if such appeals contain mixed messages that have been shown to be far less effective when respondents have no shopping history and therefore are at baseline affective states.

Theoretically, this dissertation has three main contributions. First, the proposed justification mechanism explains the observed inter-temporal behavior patterns better than other competing theories. The demonstration that complex affective responses can result from exposure to tempting offers, and that such affect may also have an inter-temporal effect is also new in consumer psychology. Finally, the idea that virtues can function as temptations—and that they are often more justifiable to give in to—is new; as much of the extant literature has addressed self-control issues only in the context of vices and vice-like behaviors. From a managerial perspective, this dissertation has specific implications for promotions managers, distance marketers making sequences of consecutive offers, and sales personnel in one-on-one selling situations. For further information contact: Anirban Mukhopadhyay, Columbia Business School, Uris Hall, Room 311, 3022 Broadway, New York, NY 10027.

76

3.4.3 (RUNNER-UP): IT'S THE TYPE OF THOUGHT THAT COUNTS: CONSEQUENCES OF BIASED VERSUS OBJECTIVE PROCESSING FOR ATTITUDE CERTAINTY.

Derek Rucker, Ohio State University Advisor: Professor Richard Petty

ABSTRACT This research examines the role of undirected versus directed thinking in persuasion. Although past research has found that directed thinking, relative to undirected thinking can influence the valence of attitudes, little research has compared attitudes that are equivalent in valence following undirected versus directed thinking. This research tests the notion that, when attitude change is equivalent, directed thinking can lead to attitudes that are stronger and held with greater certainty than undirected thinking. Specifically, it is hypothesized that sometimes thinking about a message in a directed manner can provide unique information about one’s thoughts that is not present when the same message is thought about in an undirected manner. It is posited that this unique information creates attitudes that are held with greater certainty and more predictive of behavior. For further information contact: Derek P. Rucker, Department of Psychology, The Ohio State University, 1885 Neil Avenue, Columbus, OH 43210-1222. Phone: (614) 688-5668, Email: rucker.46@osu.edu

77

4.1.1 UNDERSTANDING THEORY-BASED INFERENCE MAKING WHEN THEORIES AND DATA CONFLICT: THE ROLE OF THEORY CONFIDENCE

Dobrina G. Teifel, Cornell University

Douglas M. Stayman, Cornell University

ABSTRACT Past research has shown that the availability of an intuitive theory about attribute correlations impacts the process of making an inference about a missing attribute, above and beyond using stimulus-based empirical correlations. For instance, an intuitive theory can be used as an inference cue in the face of empirical correlations in the data which run counter to the theory (Broniarczyk and Alba 1994), or as a cue to make compensatory inferences amongst attributes (Chernev and Carpenter, 2001). This research extends previous findings to examine mechanisms underlying theory-based consumer inference making when theory-based expectations conflict with data. Specifically, we explore how processing motivation, operatinalized through NFC, and theory confidence affect inference behavior and the extent to which consumers use the theory versus the unexpected data in inferences, following exposure to unexpected (yet credible) information. An initial set of studies has explored the role of motivation on the extent of using theory versus data in inferences. Results from one experiment suggest that high (and not low) NFC subjects exhibit an unexpected tendency to make inferences aligned with their theories during an inference phase (i.e., stick with their theories) under moderate processing time conditions, in spite of appearing to have had sufficient time to notice the unexpected attribute correlation during a preceding study phase. Interestingly, they also feel more confident in their (theory-consistent) inferences than their low NFC peers. Results from a second study indicate that high NFC individuals’ theories are “stickier” and less prone to be abandoned in the face of contrary data than are those of low NFC individuals, in spite of their seemingly similar expectation levels. These effects appear to at least in part stem from metacognitive processes upon reflecting on one’s perceived resistance of a theory undermining attempt (Tormala and Petty 2002) which should reinforce the theory and lead to upholding it with greater confidence. Given that perception of resisting theory-undermining information should lead to increased theory confidence, another set of studies further

explores the effects of perceived resistance of theory-undermining information on theory confidence and subsequent inference behavior. Specifically, we consider the effects of information disconfirming prior theory-based expectations received through either a reliable or non-reliable source on the confidence with which these expectations are held and on subsequent inference behavior. We also seek to explore how processing resources and motivation influence the process. Finally, we study the role of premeasured theory confidence on inferences. One factor used is the individual’s NFC, which, as described above, has proven to influence the extent to which people “stick with” their theories. Results indicate that, even though initial theories do not differ for individuals across levels of confidence, at “high” levels of confidence, individuals inferred in line with their theories regardless of NFC, and at “low” levels of confidence inferences were more in line with the unexpected data. However, at “medium” levels of measured confidence, low NFC individuals inferred more in line with the data, while high NFC individuals inferred more in line with their theories. These findings support the results indicating that high NFC is associated with increased theory “stickiness”. In summary, it appears that the level of prior theory confidence and theory “stickiness” lead to divergent use of theory or data in inferences. Implications of these findings and possible extensions are discussed. For further information contact: Dobrina Teifel, 301 Sage Hall, Cornell University, Ithaca, NY 14853, e-mail: dgs22@cornell.edu

78

4.1.2 A META-COGNITIVE ANALYSIS OF RESISTANCE TO PERSUASION AND ATTITUDE CERTAINTY

Zakary L. Tormala, Indiana University Richard E. Petty, Ohio State University

ABSTRACT Although attitude change and persuasion researchers over the years have prioritized the study of successful persuasion, considerable attention has also focused on resistance to persuasion—the act of defending one’s attitude against persuasive attack (see Knowles & Linn, 2004). In all of this research, it has been assumed that when people are successful in resisting persuasion, their initial attitudes are completely unchanged. Accordingly, it has been assumed that when persuasion practitioners fail in their attempt to persuade, they have at least done no harm to their cause. In contrast to this view, we propose a meta-cognitive framework to understand resistance to persuasion, which suggests that when people resist persuasion, their initial attitudes can change with respect to attitude strength. More specifically, we propose that when people resist, they can perceive this resistance and become more certain of their initial attitudes. However, we submit that resistance only increases attitude certainty when situational factors suggest the resistance is diagnostic (or informative) as to the validity of the initial attitude. Stated differently, resistance should only increase certainty when people are impressed with their resistance. What determines how impressed people are? We have recently considered the role of two situational factors: perceived message strength and source credibility. PERCEIVED MESSAGE STRENGTH In one line of research (Tormala & Petty, 2002), we explored the role of perceived message strength in determining the effect of resistance on attitude certainty. We reasoned that resisting messages perceived to be strong would more likely increase certainty than resisting persuasive messages perceived to be weak. Indeed, resisting a strong message presumably is considered diagnostic with respect to an attitude’s validity. Resisting a weak message, however, should be considered less diagnostic, because it leaves open the possibility that stronger arguments exist that might not be resisted. In other words, people should be more

impressed by their successful resistance when they resist messages that seem strong. In a series of experiments, participants were presented with a counter-attitudinal message, which they were instructed to counterargue. When participants successfully resisted the message and perceived it as strong they became more certain of their initial attitudes. When they resisted the message but perceived it as weak, certainty was unchanged. Moreover, consistent with past research on attitude certainty (see Gross, Holtz, & Miller, 1995), this effect had implications for future resistance and attitude-behavior correspondence. Participants’ attitudes became more resistant to subsequent persuasion and more predictive of behavior when they resisted initial attacks perceived to be strong, but not when they resisted initial attacks perceived to be weak. SOURCE CREDIBILITY In another series of studies (Tormala & Petty, in press), we explored the role of source credibility in determining the effect of resistance on attitude certainty. Following the same logic, we proposed that people would be more impressed with their successful resistance when they resisted attacks from high credibility sources. After all, if someone resists an attack from a high credibility source (e.g., an expert), he or she can assume an attack from a low credibility source (e.g., a non-expert) would also be resisted. The same assumption cannot be made when one resists an attack from a low credibility source, because it remains a possibility that a source with more credibility might have been more persuasive or presented better arguments. In two experiments we presented participants with a counter-attitudinal message from an expert or inexpert source. When participants resisted (e.g., because we forewarned them of the persuasive intent of the message), they became more certain of their attitudes when the source was an expert, but not when the source was a non-expert. In addition, attitudes became more predictive of behavior when participants resisted the expert, but not when they resisted the non-expert. We were also interested in the potential moderation of these effects by

79

elaboration, or the extent of thinking. Consistent with other research exploring the role of meta-cognition in persuasion (e.g., Petty, Briñol, & Tormala, 2002; Tormala, Petty, & Briñol, 2002), the effects of resistance on certainty were confined to high elaboration conditions. In summary, these experiments supported our meta-cognitive approach to resistance and also uncovered a new role for source credibility in persuasion—that is, its effect on attitude certainty rather than attitude valence or extremity. CONCLUSIONS Previously, it has been assumed that when a would-be persuader is unsuccessful in an attempt to change an attitude, he or she has at least done no harm to the cause. Our findings contradict this assumption, suggesting instead that when a would-be persuader is unsuccessful in changing a target attitude, he or she may actually have been even less successful than previously realized. Specifically, he or she may have strengthened the very attitude that was targeted for change. In other words, failed persuasion attempts can sometimes backfire by making people even more committed to their original attitudes. Ironically, the probability of this result increases as the persuasive message and source have more of the trappings of persuasiveness (e.g., strong arguments, expert sources). Future research will examine circumstances in which resisting a persuasive message can cause confidence in one’s initial attitude to be undermined. REFERENCES Gross, S., Holtz, R. & Miller, N. (1995). Attitude

certainty. In R. E. Petty & J. A. Krosnick (Eds.). Attitude strength: Antecedents and consequences (pp. 215-245). Mahwah, NJ: Erlbaum.

Knowles, E.S., & Linn, J.A. (Eds.) (2004). Resistance and persuasion. Mahwah, NJ: Erlbaum.

Petty, R.E., Briñol, P, & Tormala, Z.L. (2002). Thought confidence as a determinant of persuasion: The self-validation hypothesis. Journal of Personality and Social Psychology, 82, 722-741.

Tormala, Z.L., & Petty, R.E. (2002). What doesn’t kill me makes me stronger: The effects of resisting persuasion on attitude certainty. Journal of Personality and Social Psychology, 83, 1298-1313.

Tormala, Z.L., & Petty, R.E. (in press). Source credibility and attitude certainty: A metacognitive analysis of resistance to persuasion. Journal of Consumer Psychology.

Tormala, Z.L., Petty, R.E., & Briñol, P. (2002). Ease of retrieval effects in persuasion: A self-validation analysis. Personality and Social Psychology Bulletin, 28, 1700-1712.

For further information contact: Zakary L. Tormala, Department of Psychology, Indiana University, 1101 E 10th Street, Bloomington, IN 47405, email: ztormala@indiana.edu

80

4.1.3 PERCEPTUAL SIMILARITY AND BELIEF CHANGE

Gita V. Johar, Columbia University Anne Roggeveen, Babson College

ABSTRACT This research examines the effectiveness of two refutation strategies – refutations that are perceptually similar versus different to the original claim – in changing consumers’ brand beliefs. A likely prediction is that refutations that are similar to the original claim will more effective to be effective in changing beliefs because the refutation cues the original claim in memory. However, we predict that such memorial cueing will have the opposite effect, such that perceptually similar refutations will cue the original claim in memory and result in resistance to belief updating. We use the truth effect paradigm to derive and test this proposition. The truth effect refers to the finding that repeated claims are believed more than new claims (Hawkins and Hoch 1992). Without specific knowledge about the claim, consumers depend on cues (such as how familiar the claim feels) to judge the veracity of the claim; hence, repeated claims are judged to be truer than new claims. We design two experiments to test whether repetition-enhanced belief are more likely to be changed by perceptually dissimilar (versus similar) refutations. The truth effect has shown that messages which are an exact repetition, and hence, perceptually similar to a previously viewed message, enhance belief to a greater extent than topic repetition (i.e., exposure to topic information at time 1 and detail information at time 2; see Arkes et al 1991; Begg et al 1985). Thus, the perceptual format of a message affects belief. Research also suggests that refutations of previously viewed claims are rated more false than the same refutation presented for the first time when the originally repeated claim is remembered in detail (Begg and Armour 1991). Bases on these findings, we suggest that a refutation that is perceptually similar (vs. dissimilar) to the original claim is more likely to cue the original claim and hence result in resistance to the refutational message. In two experiments, we manipulate perceptual similarity via type of claim (assertion vs. implication) and via surface similarity (look of the two ads) and examine the preceding hypothesis. In Experiment 1, matched inferred and asserted claims were created. These two claim types were shown to different groups of subjects at exposure.

Refutations of the claims, all in assertion form, were then shown at test. Hence, the refutations were perceptually different when the original claim had been inferred and perceptually similar when the original claim had been asserted. Belief in the refutation served as the dependent variable. Findings indicate that, as predicted, belief in the refutation was lower when the original claim was perceptually similar (assertion) versus perceptually different (inference). Further, recall data provide support for the underlying memorial process. Experiment 1a replicates the result with a different dependent variable – truth in the original claim – and rules out the alternative explanation that similar refutations are simply not processed. Thus, paradoxically, awareness of previous exposure is sufficient to reduce subjects’ beliefs in the refutation and maintain their belief in the disputed claim. Experiment 2 provides more direct evidence for the role of perceptual similarity which was manipulated using the similarity in the background page (picture, color and font) of the original claim and refutation. Preliminary results are consistent with the hypotheses. Data analyses are currently under way to establish the memorial process assumed to underlie the effect. REFERENCES Arkes, Hal R., Larry E. Boehm, and Gang Xu (1991), “Determinants of Judged Validity,”Journal of Experimental Social Psychology, 27 (6), 576-605. Begg, Ian M. and Victoria Armour (1991), “Repetition and the Ring of Truth: Biasing Comments,” Canadian Journal of Behavioural Science, 23 (2), 195-213. Begg, Ian M., Victoria Armour, and Therese Kerr (1985), “On Believing What We Remember,” Canadian Journal of Behavioral Science, 17 (July), 199-214. Broniarczyk, Susan M., and Alba, Joseph W. (1994),“The Role of Consumers’ Intuitions in Inference Making”, Journal of Consumer Research, Vol. 21, 393-407. Chernev, A., & Carpenter, G. (2001). The role of market efficiency intuitions in consumer choice: A

81

case of compensatory inferences. Journal of Consumer Research, 38, 349-361. Hawkins, Scott A. and Stephen J. Hoch (1992), “Low-Involvement Learning: Memory Without Evaluation,” Journal of Consumer Research, 19 (September), 212-225. Tormala, Z., & Petty, R. (2002). What doesn't kill me makes me stronger: The effects of resisting persuasion on attitude certainty. Journal of Personality and Social Psychology, 83, 1298-1313. For further information contact: Gita V. Johar, Columbia Business School, Uris Hall, 3022 Broadway, Columbia University, New York, NY 10027, email: gvj1@columbia.edu

82

4.2.1 EFFECTS OF ANGER, REGRET, DISTRIBUTIVE JUSTICE AND SOCIAL COMPARISON ON POSTPURCHASE BEHAVIORS

Carolyn Bonifield, University of Vermont

Catherine A. Cole, University of Iowa

ABSTRACT

Past research on affect has tended to adopt a general valence-based approach to studying emotions, which examines differences between positive and negative affect. However, recent literature suggests that different, but similarly valenced emotions not only have different effects on subsequent behavior, but also arise from different antecedent appraisals (Zeelenberg and Pieters 2001; Lerner and Keltner 2000, Raghunathan and Pham 1999; Smith and Ellsworth 1985). Appraisal theories of emotions propose that specific cognitions underlie different emotions and that these different emotions have distinct influences on decision making (Frijda, Kuipers, and ter Schure 1989; Smith, Haynes, Lazarus, and Pope 1993; Ellsworth and Smith 1988; Ortony, Clore, and Collins 1988). We study anger and regret, both relevant to postpurchase behaviors. While anger seems to be the resulting emotion when individuals perceive causes of product or service failure to be controllable and seller related, the appraisal literature suggests that regret is the resulting emotion when individuals’ causal ascription of failure is controllable and buyer related.

In an experiment, we examined the antecedents and consequences of regret and anger, which arise in exchange settings when consumers obtain outcomes that are worse than they expected or wanted. By studying these two emotions at the same time, we shed light on links between appraisals, the specific emotions and postpurchase behaviors. In a 2 x 2 x 2 between-subjects design, we manipulated appraisals (internal vs. external), distributive justice (low vs. high), and social comparisons (none vs. downward). We first asked consumers to complete a short survey that asked general questions about restaurants and then asked them to view a menu from which they selected and wrote down an item that they would order. They then viewed one of two professionally-produced two-part videotapes designed to simulate a restaurant service setting, one intended to generate internal

appraisals and the other intended to generate external appraisals. After watching the initial video, participants watched an intervention video that manipulated distributive justice (high/low) and social comparison (downward/none), and then completed the questionnaire which had items measuring affect, postpurchase behavioral intentions, and manipulation checks. As expected, when consumers blame the service provider for the outcome failure, they express more anger and less regret than those who blame themselves.

Much of the research on postpurchase behavioral responses has studied the direct consequences of consumers’ appraisals on complaining and repurchase intentions. Building on prior work, we developed a typology of postpurchase behaviors that includes a broad variety of behaviors and that can be separated into the types of postpurchase behaviors consumers are likely to engage in when they make primarily internal or external appraisals (Hirschman 1970, Richins 1987; Singh 1988). In an earlier study, we confirmed this categorization through factor analysis. Our categorization of postpurchase behaviors consists of retaliatory behaviors, which include complaining, negative word of mouth, a strong preference or demand for cash refunds, and a reluctance to negotiate a solution, and conciliatory behaviors, which include intentions to return to the service provider in the future, positive word of mouth, willingness to accept an exchange, sympathy for the service provider, and willingness to leave an above average gratuity. We predict and find that when compared to consumers who make internal appraisals, consumers who make external appraisals, that is, blame the service provider, are more likely to engage in retaliatory behaviors. Similarly, consumers who make internal appraisals are significantly more likely to engage in conciliatory behaviors than consumers who make external appraisals.

While research has studied the impact of consumers’ locus appraisals on postpurchase behaviors, it is likely that it is the affect generated by these appraisals, not the appraisals themselves

83

that influences these postpurchase behaviors. We expected that the emotional response to a failed outcome mediates the effect of causal appraisals on consumers’ postpurchase behaviors. This is a stronger prediction than previous recent research, which often finds that appraisals have both a direct and indirect influence through emotions on postpurchase behavior (Folkes et al. 1987; Tsiros and Mittal 2000). To test the role of anger as a mediator of the effects of appraisals on retaliatory behaviors, we conducted a mediation analysis. We found evidence of strong mediation in the control condition, where there were no interventions. To test the role of regret as a mediator of the effects of appraisals on conciliatory behaviors, we conducted another mediation analysis. However, we did not find evidence to support the role of regret as a mediator of the effects of appraisals on conciliatory behaviors in the control condition.

In the social psychology and sociology literature, exchange theorists study how perceived justice changes peoples’ reactions to positive and negative rewards in both economic and non-economic social exchanges. Moreover, because individuals can disagree about what constitutes a fair distribution of outcomes, they may use social comparisons to assess the extent to which their outcomes are fair (Folger and Kass 2000; Boles and Messick 1995). The theory of downward social comparison posits that individuals experiencing negative affect can enhance their well-being by comparing themselves with less fortunate others (Wills 1981). We predict and find that social comparisons moderate perceived justice effects. When managers offer no financial awards in the recovery effort (low perceived justice), but make downward social comparisons, consumers will be less angry and less likely to engage in retaliatory behaviors than when the managers do not offer a financial reward and do not make the social comparison. In contrast, when managers offer financial awards, downward social comparisons will have weak effects on anger and retaliatory behaviors because the financial award is sufficient to reduce anger and retaliatory behaviors. This pattern suggests that downward social comparisons may substitute for high compensation following a service failure.

Our research makes several contributions to theory and practice. First, it extends our understanding of postpurchase behaviors by examining how regret and anger differentially influence a comprehensive set of postpurchase behaviors. Moreover, it examines an intervention that reduces anger following a dissatisfactory consumption experience. In the process, this

research develops a more complete model of the antecedents of postpurchase behaviors. Furthermore, the current research provides an enhanced understanding of the causes and consequences of anger and regret. From a managerial perspective, knowing how consumers are likely to respond to dissatisfactory consumption experiences can be valuable information to manufacturers, retailers, and service providers. By selectively managing consumers’ emotional responses to dissatisfactory consumption experiences, the firm may be able to reduce certain retaliatory behaviors. For further information, contact: Carolyn M. Bonifield, Assistant Professor of Marketing, School of Business Administration, University of Vermont, Kalkin Hall, 55 Colchester Ave., Burlington, VT 05405-0157, 802-656-0516, bonifield@bsad.uvm.edu

84

4.2.2 EXPERIENCING SUSPENSEFUL DRAMA: THE SEQUENCING OF EMOTIONS AND RETROSPECTIVE JUDGMENTS

Robert Madrigal, University of Oregon

Colleen Bee, University of Oregon Monica LaBarge, University of Oregon

ABSTRACT

Experience has been defined in a variety of ways, including (a) the apprehension of an object, thought, or emotion through the senses or mind and as (b) an event or series of events participated in or lived through. Experiential consumption, therefore, focuses on experiences as consumption objects and may be thought of as the ways in which people engage in and are engaged by experiences. The type of experience considered in this study is viewers’ consumption of suspenseful drama.

Two conditions are necessary to elicit feelings of suspense. The first is uncertainty about the outcome. Uncertainty increases as the subjective probability of an event’s occurrence gets closer to its nonoccurrence (i.e., 50/50). Suspense also requires that one outcome be perceived as being desirable and the other as undesirable. Suspense represents an experience of apprehension that is based on (a) fear that a desirable outcome may occur, (b) fear that an undesirable outcome may occur, (c) hope that a desirable outcome will occur, (d) hope that an undesirable outcome will not occur, and (e) any possible combination of these hopes and fears.

A useful framework for understanding the experience and evaluation of suspense is the expectation disconfirmation model (EDM). Rather than the traditional view of EDM as a difference between pre-consumption expectations and perceived outcomes, the EDM operates in the experience of suspense as a sequence of emotions arising at different points during and immediately following the consumption of a suspenseful drama (see Kubovy 1999). The types of expectations occurring during suspenseful drama are constructed on the fly as the experience unfolds. Expectations reflect an ongoing process of probability adjustment about what may or may not happen in the future (Lazarus 1991) and serve as the catalyst for fear and hope reactions.

Suspense occurs in the moments leading to the resolution of the drama. Once an outcome is understood, suspense gives way to other types of emotions (Carroll 1990). In particular, the emotions of relief and disappointment arise when

the prospect of a particular outcome deemed to be congruent or incongruent with one’s desires changes for the better or the worse. In the case of relief, one feels pleasure when the prospect of an undesirable outcome is eliminated or changed for the better (i.e., disconfirmed); whereas disappointment occurs when the prospect for a desirable outcome has been disconfirmed (Lazarus 1991; Ortony et al. 1988).

Disconfirmation emotions give rise to two retrospective judgments. The first is outcome satisfaction which is defined as a fulfillment response or, more specifically, the degree to which the outcome is judged to be pleasant or unpleasant. Although relief and disappointment have not been considered in past research, a frequent finding in the EDM literature is that emotions predict satisfaction. The final variable we include is the consumer’s retrospective judgment of the entertainment value derived from watching the suspenseful drama. Given that experiential products are consumed for their hedonic content, the value derived from consumption should be a hedonic judgment. Thus, entertainment value is defined as a subjective evaluation of how enjoyable and fun the suspenseful drama was to watch. Whereas outcome satisfaction is a judgment about the pleasantness of the event’s resolution, entertainment value is concerned with the hedonic payoff derived from the overall experience.

We present and test a structural equation model describing a sequence of reactions to suspenseful drama that borrows from the EDM. Expectations are operationalized as the moment-to-moment fear-hope reactions felt toward the central character (i.e., protagonist) in the suspenseful drama. Fear-hope emotions are then related to the disconfirmation emotions of relief and disappointment and these, in turn, predict satisfaction with the outcome. Outcome satisfaction is then related to a second summary evaluation, entertainment value.

Data from two separate samples were tested using structural equation modeling. The context selected for Study 1 was a videotaped competitive speed-skating race and the context for the second study was an edited 14-minute film clip from a major motion picture. The movie, “Final

85

Destination,” features a protagonist, Alex, who is prescient and chronicles his race against death. The DVD version of the movie also offers alternative endings. Respondents saw one of the two endings: a favorable outcome in which Alex survives or an unfavorable outcome in which he dies.

The structural model shown in Figure 1 was tested using LISREL 8.54. The results indicated a good fit to the data from Study 1 and Study 2 (χ2’s = 40.41, 63.52; df’s = 32, 41; p’s = .15, .01; RMSEA’s = .035, .063; CAIC’s = 187.36, 211.70; CFI’s = .98; respectively). The standardized path coefficients and t-values yielded from the model for each study are shown in Table 1. As hypothesized, the linear trend of fear-hope was positively related to relief and negatively related to disappointment in each study, thus indicating that greater hope toward the end of the game predicted relief and greater fear over the final third of the game predicted disappointment. Similarly, outcome satisfaction was positively predicted by relief and negatively predicted by disappointment in both the data sets. A positive path existed between outcome satisfaction and perceived entertainment value in each sample. A negative relationship was also found between relief and disappointment in the sporting event sample

(β=-.15, t = -2.29, p < .02) but not in the movie sample (β=-.11, t = -1.18, p > .10). In addition, mediation tests revealed that the positive effect of fear-hope on satisfaction in each of the samples was fully mediated by the disconfirmation emotions of relief and disappointment.

Understanding how summary evaluations of an experience are formed is important because they play a pivotal role in predicting future consumption habits. Research addressing peoples’ retrospective judgments of experiences has included many contextual domains and has been conducted using a variety of stimuli. The integration of people’s moment-to-moment reactions with retrospective judgments has been an emerging trend in this literature. In contrast, although conceptual work abounds detailing the nature of suspense, few studies have specifically considered how feelings of suspense influence retrospective judgments (see Alwitt 2002 for an exception). Our contribution, therefore, is a detailed accounting of how the momentary feelings that people have while consuming suspenseful drama influence the emotions and summary evaluations occurring after the event’s outcome becomes known.

FIGURE 1:

THE HYPOTHESIZED MODEL

TABLE 1:

PATH DESCRIPTIONS, HYPOTHESES, STANDARDIZED COEFFICIENTS AND T-VALUES FOR STUDIES 1 AND 2

Path Path Description Study 1

(Speed Skating) Study 2 (Movie)

a Fear-Hope Relief .18 (3.47) .29 (3.60) b Fear-Hope Disappointment -.17 (-2.40) -.20 (-2.18) c Relief Outcome Satisfaction .38 (6.40) .62 (7.91) d Disappointment Outcome Satisfaction -.54 (-9.85) -.30 (-3.46) e Outcome Satisfaction Entertainment Value .22 (3.20) .53 (7.35) f Fear-Hope Outcome Satisfaction w/out Paths

a, b, c, d .16 (2.28) .27 (2.89)

g Full Model with Fear-Hope Outcome Satisfaction path

.00 (.04) .04 (.51)

Linear Trend of Fear-Hope at End

Disappointed

Outcome Satisfaction Entertainment Value

Relieved

86

One-sided t-values of 1.70 > are significant at the .05 level; 2.33 > are significant at the .01 level; and 3.2 > are significant at the .001 level. For further information contact: Robert Madrigal, Charles H. Lundquist College of Business, 1208 University of Oregon, Eugene, OR 97403 Phone: (541) 346-5163, Email: rmadriga@lcbmail.uoregon.edu

87

4.2.3 THE IMPACT OF SOCIAL SATISFACTION ON CONSUMER SATISFACTION JUDGMENT

--- A TELIC/AUTOTELIC PERSPECTIVE OF CONSUMER SATISFACTION

Xiaojing Yang, Indiana University H. Shanker Krishnan, Indiana University

ABSTRACT In a recent article, Bagozzi (2000) pointed out that the social aspects of consumer research have been overlooked (e.g. the effects of societal forces on intraindividual processes, interpersonal behavior and relationships among social collectivities/groups). He also mentioned that the level of consumer studies should also be extended to include groups of consumers (i.e., two-person dyads, families, peer or friendship groups, teams, organizations, and other social units). Consistent with Bagozzi’s concerns, most previous satisfaction research studies consumer satisfaction at an intra-personal level and seldom takes interactions among consumers into consideration. This impact of consumer interaction on consumer satisfaction remains largely unexplored with a few exceptions (e.g., Fournier & Mick 1999; Oliver 1999) and careful conceptualization is needed. The first goal of the present paper is to fill in this gap by studying consumer satisfaction in a group consumption context. Specifically, this research attempts to study the mechanism through which social satisfaction impacts consumer’s overall satisfaction judgment. In order to account for the importance of social aspects of satisfaction, social satisfaction is proposed and used to compare with attribute satisfaction in terms of their impact on overall satisfaction. Social satisfaction is defined as an affective state as a result of interacting with other consumers. Attribute satisfaction is defined as the consumer’s subjective satisfaction judgment resulting from observations of attribute performance (Spreng et al. 1996). Overall satisfaction is defined as an affective state that is the emotional reaction to a product or service experience. Overall satisfaction is viewed at a global consumption level while attribute satisfaction is viewed at an attribute level, and the latter is not the only determinant of former (Spreng et al.1996). To illustrate the importance of social satisfaction, consider the following scenario: a bunch of friends went to see a movie together and it turned out to be a terrible movie (low attribute satisfaction). However they interacted with each other in a happy fashion (high social satisfaction). They may feel happy anyways about the entire

consumption experience, despite the fact that the movie was horrible. A second goal is to study consumer satisfaction from a Telic/ Autotelic perspective. The dominant conceptual paradigm reflected in consumer satisfaction research is the disconfirmation models, in which the confirmation/disconfirmation of preconsumption product standards (e.g. expectations; desires; equity and experience-based norms) is the essential determinant of satisfaction (Fournier & Mick 1999). In life satisfaction domain, many social psychologists have studied the impact of gratifications of different types of needs on life satisfaction and individual subjective well-being (e.g. Deci & Ryan, 2000; Omodei & Wearing 1990). In contrast with this trend, consumer researchers are still overwhelmingly focusing on studying the impact of disconfirmations on consumer satisfaction, though affect-based model and performance-based model appear sporadically (Fournier & Mick 1999). Despite previous (dis)confirmation models of consumer satisfaction, this paper studies satisfaction from a Telic (the attainment of end states) /Autotelic perspective (the process through which these end states are achieved) drawing from life satisfaction literature. In addition to demonstrating the impact of attainment of agency and intimacy needs (Telic antecedents), the present paper studies how process involvement (Autotelic antecedent) impacts consumer satisfaction. Specifically, a dual-mediator process (attribute satisfaction and social satisfaction) is proposed to mediate the impact of needs attainment and process involvement on consumer overall satisfaction judgment. In an experimental setting, the participants were asked to recall a recent dining experience with other consumers and completed measures of interests. The conceptual framework was tested using structural equation modeling methodology. Preliminary results show reasonable support of our theory.

References

Bagozzi, Richard P. (2000). On the Concept of Intentional Social Action in Consumer

88

Behavior. Journal of Consumer Research, 27, 388-396

Deci, Edward L., & Ryan, Richard M. (2000). The “What” and “Why” of Goal Pursuits: Human Needs and the Self-Determination of Behavior. Psychological Inquiry, 11, 227-268.

Fournier, Susan, & Mick, David Glen (1999). Rediscovering Satisfaction. Journal of Marketing, 63, 5-23

Oliver, Richard L. (1993). Cognitive, Affective, and Attribute Bases of the Satisfaction Response. Journal of Consumer Research, 20, 451- 466

______ (1999). Whence Consumer Loyalty.

Journal of Marketing, Vol. 63, 33-44 Omodei, Mary M., & Wearing, Alexander J.

(1990). Need Satisfaction and Involvement in Personal Projects: Toward an Integrative Model of Subjective Well-being. Journal of Personality and Social Psychology, 59, 762-769

Spreng, Richard A., Mackenzie, Scott B., &

Olshavsky, Richard W. (1996). A Reexamination of the Determinants of Consumer Satisfaction. Journal of Marketing, 60, 15-32

For further information contact: Xiaojing Yang, Marketing Department, Kelley School of Business, Indiana University, 1309 East 10th Street, Bloomington, IN 47405. Email: yangxiao@indiana.edu

89

.4.3.1 VARIATIONS IN CONSUMER IMPATIENCE

Haipeng Chen, University of Miami Sharon Ng, University of Minnesota

Akshay Rao, University of Minnesota ABSTRACT Regulatory focus theory has recently been employed to address a host of consumer behavior issues ranging from preference for and attention to certain types of information, differences in behavior across cultures, and satisfaction with outcomes. We employ the principles of regulatory focus theory to understand whether and how consumer impatience (i.e., the desire for immediate relative to delayed gratification) might vary depending on whether the consumer’s goal is to achieve a desirable outcome or to avoid an undesirable outcome. The issue of impatience is akin to the topic of how people incorporate the future into their decision-making, a topic of considerable interest to scholars of consumer behavior. Various studies have examined how the future is discounted and whether discount rates vary across situations and individuals. A high discount rate implies that the future is considerably less important than the present, and people who employ a high discount rate manifest relatively high levels of impatience. Prior empirical research on discounting has been conducted principally in Western cultural contexts. Yet there is abundant evidence in the cross-cultural literature demonstrating that Easterners are more future-oriented and more patient than their American counterparts, due in part to the slower pace of life and the ubiquity of Eastern spiritual values that emphasize that one’s current life is a part of a series of births and rebirths. These differences in time perception suggest that Easterners might not discount the future to the same degree as Westerners. Thus, one objective of this paper is to examine the general issue of consumer impatience and whether the degree and manner in which people discount the future differs across cultures, as reflected in their willingness to pay for early relative to delayed consumption. Further, we draw upon regulatory focus theory to argue that Westerners’ and Easterners’ relative impatience will be impacted by the specific content of the persuasive message. In particular, we argue that subjects’ willingness to pay for expedited

delivery will be influenced by the values and norms embedded in their culture. This argument draws in part from the findings of Aaker and Lee (2001), who observe that a U.S. culture dominant individual (who is, by definition, more promotion focused) should be more persuaded by a promotion-framed message. However, in our case, we argue that not all promotion-framed messages will lead to greater willingness to pay. A promotion-framed message that evokes some distress for the individual will lead to a desire to alleviate this distress through immediate consumption. Thus, it is hypothesized that a U.S. culture dominant individual should prefer immediate consumption only when the message emphasizes the opportunity to avoid not attaining a desirable outcome (i.e., a promotion-loss). Conversely, an Asian culture dominant individual who is more prevention focused should prefer immediate consumption when the message emphasizes the opportunity to prevent an undesirable outcome (i.e., a prevention-loss). We examine these hypotheses in three studies. Studies 1 and 2 tested the main effect of culture. In Study 1, unlike traditional cross-cultural research that uses country as a proxy for culture, we primed bicultural subjects (Singaporeans) through the use of a collage, consisting of 12 cultural icons representing either the Singaporean culture or the U.S. culture, to temporarily increase the accessibility of that particular culture within the individual. For the dependent variable, subjects were asked to indicate how much they would be willing to pay for the quicker delivery of a book they had purchase on the internet. Consistent with our prediction, subjects in the U.S. primed condition were willing to pay substantially more for speed than those in the Singapore primed condition. To address the possibility that Singaporean’s perception of the U.S. culture might not be identical to those of U.S. subjects, study 2 employed a simple design in which U.S. subjects provided their responses with or without exposure to the U.S. culture prime used in study one. The prices U. S. subjects were willing to pay for quicker delivery were not statistically different across the two conditions, and the willingness to pay of the U.S.

90

subjects were statistically indistinguishable from those provided by the Singapore subjects in study 1 who had been primed with the U.S. culture. Study 3 aimed to test the key interaction prediction between cultural values and message framing on consumers impatience by adopting a two (Culture prime: U.S. prime vs. Singapore prime) by two (Regulatory focus: Promotion vs. Prevention) by two (Desirability of outcome: Desirable vs. Undesirable) between-subjects design. Consistent with our prediction, we found a significant 3-way interaction between culture, regulatory focus and desirability of outcome. Specifically, U.S. primed subjects were willing to pay significantly higher prices to assure that they could enjoy the book earlier, while Singapore primed subjects were willing to pay significantly higher prices to assure that they did not have to wait longer for the book to arrive. In addition, we also found a main effect of desirability that is consistent with the well-documented phenomenon of loss aversion, and a main effect of culture that replicated the finding of study 1. These results show that though enduring cross-cultural differences in impatience exist, consumers’ preferences are labile and can be differentially influenced by message framing. For further information contact: Haipeng (Allan) Chen, University of Miami, P.O. Box 248147, Coral Gables, FL 33124-6554, email: hchen@miami.edu

91

4.3.2 WHAT WILL THE FUTURE BRING? SUBJECTIVE ASSESSMENTS OF FUTURE PROBABILITIES BASED ON PRIOR OUTCOMES IN DIFFERENT DOMAINS

Ashwani Monga, University of Minnesota Akshay R. Rao, University of Minnesota

ABSTRACT Decisions are frequently made without definite knowledge of their consequences (Tversky and Cox 1995). The decision to submit a paper to a journal, invest in the stock market, try a new medical procedure, or bet on a game of chance is based on some expectation of the outcome. Frequently, these expectations are influenced by recent events that may or may not have a bearing on the likelihood of the focal event. That is, people might psychologically adjust their expectations of an outcome based on recent unrelated events in their lives. How do people assess the future following the revelation of an unrelated outcome? Do they update their expectations in a manner consistent with the prior outcome (the “hot hand effect”)? Additionally, does the nature of the prior event (i.e., whether it is in the loss or gain domain) matter? This is the focus of our inquiry. We report on three experiments designed to address the basic question of peoples’ expectations of the future based on the past. Consistent with Johnson and Tversky (1983) who found that inducing positive affect in subjects produced an optimism bias while inducing negative affect produced a pessimism bias, we argue that prior events that are personally experienced have an impact on affect, and this affect is a source of information that has an impact on expectations of the future (Schwarz and Clore 1983, 1988). Therefore the hot hand should be manifested for personally experienced events. Further, we examine how the nature of prior events affects expectations. Specifically, drawing from Prospect theory and regulatory focus theory, we categorize prior outcomes as follows: gains that occur or do not occur, and losses that occur or do not occur. Thus, for instance, one may win $100 (an outright gain) or not win $100 (a non-gain), or one might avoid losing $100 (a non-loss) or lose $100 (an outright loss). We predict that, depending on whether one wins or does not (a comparison of promotion outcomes in regulatory focus theory), or loses or does not (a comparison of prevention outcomes in regulatory focus theory), expectations regarding the outcome of future gambles are likely to be affected such that outright losses are likely to

be more impactful (relative to non-losses) than non-gains (relative to gains). In Study1, we established that experiencing successful prior outcomes (e.g., finding a parking spot, purchasing a gift that was difficult to find) led to significantly higher expectations of success at being granted a loan and receiving an employment offer, relative to when unsuccessful prior outcomes were experienced. Additionally, the willingness to participate in an unrelated gamble was significantly higher when prior outcomes were successful. In Study 2, subjects participated in a gamble with a 50% probability of success or failure, with real money. One half of the subjects were randomly assigned to a “lose gamble” condition, while the other half was randomly assigned to a “win gamble” condition. Further, subjects who lost the gamble were in one of two conditions: they either lost money, or they did not win money ($2 per gamble). Similarly, subjects who won the gamble condition either won money or did not lose money ($2 gamble). As half the subjects were going to lose money in the initial gambles, an outcome assured in the experimental manipulation, they were provided an initial endowment of $6. Thus, subjects with an initial endowment of $6 could lose twice (yielding a terminal state of $2) or not lose twice (yielding a terminal state of $6), while subjects with no initial endowment could win twice (yielding a terminal state of $4) or not win twice (yielding a terminal state of $0). The dependent variable was the expectation of winning a third identical gamble. Our results indicate a significant crossover interaction effect. Subjects who experienced winning or not winning (the promotion state) did not display any differences in their expectations regarding the third gamble, while subjects who experienced outright losses indicated a significantly lower expectation of winning the third lottery relative to subjects who experienced non-losses. A potential confound in study 2 is the different initial monetary states in the promotion versus prevention conditions. Therefore, in Study 3, we endowed all subjects with $4. The remaining procedures were identical to Study 2; subjects could either win or not win, or lose or not lose. The

92

pattern of results from this experiment was identical to that of Study 2. The crossover interaction effect revealed that subjects who lost had lower expectations relative to those who did not lose, while subjects who won had similar expectations as those who did not win. In sum, our results are interesting manifestations of a bias in consumer judgments that may have significant implications for risk taking in domains ranging from investments to journal submissions. For further information contact: Ashwani Monga, University of Minnesota, Carlson School of Management, 321 19th Avenue South, Room 4-300, Minneapolis, MN 55455, email: AMonga@csom.umn.edu

93

4.3.3 EFFECTS OF REGULATORY FIT ON INFORMATION SEARCH AND PERSUASION

Jing (Alice) Wang, Northwestern University Angela Y. Lee, Northwestern University

ABSTRACT Regulatory focus theory (Higgins 1997) offers an interesting and powerful framework to understand consumer information processing, judgment, and behavior. According to Higgins (2000), people experience regulatory fit when they use a strategy of goal pursuit that suits their regulatory orientation. That is, those with a promotion orientation are more eager to pursue gains than guard against non-gains, and those with a prevention focus are more vigilant guarding against losses than striving towards non-losses. More central to the current research, regulatory fit has been shown to result in higher valuation of objects (Higgins et al. 2003), more correct moral judgment (Camacho et al. 2003), as well as more favorable evaluation of products (Aaker and Lee 2001; Lee and Aaker forthcoming). The current research further examines the regulatory fit effect on information search and aims to uncover the mechanism underlying the regulatory fit effect on persuasion. In Study 1, we primed participants’ regulatory focus by presenting them with information that focuses on gains versus losses. Then, participants were presented with information boxes on a computer screen, describing different product features. These features were pre-tested to be more compatible with either a promotion or a prevention goal. Each feature was hidden in a grey box and participants had to move their mouse over each box to review the information contained. The information would disappear when participants move their mouse away from the box. Therefore, the time each feature box stays open reflects the time participants spend on reviewing the feature in the box. The results showed that promotion-focused participants spent more time reviewing information that is more compatible with a promotion goal, whereas prevention-focused participants spent more time reviewing information that is more compatible with a prevention goal. In addition, prevention-focused participants revisited prevention features more often than promotion-focused participants. In Study 2, we primed participants’ regulatory focus by asking them to write about hopes/aspirations versus duties/obligations. Then,

participants were presented with two targets, with one being more attractive along its promotion-focused features but less attractive along its prevention-focused features, and the reverse for the other target. Participants were asked to evaluate the two targets. The results showed that participants primed with a promotion focus evaluated the target with the more attractive promotion features more favorably than those primed with a prevention focus. In contrast, those primed with a prevention focus evaluated the target with the more attractive prevention features more favorably than those primed with a promotion focus. These results showed that participants put more weight on those features that fit with their regulatory goal, and provided convergent evidence that individuals pay more attention to information that is compatible versus incompatible with their regulatory concerns. Taken together, the two studies extended the current research on regulatory fit by demonstrating that the compatibility effect holds when information with different foci coexist in the same message, and that the compatibility effect on persuasion may be accounted for by more extensive information search for and more weight being placed on compatible than incompatible information. For further information contact: Jing Wang, Northwestern University, Kellogg School of Management, 2001 Sheridan Road, Evanston, IL 60208, email: wangjing@kellogg.northwestern.edu\

94

5.1.1 PERCEPTUAL SYMBOLS AND CONCEPTUAL REPRESENTATION OF PRODUCTS AND SERVICES

Ling-Ling Wu, National Taiwan University

Alan J. Malter, University of Arizona Lawrence W. Barsalou, Emory University

ABSTRACT The nature of knowledge representation has profound effects on how consumers understand, recall, and evaluate information about products and services. Standard theories in cognitive science assume that amodal symbols, redescribed from perceptual experiences, construct the representational structure. In contrast, a growing body of empirical results suggests that perceptual symbols play a central role in various conceptual tasks (Kosslyn, 1980; Finke, 1989; Wu, 1995; Mandler, 1992; Tomasello, 1992; Babbes, and Malter, 1997; Glenberg, 1997; Glenberg & Robertson 1999; Glenberg et al. 2003; Kaschak & Glenberg 2000; Soloman & Barsalou, 2001; Prinz & Barsalou, 2002). The aim of the present research is to investigate the alternative view that a system of perceptual symbols serves to represent consumer knowledge of products and services. The results of our experiments shed new light on how consumers conceptualize physical products, service environments, and the experience of consumption. THE THEORETICAL FOUNDATION OF PERCEPTUAL SYMBOL SYSTEMS Perceptual symbols are imagistic representations analogous to human perception of objects and events (Shepard & Cooper, 1982; Shepard, 1984; Finke, 1989; Barsalou 1999). According to Finke (1989), the analogous relationship between imagistic representations and perception is defined in terms of the following four principles: (1) the principle of perceptual equivalence, (2) the principle of spatial equivalence, (3) the principle of transformational equivalence, and (4) the principle of structural equivalence between imagery and perception. Based upon these principles, the experience of activating perceptual representations resembles the experience of actually perceiving an object or event in the absence of direct sensory stimulation. Physiological data demonstrate that visual imagery engages activation of visual cortex (Goldenberg et al. 1989; Cuthbert, Vrana & Bradley, 1991). Moreover, perceptual symbols are multimodal, schematic dynamic representations which are constructed based upon experiential

interaction with the outside world (Glenberg, 1997; Glenberg et al. 2003). All the above theories and empirical results suggest that perceptual symbols share a similar processing mechanism with perception. Furthermore, people always adopt a perspective when actually looking at or imagining looking at something. In fact, imagining looking at something makes it difficult to actually look at something else (Brooks, 1967; 1968). If conceptual content is represented by perceptual symbols, then product attributes viewed from the same perspective should be easier to process than attributes that can only be viewed from different perspectives when subjects are asked to verify attributes of products or services based on their knowledge. RATIONALE AND HYPOTHESES OF THE PERSPECTIVE EFFECT EXPERIMENT If conceptual representations are constructed from perceptual symbols, then consumers might utilize perceptual symbols of various perspectives when they are asked to call upon their knowledge about a product or service. For instance, we might have an image of the internal part and an image of the external part of a house because these are the two perspectives from which we typically view houses. Thus, when we think of an internal attribute of a house (e.g., 'fireplace'), and are then asked to verify other attributes of a house, the response to another internal attribute from the same perspective (e.g. 'carpet') should be faster than the response to an external attribute (e.g. 'balcony'). Conceptualization of the latter uses perceptual symbols from a different perspective than those used to conceptualize 'fireplace'. To test this prediction, we conducted a 2x2 between-subjects experiment. We manipulated the type of attribute (whether or not it would be visible from the front or external surface or whether is was contained inside or hidden behind a surface) and the relationship (same or different perspective) between the attribute used as a prime and the target attribute. As predicted, response times were considerably faster when the prime and target

95

attribute were both visible and both presented from the same perspective. These results support the view that consumers' concepts of both products and services are represented in knowledge by embodied perceptual symbols. REFERENCES Babbes, George S. and Alan J. Malter (1997).

"Embodied Cognition: Towards a More Realistic and Productive Model of Mental Representation". Advances in Consumer Research, 24: 39-41.

Barsalou, Lawrence W. (1999), "Perceptual symbol systems". Behavioral and Brain Sciences 22, 577-660.

Brooks, L. R.(1967). The supression of visualization by reading. Quarterly Journal of Experimental Psychology, 19, 289-299

Brooks, L. R.(1968). Spatial and verbal components of the act of recall. Canadian Journal of Psychology. 22, 349-368.

Cuthbert, B. N., Vrana, S. R., & Bradley, M. M. (1991). "Imagery: function and physiology." In Advances in psychophysiology.

Finke, R.A.(1989)Principles of mental imgery. Cambridge: MA: MIT Press.

Glenberg, Arthur M. (1997), “What Memory is For,” Behavior and Brain Sciences, 20(1), 1-55.

Glenberg, Arthur M. and David A. Robinson (1999), “Indexical Understanding of Instructions,” Discourse Processes, 28, 1-26.

Glenberg, Arthur M., David A. Robinson, Michael P. Kaschak, and Alan J. Malter (2003), “Embodied Meaning and Negative Priming” Behavioral and Brain Sciences, 26(5) forthcoming.

Goldenberg, G., Podreka, I., Uhl, F., Steiner, M., Willmes, K. (1989). Cerebral Correlates of Imaging Colors, Faces and a Map - I. Spect of Regional Cerebral Blood Flow. Neuropsychologia, 27(11/12), 1315-1328.

Kaschak, Michael P. and Arthur M. Glenberg (2000), "Constructing Meaning: The Role of Affordances and Grammatical Constructions in Sentence Comprehension," Journal of Memory & Language, 43, 508-529.

Kosslyn, S.(1980)Image and Mind. Harvard University Press.

Mandler, J.M.(1992).How to build a baby: II. Conceptual primitives. Psychological Review, 99, 587-604.

Prinz, J.J., & Lawrence W. Barsalou, 2002 ”Acquisition and productivity in perceptual symbol systems: An account of mundane creativity”, Creativity, computation, and cognition. Cambridge, MA:MIT/AAAI Press.

Shepard, R.N. & Cooper, L.A. (1982). Mental images and their transformations. MIT Press.

Shepard, R.N. (1984). Ecological constraints on internal representation: Resonant kenematics of preceiving, imagining, thinking, and dreaming. Psychological Review, 91, 417-447.

Solomon, Karen O. and Lawrence W. Barsalou (2000), "Grounding Concepts in Perceptual Simulation, II: Evidence from Property Verification (under review).

Tomasello, M. 1992. First verbs: A case study of early grammatical development. New York: Cambridge University Press.

Wu, L.L. 1995. Perceptual representation in conceptual combination. Doctoral dissertation. University of Chicago.

For additional information contact: Ling-Ling Wu, National Taiwan University, 50, Ln. 144, Sec. 4, Keelung Rd., Taipei, Taiwan 106, R.O.C., email: lingwu@im.ntu.edu.tw

96

5.1.2 BODY BOUNDARY ABERRATION AND BODY ESTEEM AS INFLUENCERS OF CONSUMER ATTITUDES AND INTENTIONS

Ellen Garbarino, Case Western Reserve University

José Antonio Rosa, Case Western Reserve University ABSTRACT We report on two studies that explore the relationship between boundary aberration and body esteem (Rosa and Malter 2003) and consumer attitudes and intentions. Study 1 looks at the influence of body esteem and body boundary aberration on consumer willingness to buy apparel online. Study 2 investigates the influence of body boundary aberration on consumer attitudes and expectations towards their automobiles. Study 1 shows that body esteem has a positive influence on online apparel purchase intentions mediated by involvement and concern with fit, while boundary aberration attenuates concern with fit. Study 2 shows that high boundary aberration reduces sensitivity to the fragility of automobile sheet metal and glass surfaces, and in effect causes consumers to be less concerned with auto performance in that area relative to low boundary aberration. Body boundary aberration refers to the perceived stability and resistance to external forces of our subconsciously-held edges of body containment – the wall between self as contained within the body and the outside world (Fisher 1970). Clinical research has shown that persons suffering from permeable boundaries (e.g., high boundary aberration) might be unable to distinguish between objects they are touching and part of their body (e.g., Schilder 1935). Research has also shown that people compensate for permeable boundaries through what they wear, and can become disoriented when boundary enhancers such as apparel are removed or replaced (e.g., Compton 1964). Body esteem refers to an underlying generalized disposition toward our body, often reflected in positive and negative assessments of body parts (Secord and Jourard 1953). For both studies, the boundary aberration scale (Chapman, Chapman, and Raulin 1978) was adapted for use with consumers. Statements describing body boundary beliefs and experiences such as the detachment or distortion of limbs were interspersed with statements describing other beliefs and experiences, and presented along with a 5-point Likert scale (1 = strongly agree, 5 = strongly disagree).

Study 1 involves an online survey assessing the relationship between boundary aberration and body esteem, and willingness to buy clothing online, controlling for involvement with apparel, concerns with fit, prior online purchase behavior and satisfaction with past purchases. Online retail is the fastest growing sales channel but it allows for only a limited use of embodied knowledge (Rosa and Malter 2003). For most of the frequently purchased online product categories this may be a minor limitation but apparel purchasing (consistently one of the top five categories bought online; Cox 2003) typically requires significant bodily involvement. This suggests that factors such as body boundary aberrations and body esteem may influence people’s willingness to buy apparel online. High levels of body boundary aberration were initially hypothesized to make one less likely to be willing to buy apparel online. For consumers who use apparel to define or defend their body boundaries, we expected the physiological feedback received from trying on apparel to be a necessary part of the purchase evaluation process. The results were opposite of what was expected, however, suggesting that boundary aberration is not consciously addressed by consumers, and reduces sensitivity instead of raising it. Body esteem was expected to be positively correlated with purchase intentions, because people who are satisfied with their appearance will be more confident in their ability to assess the attractiveness and fit potential of garments mentally, and hence will be less concerned with physically experiencing apparel items before purchasing them. Results show that body esteem indeed has a positive influence on purchase intentions, but only for people who have high involvement with apparel in general while at the same time having low concern with the fit of apparel. In study 2, Q-methods were used to gather consumer perceptions of different aspects of automobiles. Nine dimensions emerged from the factor analysis, with perceptions of the fragility of automobile sheet metal and glass surfaces being the

97

one dimension in which high boundary aberration respondents differed from low boundary aberration respondents. This is consistent with the results from Study 1 which showed reduced sensitivity to the edges or boundaries of wearable products among high boundary aberration respondents. REFERENCES

Chapman, Loren J.; Chapman, Jean P.; Raulin, Michael L. (1978). “Body-image aberration in schizophrenia. Journal of Abnormal Psychology, 87, 399-407.

Compton, N. (1964). Body image boundaries in relation to clothing fabric and design preferences of a group of hospitalized psychotic women. Journal of Home Economics, 56, 40-45.

Cox, Beth (2003). Rising tide lifts E-commerce boats, Cyberatlas, Jan. 2003, http://cyberatlas.internet.com/markets/retailing/article/0,,6061_1568651,00.html

Fisher, Seymour (1970). Body experience in fantasy and behavior. New York: Meredith Corporation.

Rosa, Jose Antonio and Alan J. Malter (2003), E-(Embodied) Knowledge and E-Commerce: How Physiological Factors Affect On-line Sales of experiential Products. Journal of Consumer Psychology, 13, 63-74.

Schilder, Paul (1935/1950). The image and appearance of the human body. New York: International Universities Press.

Secord, Paul F. & Jourard, Sidney M. (1953). The appraisal of body cathexis: Body cathexis and the self. Journal of Consulting Psychology, 17, 343-347.

For additional information contact: José Antonio Rosa, Case Western Reserve University, 339 Peter B Lewis Bldg., 10900 Euclid Avenue, Cleveland, OH 44106-7235, email: jar23@cwru.edu

98

5.1.3 MENTAL SIMULATION AND THE EFFECTIVENESS OF ONLINE SALES AIDS FOR EXPERIENTIAL PRODUCTS

Alan J. Malter, University of Arizona

José Antonio Rosa, Case Western Reserve University ABSTRACT This paper tests and extends Rosa and Malter's (2003) model of the effects of embodied factors on online purchases of experiential products. This model is based on the view that mental simulation is a central cognitive process (Barsalou 1999) that allows consumers to imagine themselves interacting with experiential products (e.g., apparel) and thus evaluate these products without actually touching them. In this way, consumers may be able to overcome some of the major limitations of shopping for experiential products online (i.e., using a two-dimensional computer interface). A number of online sales aids, such as "My Virtual Model," have been developed and deployed commercially to assist in facilitating these effects (Ives and Piccoli 2002). A general neurological capacity for empathy (see Preston and de Waal 2002; Iacoboni and Lenzi 2002) may enable even crude graphic representations of the body to stimulate mental simulations of action that provide a very realistic, lifelike consumption experience for online shoppers. However, individual differences in consumers' need for touch (Peck and Childers 2003a; 2003b), body image (Cash and Szymanski 1995), visual style of processing (Heckler et al. 1993) may serve to enhance or undermine the potential effectiveness of such online sales aids. We conducted a series of experiments to test theoretical predictions about factors that will influence the effectiveness of perceptual online sales aids such as "My Virtual Model" (MVM). In the first study, we established that consumers are likely to report factually accurate personal data about their height, weight, and other physical features that enable them to build a reasonably isomorphic virtual model of themselves. Consumers who prefer a more highly visual style of processing perceive their MVM as looking more like them and are more satisfied with the way their MVM looks in clothes that can be "tried on" the model. The second study tested the effect of the experience of building and using a personal MVM (compared to simply evaluating pictures of the same articles of clothing in an online catalog). It appears that the experience of trying clothes on a personal MVM shifts preferences for how

consumers would like to see clothing displayed (i.e., on a model that looks like them, compared to a photo of a super-model). However, using a personal MVM did not lead to a more positive evaluation of the sample clothing or increase shopper's readiness to purchase the items. It is possible that using a personal MVM enables online shoppers to scrutinize articles of clothing much more carefully, just as they do when actually trying on clothes in a physical store, and that extreme positive and negative evaluations cancel each other out. This effect could benefit online retailers if their priority is to achieve faster online purchase decisions and to minimize product returns by helping consumers decide which items do not suit their needs and thus move on to consider other items that might be more to their taste. Future experiments will directly tests the effect of physical changes on consumers' use of the MVM in evaluating clothing online. Our theory predicts that physical movement by online shoppers will enhance the realism of mentally simulating actions of the MVM and hence its effectiveness as an embodied decision tool. REFERENCES

Barsalou, L. W. (1999), "Perceptual symbol systems". Behavioral and Brain Sciences 22, 577-660.

Cash, T.F., & Szymanski, M.L. (1995). The development and validation of the Body-Image Ideals Questionnaire. Journal of Personality Assessment, 64, 466-477.

Heckler, Susan E., Terry L. Childers and Michael J. Houston (1993), "On the Construct Validity of the SOP Scale," Journal of Mental Imagery, 17 (3&4), 119-132.

Iacoboni, Marco and Gian Luigi Lenzi (2002), "Mirror Neurons, the insula, and empathy," Behavioral and Brain Sciences, 25, 39-40.

99

Ives, Blake and Gabriele Piccoli (2003), "Custom Made Apparel and Individualized Service at Lands' End," Communications of the Association for Information Systems, 11, 79-93.

Peck, Joann and Terry L. Childers (2003a), “To Have and To Hold: The Influence of Haptic Information on Product Judgments,” Journal of Marketing, April 67 (2).

Peck, Joann and Terry L. Childers (2003b), “Individual Differences in Haptic Information Processing: On the Development, Validation, and Use of the 'Need for Touch' scale,” Forthcoming in Journal of Consumer Research, December 2003.

Preston, Stephanie D. and Frans B.M. de Waal (2002), "Empathy: Its Ultimate and Proximate Bases," Behavioral and Brain Sciences, 25, 1-72.

Rosa, Jose Antonio and Alan J. Malter (2003), E-(Embodied) Knowledge and E-Commerce: How Physiological Factors Affect On-line Sales of experiential Products. Journal of Consumer Psychology, 13, 63-74.

For additional information contact: Alan J. Malter, The University of Arizona, Department of Marketing, Eller College of Business and Public Administration, P.O. Box 210108, Tucson, AZ 85721-0108, email: amalter@eller.arizona.edu

100

5.2.1 THE IMPACT OF COUPON VALUE, PURCHASE RECIPIENT, AND TIME OF REDEMPTION ON WHAT I THINK THEY THINK OF ME: AN INVESTIGATION OF METAPERCEPTIONS IN

COUPON REDEMPTION SITUATIONS

Jennifer J. Argo, University of Alberta Kelley J. Main, University of British Columbia

ABSTRACT

Redemption rates for coupons are less than 10% (Bawa, Srinivasan and Srivastava, 1997) and researchers have tried to identify the cause of these low rates. One potential explanation for these low rates that has not received prior attention in the literature is that individuals may be concerned about how they believe they are being viewed by others, referred to in the literature as metaperceptions (Kenny and DePaulo, 1993). To understand the role of metaperceptions and how they may influence coupon redemption rates, two studies were conducted. Study 1 identified the metaperceptions associated with coupon redemption and demonstrated that the intended recipient of the purchase moderated the impact of coupon value of consumers’ metaperceptions. Further, the study showed that different metaperceptions exist for various social presences, depending on their role in the retail context. Study 2 extended the investigation into a service situation and showed that the time of redemption also influences the impact of coupon value on consumers’ metaperceptions

INTRODUCTION Research in consumer promotions has found

that the redemption rates for coupons are less than 10% (Bawa, Srinivasan, and Srivastava, 1997). A variety of explanations have been forwarded in an attempt to understand these low rates of redemption including the tendency for coupons to expire and the effort of remembering to take them to the store (Bonnici, Campbell, Fredenberger, and Hunnicutt, 1996/97). Another potential explanation that has not received attention in the consumer behavior literature is the possibility that individuals’ unwillingness to redeem coupons may be related to their concerns about how they believe they are being viewed by others, referred to as metaperceptions. Metaperceptions are the thoughts and expectations that a person believes another will have of them (Kenny and DePaulo, 1993). Previous research in psychology has shown that metaperceptions influence individuals’ self-

perceptions, self-image, self-esteem, and behavior even in short interactions in the presence of strangers (Kenny and DePaulo, 1993). In the current context, we expect that consumers will have different metaperceptions about how they are being viewed by others when they redeem coupons depending on characteristics of the coupon and the social environment.

LITERATURE REVIEW Metaperceptions Metaperceptions are an individual’s

perceptions of how the self is viewed by others (Kenny and DePaulo, 1993). To illustrate, John believing that Susan thinks he is nice, is an example of John‘s metaperceptions. To date research has examined the metaperceptions that people have in a variety of contexts (e.g. Vorauer, Main, and O’Connell 1998); however, they have not been considered in a consumer setting. It is important to consider the metaperceptions that consumers think others in the situation (such as cashiers or other shoppers) have of them when they redeeming coupons.

One factor that may influence consumers’

metaperceptions is the value of the coupon they are redeeming. Previous research has shown that consumers tend to perceive the use of a small coupon as humiliating (Bonnici et al., 1996/97) whereas they are more likely to perceive the use of a large valued coupon as attractive (Bawa, Srinivasan, and Srivastava, 1997; Schindler, 1998). Therefore, it is expected that consumers may believe that other people will think they are cheap-shoppers (i.e., cheap-shopper metaperceptions) if they redeem a low value coupon but believe they will be perceived as a smart-shopper (i.e., smart-shopper metaperception) if they redeem a high value coupon.

However, it is also expected that the intended

recipient of the purchased product will influence consumers’ metaperceptions and their likelihood of redeeming coupons. In particular, if the purchase is intended as a gift to another person as opposed to

101

an item for oneself, this may influence how consumers believe that others view them when using a coupon. The gift giving literature has shown that the exchange of gifts communicates symbolic information in a relationship (Ruth, Otnes, and Brunel, 1999). Research in marketing has also indicated that an association exists between the type of giver/receiver of the gift and the amount of money spent on the gift (Belk, 1979). Caplow (1982) demonstrated that different relationships between the giver and the receiver determine the number and dollar value of gifts such that the closest relatives receive gifts that are the highest in cost. Since gifts can be used to show how much a consumer cares for another person (Cheal, 1987), it is expected that if a consumer were to devalue the importance of a relationship by using a coupon to purchase the gift, consumers would be expected to have greater cheap-shopper metaperceptions. However, if the item is purchased for oneself, it is expected that the value of the coupon would not matter as much.

It is also possible that the role of others in the

consumption context may influence consumers’ metaperceptions during coupon redemption. Marketing researchers who have investigated the impact of a social influence on consumers have primarily examined the influence of an interacting social presence (e.g., a salesperson greeting a consumer; Moschis 1976). However, little research effort has been directed to testing the influence of a non-interacting social presence on a consumer (e.g., another shopper who is physically present in the aisle but with whom an individual does not interact), despite findings that the mere presence of another person can have an impact on how consumers think, feel, and behave (Argo 2003; Dahl, Manchanda, and Argo 2001). Given that both an interacting and a non-interacting social influence may impact consumers, the first study determines the metaperceptions that arise in these two situations.

STUDY ONE This study used a scenario to implement a 2

(coupon value: $0.25 vs. $1.00) x 2 (purchase recipient: self vs. other) between-subjects experimental design.

Trait selection. Prior to conducting the study, twenty-one participants were asked to list a variety of words and phrases that they associated with coupon redemption. Seventeen traits were identified which were later presented to study participants.

Independent Variables. Coupon value was

manipulated by asking participants to imagine redeeming a coupon worth $0.25 or $1.00 off a $4.95 product. Purchase recipient was manipulated by indicating in the scenario that participants either purchased an art card for themselves (self) or a birthday card for their best friend (other).

Procedure. Seventy (36 males and 34 females)

undergraduates received course credit for their participation. Participants were first asked to imagine that they entered a store in search of a card. Cards were chosen as the product to be purchased as they can be purchased for oneself (i.e. an art card) or for another person (i.e. a birthday card). In addition, birthday cards have been shown to play a symbolic role in communicating one’s feelings about another person (Dodson and Belk, 1996). While selecting their product, the participants noticed another person standing next to them examining items at another display. They also noticed a coupon (with a value of either $0.25 or $1.00 off) that was available for their product. They made their purchase selection and they went to the cashier to pay for their product. After reading the scenario, participants completed a short questionnaire that asked them to complete the seventeen traits (seven-point (not at all/extremely)) identified during trait selection. Exploratory factor analysis indicated that items were related to two underlying dimensions (69% of the variance was explained). The first dimension consisted of six items: cheap, penny-pincher, financially poor, thrifty, bargain hunter, and money conscious, which were combined to form an overall cheap-shopper metaperceptions index (α = .75). The second dimension was comprised of three items: competent, capable, and clever, which were combined to form a smart-shopper metaperception index (α = .74). Lastly, they completed manipulation checks that asked them to identify the product they purchased as well as the value of the coupon, demographic information, and a suspicion probe. Three participants were unable to correctly recall the coupon value and/or the product they were purchasing and were removed from the analysis. As none of the demographic information impacted the results, it is not discussed further. Finally, no participants were aware of the experimental hypotheses.

Results and Discussion ANOVA was first conducted to

determine the impact of the manipulated variables on consumers’ metaperceptions

102

relative to the cashier (i.e., an interacting social presence). The results produced a main effect for purchase recipient for participants’ cheap- and smart-shopper metaperceptions (cheap: F (1, 66) = 5.95, p < .05; smart: F (1, 66) = 3.52, p < .05). An examination of the means indicates that participants believed that the cashier thought they were more cheap (less smart) when they were using a coupon to purchase a card for a friend (Mcheap = 3.83; Msmart = 4.39) as compared to themselves (Mcheap = 3.25; Msmart = 4.83). Coupon value was not significant (F < 1).

Analyzing consumers’ metaperceptions

relative to the other shopper (i.e., a non-interacting presence) ANOVA again produced a main effect of purchase recipient for smart-shopper with the same pattern of means as described for the cashier (F (1, 66) = 4.62, p < .05; Mbest friend = 3.97, Mself = 4.47). Additionally, a significant interaction between purchase recipient and coupon value for cheap-shopper metaperceptions resulted (F (1, 66) = 4.63, p < .05). Simple effects tests indicated that participants had more cheap-shopper metaperceptions when they purchased a card for their best friend using a $0.25 coupon (Mbest friend and

$0.25 = 4.04) as compared to an art card for themselves (Mself and $0.25 = 2.94, t (30) = 2.54, p < .05). When the purchase was an art card, cheap-metaperceptions did not differ based on coupon value (Mself and $1.00 = 3.41, t (32) = 1.30, p > .20).

The findings of this study demonstrate that the

value of the coupon only influences cheap-shopper metaperceptions when the consumer is in the presence of a non-interacting social presence. These findings are noteworthy because most previous research has not considered the impact of a non-interacting influence upon metaperceptions in a coupon redemption situation. In addition to product settings, research has indicated that coupons are often used in service situations (Hartley and Cross, 1988) ranging from hair salons to restaurants (Narasimhan, 1984); however, little is known regarding consumers’ responses to the use of coupons in these situations. One of the main differences between a product setting as compared to a service setting is the potential for relationship development in the latter situation. In a service setting, such as a hair salon, consumers are likely to develop a relationship with their service provider over time (Gabarino and Johnson 1999). If consumers plan to establish stronger or longer lasting relationship with a service provider (i.e., return for another haircut) they may avoid

threatening the relationship by devaluing the service that has been received through the use of a coupon. Another difference between a service encounter and a product purchase setting is that the latter can last for extended periods of time (i.e., consumers can spend hours in a restaurant or hair salon). Therefore, the point of time at which consumers redeem their coupon can vary to a greater extent then for product purchases. In the second study, we test the impact of coupon value in a service context and we introduce the time of coupon redemption as another factor that may influence consumers’ metaperceptions.

Finally, in a survey conducted by Bonnici et al,

(1996/1997), consumers rated embarrassment as the most important factor influencing their tendency to redeem coupons. Therefore, one final goal of the next study is to examine whether the embarrassment consumers feel when redeeming coupons mediates the impact of coupon value and point of redemption on metaperceptions.

STUDY TWO Using a scenario methodology, this study used

a 2 (coupon value: $0.75 vs. $7.50) x 2 (time of redemption: before vs. after) between-subjects experimental design.

Independent Variables. Coupon value was

manipulated by asking participants to imagine redeeming a coupon worth $0.75 or $7.50 off a $15.00 product. The time of redemption manipulation was achieved by describing in the scenario that they notice a sign that indicates they are required to present any coupon they might have to the hair stylist either before or after they receive the service.

Procedure. Sixty-three (32 males and 31

females) participated in this study for course credit. Participants received a short scenario that asked them to imagine that they had just moved to town and were going to a new hair salon for which they had a coupon (valued at either $0.75 or $7.50 off) to get a haircut. As they entered the salon, they noticed a sign with information about the salon’s policy on when coupons should be redeemed (either before or after the service). After receiving their haircut, they paid for the service and left the salon. After reading the scenario, participants completed a questionnaire using the same trait measures identified in Study 1. Additionally, to test for embarrassment, participants indicated how embarrassed, awkward, and uncomfortable (α =

103

.82) they felt on three seven-point (not at all/extremely) item scales (Dahl, et al., 2001). Finally, participants completed demographic information, were asked to recall the coupon value and the time of redemption, and completed a suspicion probe. None of the demographic information impacted the results; therefore, it is not discussed further. Although no one was aware of the experimental hypothesis, one participant was unable to correctly recall the value of the coupon and was excluded from the analysis.

Results and Discussion ANOVA revealed a significant interaction

between coupon value and point of redemption for cheap-shopper metaperceptions (F (1, 61) = 4.38, p < .05) but not for smart-shopper metaperceptions (F’s < 1). Simple effects tests indicated that participants who redeemed a $0.75 coupon before the service was provided reported significantly higher cheap-shopper metaperceptions (M = 5.07) than those who redeemed the coupon after the service (M = 4.11, t (28) = 2.25, p < .05). Further, cheap-shopper metaperceptions were significantly higher when participants redeemed a high value coupon after (M = 4.99) as compared to a lower value one (M = 4.11, t (29) = 2.09). These results demonstrated that consumers have similar metaperceptions in a service setting as a product setting and that time of redemption of the coupon can have a significant impact on metaperceptions.

One final objective of the second study was to

determine whether embarrassment mediated the impact of coupon value and point of redemption on consumers’ metaperceptions. Following Baron and Kenny (1986), the first test for mediation was significant for cheap-shopper metaperceptions (F (1, 61) = 4.38, p < .05), but not smart-shopper metaperceptions (F < 1). In order to further test for mediation, the independent variable(s) and/or their interaction must significantly influence the mediator. The results of this analysis were not significant (F’s < 1) suggesting that mediation does not exist.

GENERAL DISCUSSION In general, this research has demonstrated a

number of interesting findings. First, two distinct types of metaperceptions exist for consumers who are redeeming coupons: cheap-shopper and smart-shopper metaperceptions. Second, the intended purchase recipient moderates the impact of coupon value on consumers’ cheap-shopper metaperceptions when they did not interact with the

social presence (i.e., it was the other shopper). In particular, results indicated that participants had more cheap-shopper metaperceptions when they purchased a card for their best friend with a low value coupon as compared to purchasing an art card for themselves. However, consumers’ metaperceptions relative to an interacting social presence (i.e., the cashier) indicated that participants had cheap-metaperceptions when they used a coupon to purchase a card for friends as compared to one for themselves. Third, in the context of a service situation, our findings show that low-value coupons lead to cheap-shopper metaperceptions more when the coupon is redeemed before (vs. after) receiving the service. In contrast, consumers had higher cheap-shopper metaperceptions when they presented the service provider with a high value coupon (vs. low value) after the service was received. Finally, results of the mediation analysis indicated that embarrassment did not serve as a significant mediator of the relationship between coupon value and time of redemption on metaperceptions. While some of the initial questions concerning consumers’ metaperceptions in a coupon redemption situation have been addressed, future research could be directed to examining these issues in an actual consumption setting or in other promotional contexts. REFERENCES Argo, Jennifer J. (2003), “Social Audiences in a

Retail Setting: An Investigation of Their Influence on Consumers’ Emotions, Cognitive Performance, and Self-Presentation Behaviours,” unpublished Ph.D. dissertation, University of Manitoba.

Bawa, Kapil, Srini S. Srinivasan, and Rajendra

Srivastava, “Coupon Attractiveness and Coupon Proneness: A Framework for Modeling Coupon Redemption,” Journal of Marketing Research, 34(November 1997), 517-525.

Bonnici, Joseph, David P. Campbell, William B.

Fredenberger, and Kathryn Hunnicutt, “Consumer Issues In Coupon Usage: An Exploratory Analysis,” Journal of Applied Business Research, 13(1 1996/97), 31-40.

Dahl, Darren W., Rajesh V. Manchanda, and

Jennifer J. Argo, “Embarrassment in Consumer Purchase: The Roles of Social Presence and Purchase Familiarity,”

104

Journal of Consumer Research, 28(December 2001), 473-481.

Dodson, Kimberly J. and Russell W. Belk (1996),

“The Birthday Card Minefield,” in Advances in Consumer Research, ed. Kim Corfman and John G. Lynch, Jr., Provo, UT: Association for Consumer Research, 14-20.

Gabarino, Ellen and Mark S. Johnston (1999), “The

Differing Roles of Satisfaction, Trust and Commitment in Customer Relationships,” Journal of Marketing, 53(2), 70-88.

Gilovich, Thomas, Victoria Husted Medvec, and

Kenneth Savitsky, “The Spotlight Effect in Social Judgment: An Egocentric Bias in Estimates of the Salience of One’s Own Actions and Appearance,” Journal of Personality and Social Psychology, 78, (2 2000), 211-222.

Hartley, S.W. and J. Cross (1988), “How Sales Promotion Can Work For and Against You, The Journal of Consumer Marketing, 5 Summer, 35-42.

Iacobucci, Dawn (1998), “Services: What Do We Know and Where Shall Ee Go?” in Advances in Services Marketing and Management, ed. T. A. Swartz, D. E. Bowen, and S. W. Brown, JAI Press Inc, Greenwich, 1–96.

Kenny, David A. and Bella M. DePaulo, “Do

People Know How Others View Them? An Empirical and Theoretical Account,” Psychological Bulletin, 114(1 1993), 145-161.

Narasimhan, Chakravarthi (1984), “A Price Discrimination Theory of Coupons,” Marketing Science, 3 Spring, 128–147.

Vorauer, Jacquie D., Kelley J. Main, and Gordon B. O'Connell (1998), “How Do Individuals Expect to be Viewed by Members of Lower Status Groups? Content and Implications of "Meta-Stereotypes,”” Journal of Personality and Social Psychology, 75(4), 917 – 937.

105

5.2.2 SPATIAL POSITIONING: THE VALUE OF CENTER-STAGE

Ana Velenzuela, San Francisco State University

Priya Raghubir, University of California, Berkeley ABSTRACT Does placing a product in a central, peripheral, or extreme-end position systematically affect consumers’ attitudes toward the brand? Surprisingly, this issue has not been investigated by consumer psychologists despite the importance of shelf placement in a consumers’ brand choice decision, a manufacturer’s distribution decision, and a retailer’s shelf space pricing decision. Other than isolated studies in psychology with inconsistent position effects (e.g., Taylor and Fiske 1975, Study 1 versus Study 2), the effect of visual placement on attitudes and preferences does not appear to have been systematically researched by psychologists. While psychological “positioning” is a well-known concept to marketers, oddly, the literal, spatial analogy on which it is based, is under-researched (except for Hotelling type models of store location with a distance cost to consumers, see Eppli and Benjamin 1994 for a review). This paper introduces the important and novel concept of “spatial positioning,” in a literal sense, with psychological overtones to marketing. The introduction of the “spatial positioning” concept leads to a set of interesting theoretical questions: If there is an advantage to position, what positions have this advantage over others? While some literature in survey methodology (Schuman and Presser 1996) and impression formation (Asch 1946) predict a primacy or recency effect implying that products at either end of an array would be more likely to be chosen, the literature on visual perception would suggest that the item mid-way between the start and the middle of the array would receive most attention (e.g., Ducrot and Pynte 2002). On the other hand, the literature on spatial perceptions based on the asymmetric hemispheric view of the brain suggests that the left- or right-hand position preference is contingent on whether the stimulus is verbal or visual (Janiszewski 1990). Finally, the social psychology literature on salience and vividness predicts that items in the center would be most salient, most likely to be recalled, and liked the best (e.g., McArthur and Post 1977). This paper examines if, why and when the position of a product in an array confers an advantage. We

propose that the position in an array is informative in and of itself. Across a variety of contexts and domains, there are certain “spatial” rules that appear to govern the ordering of people, items, and things in an array. There are learned associations that the “best” option is frequently in the middle (e.g., in a medal display the gold medal is in the middle; in a beauty contest, the winner is in the middle; in an audience, the most important speaker is in the middle). People may, therefore, subconsciously believe that the option presented in the middle is the most popular or best option. In a product display context, we propose that consumers use a “Center-is-Better” rule of thumb, as they believe that the order of items is itself informative. The use of this simple learned rule leads to higher evaluations for products in the center of an array as compared to those at the ends of the array, even if such products are no better recalled. Consistent with our proposed model, Mason (1982) found that 5-symbol arrays yielded a U-shaped reaction time function with the poorest performance for symbols at the beginning or the end of the array. Understanding the reason why there is a position advantage is not only theoretically interesting, but also has managerial implications. For example, if a position advantage were due to higher attention, then alternate attention getting packaging may be used to substitute or complement a poor shelf space position. On the other hand, if it were due to higher recall of a brand in a particular position then this would imply that mnemonic devices may be called for to increase the top of mind recall of a brand. If it were due to perceptual process due to the asymmetric construction of the brain hemispheres, then there is little managers can do to reverse this effect – they can only leverage it through placing visual/ verbal material in the most appropriate position. If it were due to improved attitudes towards the brand, but these attitudes were not mediated by attention, but rather reflected an overall schema or belief such as “Center is Better,” this would imply that managers should be willing to pay a premium to have their brands occupying the premium position in a layout. Retailers would be able to use this information to bargain with manufacturers for shelf-space benefits. Finally, if position effects are due to an automatic, unthinking behavioral response (e.g., of picking up the first

106

item, last item, or central item from a display), this would imply that any sales spike enjoyed by a brand might be short-lived to the extent it enjoyed a premium position. Three laboratory experiments with frequently purchased packaged goods show that i) there is a center-stage advantage: a product in a central position is preferred over one at either end of an array (Study 1: Pretzels); ii) the effect is moderated by familiarity, such that the effects are stronger for higher (versus lower) purchase frequency (Study 2: Cookies); iii) the effect is moderated by brand preference: for preferred brands being in the center reduces the downside of being “hurt,” whereas for less preferred brands it helps (Study 3: Cream Cheese). Across studies, we found no evidence that the recall of the brands was contingent on the order in which they were presented. The effects on attitudes and intentions are interpreted in terms of consumers using a heuristic: “Center is Better” whereby they believe that by virtue of being in the central positions a product is better than others that are in extreme positions. Such a heuristic, that the “order is itself informative,” is more likely to be used when consumers do not sequentially process the items in the array. This happens in situations where there is high product familiarity, and for brands with a lower market share, which may be indicative of greater uncertainty regarding brand quality. For further information contact: Priya Raghubir, Haas School of Business, University of California, Berkeley, CA 94720-1900; phone: 510-643-1899; fax: 510-643-1420; e-mail: raghubir@haas.berkeley.edu.

107

5.2.3 WHAT’S UP DOC? DTC PHARMACEUTICAL ADVERTISING AND PATIENT DRUG REQUESTS

Rhonda W. Mack, Ph.D., College of Charleston

Julia E. Blose, Ph.D., College of Charleston Rita A. Balaban, Ph.D., College of Charleston

ABSTRACT

Since its inception in the early 1980s, direct-to-consumer prescription drug advertising (DTC) in the United States has received both praise and criticism. Ignoring the cost implications, this study examines the impact of DTC drug advertising on the patient-doctor relationship for patients who have requested a drug they have seen advertised.

BACKGROUND DTC Regulation In 1962, oversight of prescribed drug advertising was transferred from the Federal Trade Commission (FTC) to the Food and Drug Administration (FDA). Jurisdiction for non-prescribed, or over the counter drugs, remained with the FTC but was also transferred to the FDA in 1988 as part of the Federal Food, Drug, and Cosmetic Act. This act indicated that the FDA regulates labeling of drugs, also including their advertising. At that time, ads were required only to provide a brief summary describing the “drug’s adverse experience profile, contraindications, warnings, precautions, and indications for use,” (Craig 1998). Drug companies could specify the name of the drug, or the condition it treated, but were required to not name both, which consumers found confusing (Brown 2001). A clarification of the act in 1997 by the FDA allowed for the naming of the drug as well as the illness/condition in the same ad without full disclosure of risks. Companies must name only the major risks and provide statements guiding viewers to sources of information such as toll-free 800 numbers, ads, health care professionals, or Web sites in broadcast advertising (Belkin 2001). Print ads still require a brief summary of all possible downsides of the drug. Drug manufacturers are required to provide the FDA with each of their new promotions for prescription drugs when they are first used. The company, however, does not have to wait on FDA approval to run the advertisement. If the FDA does find the drug ad misleading or untruthful, a “notice of violation” is sent to the manufacturer at which

time they must either withdraw or change the ad, or even initiate corrective communications in the event of an FDA “warning” decision. No fines are imposed in this process; the FDA has no power to levy them in this area. The review process is typically lengthy due to understaffing of the FDA. A 2003 issue of Consumer Reports indicated the FDA is staffed by approximately 30 reviewers who deal with about 30,000 submissions annually. To worsen matters, the Department of Health and Human Services started requiring a legal review of all proposed FDA regulatory letters in 2002. Their purpose for this additional review is to put teeth into the FDA’s letters indicating that it will actually pursue legal recourse if the company doesn’t comply. As a result of the additional legal review step before the letters are sent out, approximately 2-11 weeks have been added to the time required for the FDA to issue its letter to the drug company. They sent out more than one hundred letters in the relevant years before 2002, with just 24 being sent out as of November of 2002. This “slow down” is especially problematic since the resulting time period of review now often exceeds the normal ad campaign time period—the ad has already run its course and the campaign finalized. Consumer Reports (2003) performed a review of FDA regulatory letters from January 1997 to November 2002 reporting there were many misleading messages either exaggerating drug effectiveness or minimizing drug risks, making false competitive claims and promoting unapproved uses of drugs and, in some cases, even advertising experimental stage drugs. Currently DTC drug ads are legal in only two countries, the United States and New Zealand. The European Commission recently allowed a provisional period of time for DTC ads for certain conditions (AIDS, asthma and diabetes) (Lerer 2002), but ended the temporary practice dissatisfied with the resulting advertisements. A report by the Health Action International Europe organization reported that one fourth of DTC drug ads in the United States and one-third of those in New Zealand had been found to be in violation of national laws (Medawar 2002).

108

Cost Implications National pharmaceutical costs have been projected to increase by 10-13 percent per year until 2007 (Wechsler 2001), with national pharmaceutical costs projected at $243 billion by that time. DTC drug spending in the United States increased to $2.49 billion in 2001 from only $859 million in 1997 according to data research company CMR (Burton 2002). Three companies spent 44% of the DTC funds in 2002, with the biggest spender Nexium at $211 million followed by Clarinex and Allegra (DTC Perspectives 2003). CMR reported that General Motors Corp. spent $55 million for Prilosec (the little purple pill) in one year, with their drug costs increasing 14% to $1.3 billion. Prilosec costs an estimated 13 times more than its generic option. A report published in DTC Perspectives indicated that DTC advertising totaled $2.7 billion in 2001 for $172 billion in US pharmaceutical sales that year pointing out that DTC accounts for a small percentage of drug sales and less than 1% of the country’s total annual healthcare costs (Tacconi 2003). A reported 24 million people are diagnosed of previously untreated conditions because of DTC ads according to the article. This reportedly leads to the patients living longer and happier lives partly due to the ads motivating them to visit their physician and discuss the conditions that possibly would have gone undiagnosed. Putting the outlay for DTC pharmaceutical advertising into perspective, Glasgow (2002) notes that 2000 DTC spending was 2.2% of product sales as compared to 11.8% of product sales expenditures for promotion to professionals in the industry. Empowering the Consumer

The flip side of the cost of DTC pharmaceutical advertising deals with the value of information supplied to the consumer. Given the increasing scarcity of the physician’s time with a patient, DTC drug ads may serve a major informational role. Some suggest that DTC ads educate consumers on medical conditions possibly aiding in increased office visits, better informed consumers regarding a drug’s side effects, treatment options and possibly undiagnosed medical problems. (Roizen 2002, Burton 2002). At the 2002 DTC National Conference and Exposition, Dr. David Kessler, former FDA commissioner, shared data indicating that 30% of consumers who are exposed to DTC ads visit their doctor in response to the ad. (Youner 2002) Forty-four percent of the doctors subsequently prescribe a

prescription for the advertised product. Bloom (1999) reported that the pharmaceutical requests of consumers changed significantly after the FDA eased regulations on DTC drug advertising in 1997 (both the name of the product and the condition it treats can be included in the ad). Of an estimated 163 million consumers who had seen DTC ads, 54 million consumers talked with their physicians about drugs they had seen advertised with a resulting 12 million prescriptions written. While Bloom reports that approximately 30% more requests were made by consumers for brand name drugs than prior to 1997, she also states that consumers find DTC ads to be unclear feeling the ads do not provide good information about the side effects or benefits of the drug. Most consumers do not know what the advertised medicines are intended to treat, after they have seen the ads.

Many doctors may not know the costs of

the drugs they are prescribing (Bloom 1999). One study found that doctors under-estimated the cost of brand-name drugs (what an uninsured patient would pay) more than 90% of the time and they overestimated the cost of generic drugs. This could have major impact given claims that DTC ads are largely for more expensive brands.

The American Cancer Society uses DTC

ads to promote colon cancer testing (Mickelson 2002) supporting the value of the informational role the ads can play. A National Medical Association survey of African-American doctors’ acknowledged the positive value of drug ads in “raising awareness of medical conditions and treatment” (Allison-Ottey 2002) with over half of the responding physicians indicating the drug ads provided a positive benefit for their patients. Others point to the negative informational role that DTC ads may result in (Wilkes et al (2002) reporting that the side effects (valuable information) is usually buried somewhere in the advertisement’s narrative and questioning the value and educational impact of the ad. Physician-Patient Relationship An additional concern about DTC pharmaceutical ads revolves around the patient-physician relationship. The traditional physician/patient role has been impacted greatly by DTC ads. Previously, in somewhat of a simplified fashion, the patient would enter the doctor’s office, describe his symptoms, be examined, answer additional relevant questions from the doctor, and receive a diagnosis and prescription for a pharmaceutical product deemed appropriate by the

109

doctor. With DTC, the patient now enters the doctor’s office, describes his symptoms, provides the diagnosis from the ad he/she has seen, and requests the advertised pharmaceutical product. The resulting last step in this scenario can vary from the doctor quickly complying with the patient’s request, to some effort to educate the patient, to denying the patient’s request and prescribing an alternative (maybe a better one, maybe a less expensive one). The patient and the doctor may, or may not, be satisfied with their interaction and their individually perceived “relationship” would be formed from the initial and possible continued interactions of the patient and physician. Rosenberg (2001) raises the likelihood that DTC ads may introduce possible tension between the doctor and the patient. This may be due to a number of reasons such as the physician not thinking that the requested drug is the best drug for the patient or that other options than drugs should be considered first (lifestyle management). A pre-visit, self-diagnosed patient introduces different and new dynamics to this traditional relationship. Glasgow (2002) points to the possibility that patients requesting a prescription that the physician grants, may better adhere to the drug therapy recommended, thus improving health care. The National Medical Association survey referred to earlier also reported that drug advertising has a positive effect on the doctor-patient relationship while at the same time calling for drug companies to provide information to doctors of new marketing campaigns directed to patients. (Allison-Ottey 2002) One earlier report, however, indicated that the patient may actually mistrust the doctor if he does not comply with their request for a particular drug which they are convinced they need (Maguire 1999).

One interesting concept dealing with

process production that may be relevant to the patient-physician relationship is the degree to which the patient/customer participates in the production of the physician-visit/diagnosis process. Bendapudi and Leone (2003) investigated the psychological implications of co-production of a process and found psychological implications regarding satisfaction with the process and responsibility (taking credit or blame for the outcome) for the process/outcome dependent upon whether or not the customer was a part of the production of the process. In examining previous literature, the authors make an interesting observation about co-production: “The customer is usurping the privileged status previously accorded to the producer.” Quoting Firat, Dhalakia, and

Behkatesch (1995, p. 50) they point out that the customer is demanding more of a role in production, “the consumer may be finding the potential to become a participant in the customization of his/her world.” One may see how these ideas would transfer nicely to the patient/physician relationship with respect to the increasing number of self-diagnoses given patients’ growing exposure to DTC pharmaceutical ads informing them of medical problems and treatments. THE STUDY The purpose of this study was to begin exploring the impact of DTCrespondents who had asked their physician for a prescription drug they had seen advertised. A final sample of 224 usable surveys resulted from the survey. (Table 1 presents a brief demographic profile of the study respondents.) Respondents were asked questions about how many times they had requested an advertised drug, whether or not their doctor had prescribed the requested medication, and several questions about their opinions of the office visit, their doctor, and their behavior following the visit. RESULTS Requests, Satisfaction, Perceived Effectiveness and

Compliance Over three-fourths of the respondents had asked their physician to prescribe a drug they had seen advertised “1 to 3 times” and over two-thirds indicated their requests had been granted. Table 2 provides relevant frequency responses and additional data presentation. The majority of the respondents indicated they were either “very satisfied” or “somewhat satisfied” with their office visit to the doctor. Almost two-thirds indicated they fully followed the doctor’s instructions following the visit, yet only 44.2% indicated they felt the visit was “very effective”. Over 86% of the respondents, however, said they would return to this doctor again. Opinions of Doctor’s Compliance Role and DTC Ads Regarding the respondents’ perceptions of the doctor’s role in responding to their drug prescription request, just over half agreed they would be upset if their doctor did not prescribe the drug requested. When asked whether or not they thought their doctor should prescribe a drug when

110

requested, just over one-third agreed that he/she should comply with the request. With respect to perception of DTC ads, over 40% indicated they are often confused by the ads (women were significantly more likely to indicate so) and over one-third felt the drug ads help inform them of their health care problems. Less than 27% indicated they feel the drug ads fully inform them of the drug’s side effects. Influence of Physician Prescription Decisions To test whether doctors’ responses to drug prescription requests significantly influence the patient-physician relationship, logistic regression analysis was used. Four models were estimated. In each, the independent variable of interest is whether the physician in question prescribed the medication the patient requested. Several additional predictor variables describing the patient were also included as controls in each model: patient gender, patient age, patient education level and the party paying for the majority of the patient’s prescriptions. These categorical predictor variables were coded as dummies with certain values set at the reference values, to which the remaining values were compared. For patient gender, the reference value is male. For patient age and education, the youngest age group and the lowest education level serve as references. For paying party, Medicare is the reference value. The four dependent variables of interest are: whether the patient was satisfied with the visit, whether the patient felt the visit was effective in treating their medical condition, whether the patient followed the physician’s instructions and whether the patient intends to return to the physician. The original survey responses for the first three of these four variables (satisfaction, perceived effectiveness and patient compliance) were recoded and dichotomized. Patients indicating they were very satisfied or somewhat satisfied were grouped together and compared to those who were indifferent, somewhat dissatisfied or very dissatisfied. Patients indicating the physician was very effective or somewhat effective were compared to those patients indicating the physician was not effective at all. Similarly, patients indicating they fully followed the physician’s instructions or followed most of the instructions were compared to those that indicated they didn’t follow the physician’s instructions very closely. Analysis was conducted using this recoded data since the purpose of this exploratory study was to determine the general impact of the physician’s

decision to grant drug requests on the patient outcomes described (i.e. did the physician’s decision influence whether a successful outcome was observed?). There was less interest in determining whether the physician’s decision resulted in a somewhat dissatisfied patient, for instance, as opposed to a very dissatisfied patient. This finer level of data was collected to be used for an additional research project that will compare the specific responses and relative sensitivity of different types of patient groups. The results of the analysis are displayed in Table 3. The overall explanatory power of the models, as measured by Nagelkerke’s R square, ranges from thirty-one to thirty-eight percent. The independent variables included in each of the models explain a significant amount of the variation in each of the outcomes of interest. This is evidenced by the significant improvement of the likelihood measure reported for each model which represents the probability of the observed results, given the parameter estimates. Since this value is typically a small number less than one, it is customarily reported as -2 times the log of the likelihood (Hair, Anderson, Tatum and Black, 1992). The closer this value is to zero, the better the model. In this study, the introduction of the independent variables into each of the four models beyond the constant results in a significant improvement in the likelihood (at the .01 level of significance for one model and .0001 level of significance for the remaining three). The resulting estimated influence of each of the predictors on the outcome variables are reported in terms of Exp(B). In Model I, these values represent the factor by which the odds a patient will be satisfied with a visit will change when the independent variable increases by one unit while all others are held constant. Note these odds are defined more specifically as the ratio of the probability that the patient will be satisfied with a visit to the probability the patient will not be satisfied (Norussis, 1994). Exp(B) values greater than one signify an increase in the odds a patient will be satisfied, while a value less than one signifies a decrease in these odds. Thus, the large positive Exp(B) value of 11.2 in Model I suggests whether a physician grants a prescription request is significantly and positively related to whether a patient will be satisfied with a visit. In fact, this value suggests patients are roughly eleven times more likely to be satisfied with a visit if a physician prescribes the requested

111

medication than if a physician does not grant the request. In the remaining three models, the influence of whether a physician grants the requested prescription is also significant and positive. When a physician prescribes a requested medication, patients are roughly eighteen times more likely to feel the visit was effective. They are also roughly eleven times more likely to follow the physician’s instructions and they are roughly thirteen times more likely to return to such a physician. The results of Model II also suggest female patients are significantly less likely to feel the visit was effective when compared to male patients. None of the other predictors are significant in any of the four models. While certainly there are likely to be other factors that influence the patient outcomes examined here, the relatively large, positive Exp(B) values corresponding to physician prescription decisions across the board provide strong preliminary evidence that whether a physician prescribes the medication a patient has requested is a critical determinant of whether a particular visit will lead to a beneficial relationship in the future for both parties. CONCLUSIONS AND IMPLICATIONS Most of the respondents in this study reported that their drug prescription requests were complied with by their doctor. It is interesting to note that these respondents requested a drug from their doctor even though they appear to not have strong beliefs regarding the information value and helpfulness of DTC drug ads. More interesting is the apparent influence the doctor’s compliance with their prescription request has on satisfaction, return visits, perception of the effectiveness of the office visit and, alarmingly, the patient’s compliance with the doctor’s health care instructions. While such relationships (satisfaction, return, perceived effectiveness) might be typical in service delivery relationships such as restaurants, hotels, etc… one might expect that the physician’s judgment would hold more credence value in a health care decision dealing with medical diagnosis and pharmaceutical prescriptions. Regarding the actual patient’s following the doctor’s orders, the possibility that the patient may not comply if the doctor did not prescribe the drug requested, prescribed another alternative drug, or even an alternative treatment could greatly

influence the individual patient’s health outcomes. Assuming a positive relationship between following orders/patient compliance and quality of healthcare, the patient’s health may decline, or not improve, possibly leading to additional health care treatment, and thus, costs. The potential accumulated effect of such costs, direct and indirect, on a national level could be astounding. One recent estimate is that patient noncompliance costs the United States economy approximately $100 billion each year. (Bonacossa 2003) Implications obviously relate to the role of the pharmaceutical industry in its design of information that is useful and ethical in DTC ads. In addition, possible public awareness campaign initiatives may need to be considered by objective sources educating the consumer on the requirements of drug ads as well as stressing the importance of the medical professional’s role in prescription decision making. Finally, an examination of the office visit and the information provided to the patient when he/she makes a drug request should be conducted. While the physician has been put in somewhat of a difficult situation by increasing DTC pharmaceutical ads, this will more than likely only increase in the future. The doctor’s role in providing explanation/education upon the patient’s request should probably be given careful attention in not only improving the doctor-patient relationship but also in improving the patient’s health outcomes. REFERENCES Allison-Ottey, S. (2002), “African-American Doctors Say DTC Ads Raise Disease Awareness, Bolster Doctor-Patient Ties,” DTC Perspectives, l (4), 43-45. Belkin, L. (2001), “Primetime Pushers,” Mother Jones, 26 (12), 30. Bendapudi, N. and Leone, R. P. (2003) “Psychological Implications of Customer Participation in Co-Production,” Journal of Marketing, 67 (1), 14-39. Bloom, M. (1999), “Direct-to-Consumer Advertising Provides Challenge to Managed Care,” Journal of Managed Care Pharmacy, 5 (2). Bonacossa, P. (2003), “Driving Patient Compliance,” Pharmiweb, Monday, February 3.

112

Brown, A. B. (2001), “The Direct-to-Consumers’ Advertising Dilemma,” Patient Care, 35 (6), 22. Burton, T. (2002) “Brand-Name Backlash: Companies counter Push to Costly Pills,” Wall Street Journal, Sunday, March 13, 239 (50), B1. Consumer Reports (2003) “Free Rein for Drug Ads?” February. Craig, R. P. (1998) “The Patient as a Partner in Prescribing Direct-to-Consumer Advertising,” Journal of Managed Care Pharmacy, 4 (1), pNA. DTC Perspectives (2003), “Factoids,” 2 (2), 10. Firat, A. F., Cholakia, N., and Venkatesh, A. (1995), “Marketing in a Postmodern World,” European Journal of Marketing, 29 (1), 40-56. Glasgow, C. , Schommer, J. C., Gupta, K., and Pierson, K. (2002), “Promotion of Prescription Drugs to Consumers: Case Study Results,” Journal of Managed Care Pharmacy, 8 (6), 512-518. Hair, Jr., J. F., Anderson, R. E., Tatham, R. L., and

Black, W. C. (1992). Multivariate Data Analysis with

Readings (3rd ed.). New York: Macmillan.

Lerer, L. (2002), “DTC Europe: Dispatches from Across the Pond,” DTC Perspectives, 1 (3), 28-30. Maguire, P. (1999), “How Direct-to-Consumer Advertising is Putting the Squeeze on Physicians,” ACP-ASIM Observer, American College of Physicians-American Society of Internal Medicine, March. Medawar, C. (2002), “The Politics of Direct-to-Consumer Promotion of Prescription Medicines,” DTCA Symposium Report, January 10, Brussels, 7-10. Norusis, M. J. (1994). SPSS Advanced StatisticsTM 6.1. Chicago: SPSS. Tacconi, L. (2003), “DTCA: The New Consumer Marketing Frontier,” DTC Perspectives, 2 (1), 18-22. Wechsler, J. (1999), “The DTC Dilemma: As Pharmaceutical Companies Promote Products Directly to Consumers, Patients are Demanding Medications That May Be Inappropriate or Overly Expensive,” Pharmaceutical Executive, 19 (11), 22,24,26-27. Youner, J. (2002), “News from the DTC National Conference and Exhibition,” DTC Perspectives, 1 (4), 10-13.

TABLE 1 DEMOGRAPHIC PROFILE OF RESPONDENTS

GENDER: Male 40.2% Female 59.8% AGE: 24 years or younger 30.8% 25-45 years 33.5% 46-60 years 27.7% 61-74 years 5.8% 75 years or more 2.2% EDUCATION: Some High School 3.6% High School 17.0% Some College 31.4% Bachelor’s Degree 36.3% Graduate Education 11.7%

113

TABLE 2 DATA SUMMARY

Did doctor prescribe requested drug? Yes 68.3% No 31.7% Satisfaction with office visit? Very Satisfied 28.6% Somewhat Satisfied 32.1% Indifferent 21.9% Somewhat Dissatisfied 13.4% Very Dissatisfied 4.0% Followed doctors instructions after visit? Followed Fully 61.7% Mostly followed 29.7% Did not follow 8.6% Perceived effectiveness of office visit? Very Effective 44.2% Somewhat Effective 42.0% Not Effective At All 13.8% Return to this doctor again? Yes 86.6% No 13.4% Strongly Strongly Agree Disagree 1 2 3 4 5 Upset with doctor if he did not prescribe requested drug. 17.0 33.5 32.1 10.7 6.7 Think doctor should prescribe drug when asked to. 13.8 23.7 37.1 18.8 6.7 I am often confused by drug advertisements. 12.1 28.1 40.2 13.8 5.8 Feel drug ads fully inform of drug’s side effects. 6.7 20.2 34.1 30.5 8.5 Drug ads help inform me of health care problems. 7.1 26.8 37.5 19.2 9.4

Table 3

Results of Logistic Regression Analysis for Patient Outcomes (Logistic regression reference values are: Request Granted = No, Gender = Male, Age = Less than 24, Education Level = Some High School, Paying Party = Medicare). Effective Follow Satisfied Visit Instructions Return Predictors Values Exp(B) Exp(B) Exp(B) Exp(B) Request Granted Yes 11.20** 18.91** 11.51** 12.72** Gender Female 1.15 .34* 1.89 .98 Age 25 - 45 years .53 .95 2.00 1.60 46 - 60 years .62 .80 4.43 .50 61 - 74 years .57 .55 1.23 .45 75 Plus .16 .18 1.67 6.4E08 Education Level High School or GED .29 4.09 2.86 .80 Some College .55 3.50 5.62 3.20 Bachelor’s Degree .48 2.69 7.99 3.84 Graduate Degree 2.55 10.19 10.96 3.08 Paying Party Medicaid .28 3.18 9.4E08 10.85 Out-of-pocket 1.55 .97 5.05 4.03 Self-insured .99 .63 2.21 .91 Employer-insured 2.15 .55 5.44 2.17 Constant .65 1.64 .08 .51 Nagelkerke’s R-square .38 .37 .32 .31 * = significant at .05 level, ** = significant at .0001 level

114

5.3.1 THE GOAL-DRIVEN SHOPPER: THE INFLUENCE OF GOAL SETTING AND GOAL

COMPLETION ON CONSUMER SPENDING

Leonard Lee, Massachusetts Institute of Technology Dan Ariely, Massachusetts Institute of Technology

ABSTRACT Conditional coupons are pervasively used as a promotional device in today’s retail marketplace. These coupons typically entitle customers to a discount or fixed monetary rebate if they spend at least a stipulated amount (e.g. “Spend $X or more and get $Y off). Given the popularity of these coupons, it is important to try to understand how they work. A straightforward answer stems from the economic incentives that such coupons create by decreasing the cost of buying additional (marginal) products. In addition to such incentives, we propose that these coupons, particularly with their precondition values, can provide concrete spending goals for consumers, changing their shopping agendas. To test this hypothesis, a series of five field experiments was conducted at a local convenience store. In each experiment, we randomly distributed different types of promotional coupons to customers as they entered the store, and analyzed the influence of these coupons on both the amounts and patterns of their subsequent spending. In Experiment 1, we demonstrated the basic goal effects by distributing three different types of coupons, all having a face value of $1. One type of coupon entitled customers to an instant rebate of $1 if they spent at least $5 (Conditional-$5 coupon), another had the same rebate value but with a precondition value of $10 (Conditional-$10 coupon), and the third type of coupon offered a rebate regardless of spending (Unconditional coupon). We found that customers who were given either type of conditional coupons spent significantly more than those who received either the unconditional coupons or no coupons at all (all p’s < 0.0001). Most central to the goal explanation, we found that customers who were given conditional coupons tend to spend amounts that clustered above the respective precondition values ($5/$10) on the coupons (which we shall call the goal-clustering effect), providing initial support for the goal hypothesis.

Experiment 1 suffered from one major problem –a large number of customers who did not qualify for the rebate did not return the coupons (despite being given lottery incentives to return them as they exited the store). This self-selection problem poses a serious challenge to the interpretation of the results from Experiment 1. In Experiment 2 (as well as later experiments), we corrected this bias by recording customer spending and disbursing the rebates directly to customers at the cashiers. The results from Experiment 2 replicated the results from Experiment 1 – there was a significant increase in average spending when customers were given conditional coupons (p < 0.0001), as well as a similar goal-clustering effect. Nonetheless, the effects described thus far could be explained by either the goal perspective or standard economics. From an economic perspective, conditional coupons that provide rebates have the effect of reducing the marginal cost of buying an additional item (assuming that the purchase of this item would allow the customer to meet the precondition on the coupon), and can thus induce more spending. To provide an initial test that contrasts the goal perspective and the economics perspective, we included a conditional coupon that had a very low requirement level in Experiment 3. From an economic perspective, setting a lower precondition value should only make the rebate easier to obtain, and should still lead to an increase in overall spending. On the other hand, the goal hypothesis would predict a decrease in spending, given that the lower precondition value would induce the corresponding setting of a lower goal. In addition, in order to test the generality of goals, we introduced a different type of goal – one that was dependent on the number of items purchased and not on the amount spent. As in the earlier experiments, the results showed a similar goal-clustering effect with this new type of goals. More notably, we also found a significant decrease in the number of items bought by customers given the low precondition (1-item) coupon (M = 2.11) relative to those not given any coupons at all (M = 2.69, p = 0.03). The next two experiments were designed and conducted to test more directly the goal

115

explanation. In Experiment 4, we distributed two types of coupons – a Conditional-$6 coupon and an Unconditional coupon. However, half the customers who entered the store were asked to estimate how much they thought they would spend in the store before receiving the coupons. Consistent with the goal hypothesis, this prior questioning significantly attenuated the degree to which customers used the coupon preconditions to set goals for their spending (p < 0.05 for interaction effect). We believe that the initial questioning had the effect of increasing customers’ psychological commitment to their own spending goals, hence making it less likely for them to use the coupon preconditions to set goals for their spending. A more implicit test of the goal explanation was carried out in Experiment 5 by manipulating the location at which the coupons were distributed. Instead of handing out all the coupons outside the store, we distributed half the coupons along the aisles at the back of the store and to customers who were holding at least one item in their hands (Presumably, at this stage, consumers had a more defined goal for their spending at the store). Interestingly, we found that customers who received the coupons inside the store were less influenced by the coupons – there was a significant interaction effect between coupon precondition and the location of coupon distribution on spending (p = 0.04). Arguably, customers who received coupons inside the store were more decided on what to buy and how much to spend, and were thus less influenced by the coupons. In sum, the results from this set of experiments showed that consumers have vaguely defined objectives when entering a store, and at that point, different persuasion efforts that allude to some goal can have a significant effect. More generally, these results provide additional support for the goal view that motivation and planned actions in day-to-day life are directed at the fulfillment of some end state. For further information contact: Leonard Lee, MIT Sloan School of Management (Marketing), 38 Memorial Drive, Room E56-345E, Cambridge, MA 02139, Email: leonardl@mit.edu

116

5.3.2 TIS BETTER TO GIVE THAN TO RECEIVE: PREFERENCE ESTIMATES CONDITIONED ON OWN AND OTHERS' PREFERENCES

Andrew D. Gershoff, Columbia University

Susan M. Broniarczyk, University of Virginia ABSTRACT Consumers are both givers and receivers of advice. Sometimes they are called on to help others by acting as agents who give recommendations (West 1996). Other times they must call on others for decision making assistance and they become the recipients of others’ agent services (Gershoff, Broniarczyk, West 2001). Since information is neither perfect nor costless, and the future is uncertain, the recommendations that they give and receive are not always accurate. Therefore, both when giving and receiving advice, consumers often must make estimates about what their own and others’ preferences will be (Loewenstien and Schkade 1999; West 1996). Frequently these estimates are conditional probabilities involving both one’s own and another’s preferences or evaluations (Gershoff et al. 2001). For example, for the task of giving a recommendation to a friend, a consumer might consider the conditional probability that the friend will like an alternative given that the consumer liked it; P(other +|self +). For the task of receiving a recommendation from a friend, the consumer might consider the conditional probability that he or she will like an alternative given that the friend liked it; P(self +|other +). While both tasks require consideration of both one’s own and other’s preferences, the conditional probabilities to be estimated differ in that one is the inverse of the other. Research has shown that that people’s ability to assess a conditional probability may depend on both the task (Gershoff et al. 2001) and information presentation (Gigerenzer and Hoffrage 1995). One error that has been observed is the occasional mistaking of P(A|B), with its inverse, P(B|A) (Bar-Hillel 1983). An explanation for this finding is that a category that is more accessible or perceived to be more diagnostic tends to be adopted as the denominator, or sample space, for the conditional probability estimate (Sherman, McMullen, Gavanski 1992; Brase Cosmides, and Tooby 1999). Depending on whether a consumer is giving or receiving a recommendation, the appropriate sample space for assessing the conditional

probability differs. For a consumer giving a recommendation it is the set of alternatives that the consumer has given a positive rating. For a consumer receiving a recommendation, it is the set of alternatives that the other individual has given a positive rating. Where a categorization schema for the relevant sample space is difficult to access, people may rely on a more accessible categorization schema or one that appears diagnostic because it is related to the task at hand and because it is relevant for similar tasks (Hilton and Fein 1989). Compared to information about another’s preferences, information about one’s own preferences is likely to be more accessible, and generally more relevant. Indeed, information that is self-relevant has been shown to be easier to recall and to be perceived as more important than information that is not self-relevant (Bower and Gilliagan 1979). Categorizing a set of alternatives in terms of one’s own likes and dislikes is likely to be a more natural way to categorize the alternative space than in terms of another’s likes and dislikes. Thus, it is predicted that people will be less accurate and more likely to make an inverse conditional probability error when assessing P(self+|other+) (assessing a received recommendation), compared to assessing P(other+|self+) (assessing a given recommendation). One hundred and forty-nine participants in study 1 both provided their own liking and disliking ratings and learned the ratings of another participant for a set of 50 posters. Next participants provided estimates of P(self +| other +) and P(other +| self +). The order of estimating the two probabilities was counterbalanced. As predicted participants performed worse and were more likely to make estimates consistent with the inverse conditional probability for estimates of P(self +| other +) compared to P(other +| self +). The correlation between participants’ estimates of P(self +| other +) and actual (r = .271; z = 1.96, p = .05) was significantly less than the correlation between participants’ estimates of P(other +| self +) (r = .467) and participants showed more error in predicting P(self +| other +) (M = .2215; F(1,145) = 7.002, p < .01) than P(other +| self +) (M = .1825).

117

Finally, estimates P(other +| self +) were more highly correlated with the inverse (r = .607) than with the actual conditional probability (r = .271; Z = 3.64, p <.001) while for estimates of P(self +| other +), there was no difference (Z = .011, p = .99). Study 2 sought to rule out the possibility that the error observed in study 1 for conditional probability estimates of one’s own preferences conditioned on another’s preferences occurs for all estimates of one’s own preferences, regardless of the conditioning information. Thus in study 2, one hundred and twenty four participants made estimates of the probability that they would like an alternative conditioned on an attribute of the alternative (the subject of the poster, or the catalog where it is sold) rather than conditioned on another’s preference. Because thinking about a set of alternatives in terms of commonly used descriptive attributes is likely to be a natural way to categorize the alternative space, in study 2 it was predicted that estimates would show less error and less consistency with the inverse conditional probability for estimates of one’s own preferences conditioned on attribute information, compared to estimates of attribute properties conditioned on one’s own preferences. As predicted, estimates of P(self +| attribute) were more highly correlated than with the actual probability (r = .632, p < .001) than estimates of P(attribute| self +) (r = .104, p = .250; z = 4.981, p < .001), and participants had less error in their estimates of P(self +| attribute) (M = .1338) compared to their estimates of P(attribute| self +) (M = .1617) (F(1,122) = 3.870, p = .051). Finally, analysis of the relationship between participants’ estimates and the inverse conditional probabilities suggests more use of inverse conditional probabilities for P(attribute |self +) than for P(self +|attribute). Specifically, estimates of P(attribute | self +) were more highly correlated with the inverse (r = .356) than the actual conditional probability (r = .104) (Z = 2.08, p < .05). Conversely, estimates of (self +|attribute) were more highly correlated with the actual (r = .632) than the inverse (r = .110; Z = 4.93, p <.001). The studies presented have found that consumers are prone to systematic errors in conditional probability estimates involving their own preferences consistent with differences in the accessibility and diagnosticity of the relevant sample space for the task. The probability estimation errors were consistent with better

accuracy associated with the task of giving compared to receiving recommendations (Study 1). However the errors associated with estimates of one’s own preference diminished when the conditioning information was attribute based (Study 2). References Bar-Hillel, Maya (1983), “The Base Rate Fallacy

Controversy,” in Decision Making Under Uncertainty, ed. Roland W. Scholz, Amsterdam: North-Holland, 39-61.

Brase, Gary L., Leda Cosmides, and John Tooby (1998), “Individuation, Counting, and Statistical Inference: The Role of Frequency and Whole-Object Representations in Judgment Under Uncertainty,” Journal of Experimental Psychology: General, 127 (1), 3-21.

Bower, G. H., & Gilligan, S.G. (1979). Remembering information about one's self. Journal of Research in Personality, 13, 420-432.

Gershoff, Andrew D., Susan M. Broniarczyk, and Patricia M. West (2001), “Recommendation or Evaluation? Task Sensitivity in Information Source Selection,” Journal of Consumer Research, 28 (December), 418-438.

Gigerenzer, Gerd and Ulrich Hoffrage (1995), “How to Improve Bayesian Reasoning without Instruction: Frequency Formats,” Psychological Review, 102 (October), 684-704.

Loewenstein, George and David Schkade (1999). "Wouldn’t it be Nice? Predicting Future Feelings," in E. Diener, N. Schwartz, & D. Kahneman (Eds.) Foundations of Hedonic Psychology: Scientific Perspectives on Enjoyment and Suffering. New York: Russell Sage Foundation Press.

Sherman, Steven J., Matthew N. McMullen, and Igor Gavanski (1992), “Natural Sample Spaces and the Inversion of Conditional Judgments,” Journal of Experimental Social Psychology, 28, 401-421.

West, Patricia M. (1996), “Predicting Preferences: An Examination of Agent Learning,” Journal of Consumer Research, 23 (June), 68-80.

For further information contact: Andrew D. Gershoff, Columbia Business School, 3022 Broadway, 520 Uris Hall, New York, NY 10027, Phone:(212) 854-0344, e-mail: ag644@columbia.edu

118

5.3.3 SHOOTING THE PRINCIPAL WHILE EMBRACING THE AGENT: HOW RESPONSIBILITY ACCEPTANCE INFLUENCES PRINCIPAL AND AGENT EVALUATIONS WHEN CONFRONTED

WITH BAD NEWS

Susan Hogan, Emory University

ABSTRACT This research suggests that individuals often form two separate evaluations for agents and principals when confronted with bad news. Specifically, while agents derive more favorable evaluations from clients when they accept responsibility for bad news, their employers’ evaluations are most favorable when the agent does not accept responsibility. INTRODUCTION When someone consciously decides that an injustice is too small to deserve a response, that someone still exacts a price from the perpetrator. Typically, the consequence is a slight change in the level of respect for, positive attitude toward, or commitment to the perpetrator of the injustice. These incremental responses are the consequence of a small, but nagging internal voice that demands attention (Sheppard, Lewicki, & Minton, 1992).

Responses to bad news can be quite

passive in nature, as is often the case for small transgressions versus large transgressions (Farrell, 1983; Rusbult, Zembrodt, & Gunn, 1982). A customer at a dry cleaner who receives a shirt back with two broken buttons may not bother to complain and may continue taking his clothes to this store. Simultaneously, however, he might unconsciously become less loyal to the dry cleaner and also find himself mentioning the problem to friends who potentially could have been future customers. Similarly, a retailer who receives product shipments late from a manufacturer may not stop distributing this manufacturer’s product. However, she may feel less positively toward the manufacturer’s representative, as well as the firm he or she represents. Over time, she may possibly allocate less shelf space for the firm’s products when the inventory is high and space is at a premium. Consequently, these passive responses have the important effect of slightly changing one’s attitude, which should not be ignored (Adams, 1965).

Fortunately, one problem usually will not

determine the fate of the business relationship.

People do understand that sometimes mistakes are made. However, as was found in the preliminary research conducted for this paper, people generally want to know the cause of a problem. This information can help them understand who to hold responsible as well as potentially how to avoid the same problem in the future. Often, however, especially in cases in which multiple parties are involved, the cause of the problem is unclear. Consequently, in these situations, who should be held accountable for the problem is also often unclear (Folkes, 1984). Thus, in these instances, how to assign initial responsibility for a problem is often left to the discretion of either the firm delivering the bad news (reporting the problem) or the individual agent within the firm, who has primary contact with the client.

The behaviors of these agents (e.g.,

salespeople and other customer-contact personnel) during encounters with customers, as well as the quality of the relationship between the customer and the agent, have been found to influence the impressions and opinions formed toward both the agent and the agent’s firm, the principal. Further, such agent behaviors also influence the probability of continued business between the customer and agent’s firm (Crosby, Evans, & Cowles, 1990; Normann 1983; Weitz, Sujan, & Sujan, 1986; Zemanek, 1977). Consequently, the decisions the firm or agent makes regarding whether to accept responsibility for the cause of a problem can have significant implications for both the agent and the firm.

CONCEPTUAL BACKGROUND AND HYPOTHESIS DEVELOPMENT Evaluations Toward Individuals Accepting or Denying Responsibility

Forsyth, Berger, and Mitchell (1981) explored how perceptions about group members formed by other members of the same group vary due to the amount of personal responsibility each group member took for the group’s negative outcomes. Group members who took low responsibility (blamed others) were seen as having self-serving motives and were less liked than group members who exhibited moderate or high personal

119

responsibility by either allocating responsibility equally or taking blame for the failure. This dislike for the “buck passer” supports Tetlock’s (1980) finding that teachers who were counter-defensive were perceived as more professionally competent, concerned for others, and likable than were defensive teachers. Thus, it appears that individuals are viewed most positively when they take high personal responsibility for negative outcomes and least highly when they take low responsibility or blame others. Given these findings, it seems logical to expect that negative information recipients will evaluate the agent most positively when the agent takes high personal responsibility for the bad news.

H1: Agents will receive the highest evaluations when they accept personal responsibility for the bad news.

The above hypothesis provides guidelines regarding the best course of action for an agent to take when deciding whether to accept personal responsibility for a problem at the client location, in order to be personally seen in the best light by the client. However, often these individual agents find themselves in situations in which they not only need to be concerned about how they themselves are viewed by others, but also how their behavior and actions influence the perceptions others form about their employer, the principal. Thus, a question that remains is whether the best course of action that agents should take to be personally viewed most highly by the client (e.g., taking personal responsibility) is the same optimal course of action they should take to ensure that their employer, the principal receives the most favorable client evaluation as well.

Understanding Evaluations toward the Principal

The existing literature suggests that, when

consumers perceive product success to be controlled by a firm (e.g., a difficult stain is removed due to a local dry cleaner’s expertise), consumers are more willing to compliment the firm and to recommend it than when the positive result is not due to that particular firm’s action (e.g., the dry cleaner used a standard formula to remove the stain) (Folkes, 1984). Conversely, respondents feel that a firm owes consumers both refunds and apologies when the product failure is identified as being seller-related, but not when the product failure is identified as consumer-related (Curren & Folkes, 1987; Folkes, 1988). That is, there is a tendency to match reasons that are similar to or

representative of the action to the action itself (Belk & Painter, 1983; Belk, Painter, & Semenik, 1981; Folkes 1988; Kahneman, Slovic, & Tversky, 1982). Consequently, I anticipate that the opinions and perceptions formed toward firm related issues will follow a pattern of results similar to that found in the perceived outcome controllability and firm blame attribution literature.

H2: The evaluations formed toward the principal, with regard to re-patronage intentions and perceived firm blame attribution will be the most negative when the agent accepts high personal responsibility for the bad news.

From the perspective of providing guidelines to an agent regarding the best course of action to take when deciding whether to accept personal responsibility for a problem at the client location, these two hypotheses offer contradictory advice. That is, if the agent chooses to take personal responsibility for a problem, in order to receive the highest personal evaluative responses from the client, this action will not have the same positive effect on firm evaluations. A question that remains, then, is which response will win out, in terms of the customers’ overall response—that formed toward the agent or that formed toward the firm? Understanding the Customer’s Overall Response

The social psychology literature contains

many theories and models suggesting that individuals often negotiate responses to new information by processing the information through concurrent, but qualitatively different, processes. Both the Elaboration Likelihood Model (ELM) and impression formation (inferences) literature contain examples of such dual processing perspectives (Chaiken & Trope, 1999; Uleman, 1999). For example, the ELM model suggests that individuals process personally relevant information via an analytic (central) route. Simultaneously, however, individuals might process less important information, such as attractiveness of the message source or quality of the advertisement, through a more emotional (peripheral) route (Petty & Cacioppo, 1983).

The organizational justice literature also

supports this idea, suggesting that individuals often form both emotional and rational responses to an injustice. Further, individuals have the ability to “negotiate” a response to the injustice, often between two objects, to achieve some sort of

120

compromise reaction or combined course of action (Sheppard, Lewicki, & Minton, 1992).

Based on this evidence, I predict that

individuals receiving bad news information from an agent will not necessarily form the same evaluations toward the principal and agent. Rather, individuals may form separate evaluations toward the agent and the firm, based on the same information.

H3: Individuals, when presented with bad news by an agent, will not always form the same evaluation toward the principal as they form toward the agent. STUDY 1

The purpose of Study 1 was to explore whether an agent’s acceptance or denial of personal responsibility for a problem in an organizational context influenced the client’s evaluations of both the agent and the principal. Additionally, in cases in which the agent did not accept personal responsibility, I was interested in whether and where the agent assigned responsibility influenced these evaluations.

Type of Bad News

To determine what were considered common day-to-day (non-critical) bad news problems in organizational environments, seven in-depth interviews were conducted with industry professionals from six different firms. The industries that employed these individuals included automotive, pharmaceutical, manufacturing, high-tech, banking, and broadcast media. Of the bad news situations mentioned, all seven respondents mentioned at least one bad news situation that had to do with a delivery date delay (e.g., computer hardware, software projects, parts components, or retail products).

Design and Procedures

This study was a 3 (responsibility) between subjects design. Respondents participating in this study were presented with a three-paragraph scenario asking them to adopt the role of an information technology manager (information recipient), whose company had hired a third party firm for a one-year contract to perform on-site software customization and installation. They were then informed by an agent of this third party firm that a current project was going to be delayed. In

all conditions, the agent also provided a new projected project completion date. Respondents were then asked to answer a series of 7-point semantic scale questions regarding their perceptions and feelings toward the agent and principal.

Responsibility Manipulation

Responsibility assignment for the bad news was manipulated by having the agent accept personal responsibility, assign responsibility to a fellow co-worker, or assign responsibility to an employee of the information recipient’s company.

Responsibility was assigned to individuals, as opposed to organizations or non-human entities, based on information gathered from the exploratory interviews as well as an industry pretest. Specifically, the majority of the industry professionals interviewed believed that, when in the position of delivering bad news, they would be further queried for a name within the company, rather than just getting away with naming a department or another part of the business. Similarly, when in the position of receiving bad news, most would press the agent delivering the bad news for an individual that they could hold accountable (responsible). Thus, assigning responsibility to an employee within the client’s firm, as well as a fellow co-worker, was meant to represent acceptable and common places where the agent could deny responsibility for the problem, but yet blame an individual rather than an institution. Agent and Firm Dependent Measure Questions

Respondents rated the agent delivering the bad news on five personality attributes: honesty, trustworthiness, credibility, competency, and professionalism. These five personality attributes were identified in a pretest as the most important qualities respondents would like to see in an agent with whom they have a close working relationship. The idea of measuring customer evaluations of agent attributes is consistent both with research on various dimensions of overall customer satisfaction (Groonroos, 1988; Parasuraman, Zeithaml, & Berry, 1988) and with research on the relationship between customer satisfaction and repurchase intentions. Specifically, evaluations of agent attributes have been identified as the questions most predictive of client satisfaction and repurchase intentions (Mittal and Kamakura, 2001; Mittal, Ross, & Baldasare, 1998). These five agent attributes were combined and averaged into a

121

dependent measure: agent evaluations (Cronbach’s Alpha = 0.92).

Respondents also were asked three

questions related to re-patronage intentions: likelihood of doing future business, contract renewal, and switching vendors. Additionally, they were asked one question regarding whether they blamed the agent’s firm for the bad news. This last question was asked to determine the extent to which respondents held the firm responsible for the bad news and whether this firm blame attribution was influenced by whether the agent accepted responsibility for the problem. The three re-patronage intention questions, as well as the question regarding blame attributed toward the firm, were combined to create a new dependent measure: firm evaluations (Cronbach Alpha’s = 0.75).

STUDY 1 RESULTS

Eighty-five respondents completed this study as part of a business course requirement. The average reported age of the respondents was 19.4 years old; forty-three were females, forty-two were males. As expected, the five agent attributes, combined to create an agent evaluation dependent measure, all loaded onto one factor of a factor analysis. Additionally, the three re-patronage intention questions, as well as the question regarding blame attributed toward the firm, all loaded onto the second factor and were combined to create a firm evaluation dependent measure.

Agent Evaluations

The significant main effect for responsibility, F(2, 82) = 6.10, p < 0.0034, suggests that evaluations vary based on where the agent assigns responsibility for the problem when delivering bad news. With regard to agent evaluations, (Figure 1), the agent was evaluated most positively when the agents took personal responsibility for the bad news they were delivering (M takes responsibility = 1.65) as opposed to passing responsibility to either a fellow co-worker (M passes responsibility to fellow co-worker = 1.08) or to an employee at the client’s location (M passes responsibility to client’s employee = 0.61). This result provides support for H1.

Firm Evaluations

However, as anticipated, and can be seen

in Figure 1, the firm evaluations were the least

favorable when the agents took personal responsibility for the problem (M takes responsibility = -0.16) versus passing responsibility to a fellow co-worker (M passes to fellow co-worker = 0.00) or passes responsibility to an employee at the client’s location (M passes to client’s employee = 0.53). This significant main effect for firm evaluations, F (2,82) = 4.97, p<.0092, provides support for H2. Moreover, firm evaluations were most favorable when agents passed responsibility for the problem to an individual outside their own firm (in this case, within the client’s firm) as opposed to merely passing responsibility to another individual within their own firm. The contrast effects are given in Table 1.

These results suggest that there is a potential principal-agent conflict, in that the agent has an incentive to take responsibility for a problem when the cause of the problem is hard to identify. In contrast, the firm has an incentive to encourage the agent to pass responsibility for the problem.

DISCUSSION

Contributions and Implications

The results presented in this paper suggest that the individuals within client organizations, upon receiving bad news from an agent with whom they have a business relationship, often form two separate evaluations: one toward the agent and another toward the agent’s employer, the principal. Moreover, these evaluations formed toward agents and principals by organizational clients are influenced by whether the agent accepts personal responsibility for a problem.

Identification of a Principal-Agent Conflict. There is strong evidence to suggest that agents are evaluated most highly when they take personal responsibility for a negative event. Thus, given situations in which the agent could potentially blame another party, it may make sense for an agent to take responsibility when faced with the decision of whether to accept responsibility for a problem as a means to be seen in a favorable light by a client. Thus, this choice to take responsibility is not only the morally kind, or “big” thing to do, but also the smart thing to do. However, this action does not bode well for the agent’s employer, the principal. Rather, the best thing agents can do for the sake of their employer is to not accept responsibility, but rather, as the situation explored in this study suggests, pass responsibility to an

122

employee within the client’s firm. Thus, these results suggest that there is a potential principal-agent conflict of interest due to the fact that situations may arise in which agents are put in a position in which they have to choose between maximizing their own personal goals versus the goals of their employer. Limitations and Future Research Replicating Study on Industry Professionals. In the study presented, the participating respondents were similar to those who participated in the prior interpersonal studies I referenced and built on to develop my hypotheses. Using this similar type of respondent population (college students) enabled me to determine whether findings from prior interpersonal research held when the context was changed to a business (inter-organizational) setting while minimizing the possibility for alternative explanations. What remains to be seen however, is whether business professionals evaluate agents and firms in a manner similar to how the current respondents also evaluate agents and firms. One might argue that, whereas college students, acting as respondents, form separate evaluations toward the firm and agent, business professionals may not,

having over time learned to integrate their feelings toward the agent and the firm, such that there is little distinction in how they view the two. However, if this is the case, it is still important to determine in which direction these evaluations have moved. That is, are business professionals’ evaluations more closely aligned with how the current respondents evaluated the firm or the agent? Thus, a worthwhile area for future research would be to replicate the current studies on business professionals. Which Evaluations Matter Most and Why? The study presented adds support to the existing body of literature suggesting that people process information and make evaluations through more than one cognitive process. What has not been addressed is which type of client evaluation (e.g., agent attribute evaluations, firm repurchase intentions, or perceived blame attributions) is the most indicative of a client’s true overall satisfaction and actual repurchase intentions. Similarly, what has not been explored in detail is the process that might explain the observed results. That is, what is it about the agent accepting or assigning responsibility that causes these separate evaluations?

Table 1

Study 1 Responsibility Assignment Contrast Effects __________________________________________________________________________

Agent Evaluations Firm Evaluations

Contrast DF F Value Pr > F F Value Pr > F Pass Responsibility vs. Take

Responsibility 1 8.96 0.004 4.35 0.04

Take vs. Pass to Co-worker 1 2.91 0.09 0.41 0.53 Pass to Co-worker vs. Pass to

Client’s Employee 1 1.95 0.17 4.31 0.04

Take vs. Pass to Client’s Employee 1 12.16 0.0008 9.29 0.0031 Within Agent’s firm vs. within

Client’s firm 1 7.84 0.006 8.87 0.004

123

Figure 1

Study 1: Responsibility’s Effect on Evaluations

1.65

1.080.61

-0.16 00.53

-3-2.5

-2-1.5

-1-0.5

00.5

11.5

22.5

3

TakeResponsibility

FellowCoworker

Client'sEmployee

Neg

ativ

e --

Pos

itive

Agent EvaluationsFirm Evaluations

REFERENCES

Adams, J. Stacy (1965). Inequity in Social Exchange. In Leonard Berkowitz (Ed.), Advances in Experimental Social Psychology (Vol. 2, pp. 267-99). New York: Academic Press. Belk, Russell, & Painter, John (1983). Effects of Causal Attributions on Pollution and Litter Control Attitudes. In Non-Profit Marketing: Conceptual and Empirical Research. eds. F. Kelly Shuptrine and Peter Reingen. Bureau of Business and Economics Research. Arizona State University, Tempe, AZ 85287, 22-25. Belk, Russell, Painter, John, & Semenik, Richard. (1981). Preferred Solutions to the Energy Crisis as a Function of Causal Attributions. Journal of Consumer Research, 8, 306-312. Chaiken, Shelly, & Trope,Yaacov. (1999). Dual Process Theories in Social Psychology. New York: Guilford Press. Crosby, Lawrence A., Evans, Kenneth R., & Cowles, Deborah. (1990). Relationship Quality in Services Selling: An Interpersonal Influence Perspective. Journal of Marketing, 54, 68 – 81.

Curren, Mary T., & Folkes, Valerie S. (1987). Attributional Influences on Consumers’ Desires to Communicate About Products. Psychology and Marketing, 4, 31-45. Farrell, Dan.(1983). Exit, Voice, Loyalty, and Neglect as Responses to Job Dissatisfaction: A Multidimensional Scaling Study. Academy of Management Journal, 26, 596-607. Folkes, Valerie S. (1984). Consumer Reactions to Product Failure: An Attributional Approach. Journal of Consumer Research, 10, 398-409. Folkes, Valerie S. (1988). Recent Attribution Research in Consumer Behavior: A Review and New Directions. Journal of Consumer Research, 14, 548-565. Forsyth, Donelson, R., Berger, Rick E., & Mitchell, Tom. (1981). The Effects of Self-Serving vs. Other-Serving Claims of Responsibility on Attraction and Attribution in Groups. Social Psychology Quarterly, 44, 59-64. Groonroos, Christian. (1983). Strategic Management and Marketing in the Services Sector

124

(Report No. 83 – 104). Cambridge, MA: Marketing Science Institute. Kahneman, Daniel, Slovic, Paul, & Tversky, Amos. (1982). Judgment Under Uncertainty: Heuristics and Biases. Cambridge: Cambridge University Press. Mittal, Vikas, & Kamakura, Wagner A. (2001). Satisfaction, Repurchase Intent, and Repurchase Behavior: Investigating the Moderating Effect of Customer Characteristics. Journal of Marketing Research, 38, 131-42. Mittal, Vikas, William T. Ross, & Baldasare, Patrick M. (1998). The Asymmetric Impact of Negative and Positive Attribute-Level Performance on Overall Satisfaction and Repurchase Intentions. Journal of Marketing, 62, 33-47. Normann, Richard. (1983). Service Management, New York: John Wiley & Sons, Inc. Parasuraman, A., Valerie Zeithaml and Leonard L. Berry. (1985). A Conceptual Model of Service Quality and Its Implications for Future Research. Journal of Marketing, 49, 41-50. Petty, Richard E., & Cacioppo, John T. (1983). Central and Peripheral Routes to Advertising Effectiveness; The Moderating Role of Involvement. Journal of Consumer Research, 10, 135–46. Rusbult, Caryl E., Zembrodt, Isabella M., & Gunn, Lawanna K. (1982). Exit, Voice, Loyalty, and Neglect: Responses to Dissatisfaction in Romantic Involvements. Journal of Personality and Social Psychology, 43, 1230–42. Sheppard, Blair H., Lewicki, Roy J. & Minton, John W. (1992). Organizational Justice: The Search for Fairness in the Workplace, (Series title: Issues in organization and management series). New York: Lexington Books. Tetlock, Philip E. (1980). Explaining Teacher Explanations of Pupil Performance: A Self-Presentation Interpretation. Social Psychology Quarterly, 43, 283-90. Uleman, James S. (1999). Spontaneous versus Intentional Inferences in Impression Formation. In Dual Process Theories in Social Psychology, ed. Shelly Chaiken and Yaacov Trope, New York: Guilford Press.

Weitz, Barton A., Sujan, Harish, & Sujan, Mita. (1986). Knowledge, Motivation, and Adaptive Behavior: A Framework for Improving Selling Effectiveness. Journal of Marketing, 50, 174-91. Zemanek, James E. Jr. (1997). Manufacturer Influence versus Manufacturer Salesperson Influence over the Industrial Distributor. Industrial Marketing Management, 26, 59-66.

125

6.1.1 PREFERENCE FLUENCY AND ITS EFFECT ON NO-CHOICE, COMPROMISE, AND ATTRACTION EFFECTS

Nathan Novemsky, Yale University

Ravi Dhar, Yale University Itamar Simonson, Stanford University

Norbert Schwarz, University of Michigan ABSTRACT The notion that preferences are constructed rather than retrieved from memory has been an important influence on consumer choice research (e.g. Bettman, Luce and Payne, 1998). While constructionist ideas were flourishing in choice research, researchers studying attitudes were suggesting that attitudes are also constructed rather than retrieved from memory. In fact, there are many parallels between the two bodies of research. Context effects based on temporary and often normatively irrelevant cues were having effects both on the construction of preferences and on the construction of attitudes. However, many ideas that might be applied to both attitudes and preferences have not crossed the boundary between these two fields. In the research described below, we attempt to introduce a key idea from attitude research into preference contexts. In the attitude literature, a line of research known as feelings-as-information (Schwarz, 1990; 2001) shows that individuals use their current subjective feeling state to inform their attitudes toward various objects. In our research we attempt to induce feelings (e.g. difficulty) in individuals making decisions and see how that affects their preference construction. We examine the role of feelings in three preference effects: the compromise effect, the attraction effect, and the no–choice effect. In several studies, we investigate the compromise effect. Demonstrations of this effect involve adding a third option to a two-option choice set in such a way as to increase the share of one the original options. For example, consider a choice between a high quality/high price option (B) and a low quality/low price option (C). When a very high price/very high quality option is added to the choice set, consumers seem to often choose the middle or compromise option (B) even more than they chose it in the two-option set. In our studies, we expand this design to include an inducement of a subjective feeling of difficulty. We manipulate the subjective difficulty of the choice by asking respondents if they can think of 2 or 10 reasons for their choice.

Most people cannot think of 10 reasons in the choices we offer, so the 10-reason condition induces a strong subjective feeling of difficulty. When subjective difficulty is increased, we observe a greater compromise effect. We also conducted several studies looking at the effect of adding a no-choice option to a decision. Similar to the compromise effect, we find a greater preference for the no-choice option when the subjective difficulty is high. In one study, we induce a feeling of difficulty by describing the options in a hard to read font and this font increases the compromise effect. Interestingly, subjective difficulty does not always increase context effects in choice. We investigated another preference effect, the attraction effect (Huber, Payne, & Puto, 1982). In the attraction effect, adding a third option to a two-option choice set that is dominated by one of the original options increases the share of the dominating option relative to the two-option set. Using the 2 versus 10 reasons manipulation we varied the perceived difficulty of choices related to the attraction effect. We find that the attraction effect, as opposed to the compromise effect or no-choice effect, is attenuated when the subjective difficulty of the choice is increased. We believe that the attraction effect is reduced because it is a perceptual effect. That is, the attraction effect occurs because the dominated option makes one option more attractive. Choosers are generally not aware that the dominated option is having an effect on the choice. Rather, the chooser just perceives the dominating option as more attractive. Increasing the subjective difficulty interrupts this effect by encouraging the chooser to think more about the choice. By contrast, the compromise and no-choice effects may be the result of more deliberate strategies, whereby the chooser feels the difficulty in making the choice and deliberately chooses the middle option or the no-choice option. In the studies above, the difficulty in generating reasons for the choice is attributed to the difficulty of the decision-maker in making the decision. In

126

follow-up studies, we manipulated whether the difficulty of generating reasons was attributed to the decision-maker having difficulty with the decision or to something else. For example, by telling respondents that other people on average can generate only a few reasons, the difficulty in generating reasons is attributed to the fact that too many reasons were requested not to the fact that the respondent is having difficulty with the decision. When the attribution for the feeling of subjective difficulty is directed away form the decision the influence of the number of reasons on the context effects is eliminated. Similarly, when attention is drawn to the difficult font in the as the source of difficulty in the no-choice study, the font has no effect on the share of no-choice. We have found that the subjective experience of the choice can have a substantial impact the option chosen. These studies also highlight that some context effects in choice may operate through very different mechanisms. We hope that more ideas cross the attitude/choice boundary in future research. There are many parallels between these two areas that have not been explored. For further information contact: Nathan Novemsky, Yale School of Management, 135 Prospect Street, New Haven, CT 06520, email: nathan.novemsky@yale.edu

127

6.1.2 THE PSYCHOLOGY OF MATCHING AND THE PROMINENCE EFFECT

Shane Frederick, Massachusetts Institute of Technology

Eldar Shafir, Princeton University

ABSTRACT Discussions about the relative importance of different variables arise often in academic and everyday discourse. Which is the more important determinant of SAT scores: nature or nurture? What is the more effective way to lose weight: diet or exercise? What do consumers care more about in a yogurt, taste or texture? However, despite widespread references to attribute importance, there is no consensus about what it means for one variable (attribute, dimension, etc.) to be more “important” than another. Attribute importance is often conceptualized as the degree to which an overall evaluation (a rating, ranking, choice, willingness to pay, etc.) is affected by a change in attribute level. However, relative importance comparisons are bedeviled by the problem of decomposing the impact of “range” and “weight.” For example, if Car A ($15,000, 26 mpg) is preferred to Car B ($30,000, 27 mpg), it could be because price varies “more” than gas mileage, not because price is the more “important” attribute. Despite this ineradicable ambiguity, many notions of attribute importance (“swing weights,” “part worths” etc.) are based on such evaluations. (For further discussion of this conceptual hurdle, see Shanteau, 1980; Keeney & Raiffa, 1976, pp. 271-273; Goldstein & Beattie, 1990, pp. 114-115.). However, several conceptions of attribute importance may exist independently from specific tradeoffs of particular alternatives. For example, most will confidently assert that saving lives is more important than saving money, without any reference to a specific number of lives or amount of money. Thus, one attribute may be considered more important than another if it evokes a stronger or more favorable emotional reaction, or is more perceptually salient, or is believed to be correlated with other valued attributes, or is believed to vary more across the alternatives typically considered. The welter of alternate theoretical conceptions and empirical measures of relative importance is well documented (see Goldstein 1990; Goldstein & Beattie, 1991; Kruskal 1984; Kruskal & Majors, 1989; Budescu, 1993; Goldstein, Barlas, & Beattie, 2001), as is the modesty of convergence across

such measures (see, e.g., Jaccard, Brinberg, & Ackerman, 1986; Borchedering, Eppel, & Von Winterfeldt, 1991; Aschenbrenner, Bockenholt, Albert, and Schmalhofer, 1986). These facts complicate an interpretation of the prominence effect, which is empirically verifiable (or refutable) only by reference to a measure of attribute importance that exists apart from the observed choice-matching discrepancy it is invoked to explain. If the relative importance of attributes is inferred from the choice matching discrepancy itself (as is sometimes done), the “prominence hypothesis” is reduced to a tautology -- “predicting” that the attribute that is weighted more heavily in choice than matching is weighted more heavily in choice than in matching. But if some other measure of attribute importance must be used, which of the various measures should it be, and what meaning can be given to the prominence effect if different measures of relative importance disagree? In a series of studies, we found that judgments of relative importance had one of six different bases, including: (1) preference between extreme exemplars; (2) strength of semantic/conceptual association with target domain; (3) strength of diagnostic or causal relations with other valued attributes; (4) magnitude of ecological variation; (5) moral priority, and (6) logical entailment. Although the ranking of attribute “importance” may agree across these different senses, the possibility of divergence calls into question the conceptual validity of the prominence effect, and, at a minimum, raises the issue of which of these distinguishable senses of importance correlates most strongly with choice-matching discrepancies or other phenomena attributed to the concept of “prominence.” For further information contact: Shane Frederick, Sloan School of Management, Massachusetts Institute of Technology, Room E56-317, 38 Memorial Drive, Cambridge, MA 02142-1347, email: shanefre@mit.edu

128

6.1.3 THE DISTINCTION EFFECT: MIS-PREDICTION AND MIS-CHOICE DUE TO JOINT-

SEPARATE-EVALUATION INCONSISTENCY

Christopher K. Hsee, University of Chicago Jiao Zhang, University of Chicago

ABSTRACT Suppose that a person, who shops for a digital camera, compares two models. One fits in his pocket and the other does not. But the one that fits in his pocket features only a 3-megapixel resolution, and the one that does not fit in his pocket features a 5-megapixel resolution. To decide which camera to purchase, the person predicts the difference in his consumption experience between having a camera that fits in his pocket and one that does not, and the difference in his consumption experience between having a 3-megapixel camera and a 5-megapixel camera. Is he able to make these predictions accurately? Is he able to choose the camera that will give him the greatest overall satisfaction? More generally, if people are presented with two alternative values of an attribute, are they able to accurately predict the difference these values will make to their happiness? If people are faced with two options, one having a better value on one attribute and the other having a better value on another attribute, are they able to choose the option that will bring them the greatest overall happiness? This research explores a potentially important source for failure to accurately predict future experience or failure to make experientially optimal choices. When people predict the experiential consequences of alternative values or choose between alternative options, they are in comparative, or joint evaluation (JE) of these values or options. But when people actually experience a product, they are typically in single evaluation (SE) of that product alone. Building on previous research on evaluability (Hsee, 1996; Hsee et al., 1999), we propose that the "utility function" of most attributes differs systematically between JE and SE. Generally speaking, the utility function of an attribute will be flatter in SE than that in JE except around a reference point. In other words, in JE, people will be sensitive to any variations on the attribute, but in SE, people will be less sensitive to variations on the attribute, except that the variation involves a switch

of valence, namely, from one side of a reference point to the other. When people in JE make predictions or choices for products to be experienced in SE, they do not (sufficiently) adjust for the difference between JE and SE, and tend to resort to their JE preferences to make their predictions and their choices. As a result, they may overpredict the experiential difference between alternative values and make experientially suboptimal choices. Our theory specifies when people in JE are likely to make overpredictions and when they are not. Generally speaking, people are likely to overpredict the experiential difference between attribute values that differ only in degree or only quantitatively, but are less likely to overpredict (and sometimes even underpredict) the experiential difference between attribute values that differ in valence or qualitatively. We refer to the overprediction of the experiential impact of a quantitative difference as the “distinction effect.” In the camera case we introduced earlier, our theory suggests that buyers are likely to overpredict the difference in their future consumption experience between having a 3-megapixel camera and having a 5-megapixel camera (merely a quantitative difference), but are unlikely to overpredict the difference between having a camera that fits in their pocket and one that does not (a qualitative difference). Likewise, people are likely to overpredict the difference in subjective well-being between earning $80,000 a year and earning $70,000 a year, but less likely to overpredict the difference between having friendly colleagues and unfriendly colleagues. This research yields contributions in at least two areas. First, it contributes to the basic utility function literature by showing that the utility function for decision and prediction can differ systematically from the utility function for experience. Second, it contributes to the affective forecasting literature by providing a novel explanation for affective overprediction and also specifying when people overpredict and when they do not.

129

For further information contact: Christopher K. Hsee, The University of Chicago Graduate School of Business, 1101 East 58th Street, Chicago, IL 60637, email: christopher.hsee gsb.uchicago.edu

130

6.2.1 IMPLICIT ATTITUDES TOWARD GREEN CONSUMER BEHAVIOR

Delphine Vantomme, Ghent University Maggie Geuens, Ghent University Jan De Houwer, Ghent University

Patrick De Pelsmacker, Ghent University

ABSTRACT

Traditionally, attitudes toward green consumer behavior are measured by means of explicit attitude measures (surveys). In general, those measures indicate growing environmental concern and willingness to change consumption patterns (Yam-Tang & Chan, 1998). However, when it comes to purchasing and consuming products and services, buying behavior is often inconsistent with these attitudes. In fact, the market share of the majority of environmentally friendly low-involvement products amounts to less than 1% (Roozen, 1999). Thus, these findings suggest that explicit measures may not be appropriate for research on green consumer behavior. Especially social desirability concerns are assumed to bias explicit attitude measurement as buying environmentally friendly products is an ethical consumer act (La Troobe, 2000; Roozen & De Pelsmacker, 1998). Further, it has been shown that respondents are not always able to report their attitudes and convictions accurately because these attitudes are not analyzed in a conscious and deliberate manner (Greenwald & Banaji, 1995). Recently, interesting insights on alternative methods came from psychological research on the measurement of automatic evaluations (Banaji, 2001; Fazio, Sanbonmatsu, Powell, & Kardes, 1986). This type of research is based on a distinction between automatic or implicit attitudes on the one hand and explicit attitudes on the other (Fazio, 1990; Wilson, Lindsey, & Schooler, 2000). Explicit attitudes are attitudes that operate in a conscious mode and are typically measured by self-report tasks. Implicit attitudes are “introspectively unidentified (or inaccurately identified) traces of past experience that mediate favorable or unfavorable feeling, thought, or action toward social objects” (Greenwald & Banaji, 1995). In general, it has been suggested that implicit attitudes guide spontaneous behavior whereas more deliberative behavior is influenced more by explicit attitudes (Wilson et al., 2000).

Implicit attitudes are measured by indirect measures that use reaction time as an indicator of automatically activated attitudes. Probably the most well known implicit measurement technique is the

Implicit Association Test (IAT), developed by Greenwald, McGhee and Schwartz (1998). The IAT is a computerized task that measures the strength of association between two contrasted target concepts (e.g. green and traditional products) on the one hand and an attribute dimension (e.g. positive and negative words) on the other hand. The idea behind the IAT is that it should be easier to map two concepts onto a single response when those concepts are somehow similar or associated in memory than when the concepts are unrelated or dissimilar. Because respondents cannot control the influence of attitudes on their reaction speed, the advantage of this measurement method is that it circumvents reliance on the willingness or ability of respondents to express their opinions.

1. Objectives and hypotheses The purpose of the current research was to

examine the usefulness of an implicit attitude measure in green consumer behavior. Therefore, we compared implicit attitudes as measured by the IAT with explicit attitudes and behavioral intentions toward cleaning products. Because cleaning products are low-involvement products that do not involve long effortful considerations on which brand to choose, we expected implicit attitudes to have an important impact on purchase behavior. It was hypothesized that explicit measures would reveal strong positive attitudes toward environmentally friendly cleaning products (due to measurement bias); while the implicit measure would show less positive or even negative attitudes. Further, because both implicit and explicit processes may operate in the purchase decision of low-involvement products (Petty & Cacioppo 1986; Fazio, 1990), we expected that implicit and explicit attitudes would relate to behavioral intention.

2. Methodology Participants. 60 undergraduate students (26

women, 34 men, Mage= 21.53) at Ghent University participated in the experiment in exchange for a movie ticket.

Learning phase. Upon arrival at the laboratory respondents were taught two fictitious brands of all-purpose cleaners (2 non-words, Matu and Giko) together with their characteristics in a learning phase. For half of the respondents, Matu referred to an environmentally friendly cleaning product and

131

Giko to a “traditional” cleaning product. For the other respondents the reverse was true. Respondents were instructed to memorize the brand names and their accompanying characteristics.

Measures. Implicit attitudes toward the fictitious brands were measured by means of an IAT. The explicit measures consisted of (1) explicit measures of attitudes and behavioral intentions toward the two fictitious brands (seven-point scales) and (2) behavioral intention measure toward real cleaning products.

3. Results and conclusions In line with previous research, the explicit attitude measure showed strong positive attitudes toward the green product. Contrary to expectations, the implicit attitude measure (IAT) revealed more positive implicit attitudes toward the green product than toward the traditional product. Our results thus provide no support for the hypothesis that implicit attitudes toward green brands are less positive or negative. This implies that we did not find support for the assumption that positive explicit attitudes result from drawbacks of explicit measurement such as social desirability bias or ‘rationalization’ of introspectively inaccessible attitudes. Instead, the results suggest that the weak attitude-behavior relationships often found in green consumer behavior research can be ascribed to intrinsic features of environmentally friendly products such as their discordant character rather than to measurement problems of explicit attitude measures.

With respect to behavior, the results showed that interindividual differences in implicit attitudes were significantly related to interindividual differences in purchase intentions of both fictitious and real green brands. Those participants who said that they would purchase the green product had a more positive implicit attitude toward the green brand (as compared to the traditional brand) than participants who said that they would buy the traditional product. What is even more interesting is that implicit attitudes were related to purchase intentions even when explicit attitudes were not. That is, purchase intentions with regard to real products were related to implicit but not explicit attitudes toward the fictitious green and traditional brands. These findings strongly suggest that measuring implicit attitudes could allow one to predict interindividual differences in green consumer behavior better than measuring explicit attitudes. REFERENCES Banaji, M.R. (2001). Implicit attitudes can be

measured. In H.L. Roediger & J.S. Nairne (eds.),

The Nature of Remembering: Essays in Honor of Robert G. Crowder (pp. 117-150). Washington, DC: Am. Psychol. Assoc.

Fazio, R.H. (1990). A Practical Guide to the Use of Response Latency in Social Psychological Research. In C. Hendrick & M.S. Clark (Eds.), Research Methods in Personality and Social Psychology (pp.74-97). California: Sage Publications, Inc.

Fazio, R.H., Sanbonmatsu, D.M., Powell M.C., & Kardes, F.R. (1986). On the Automatic Activation of Attitudes. Journal of Personality and Social Psychology, 50, 229-238.

Greenwald, A.G. & Banaji M.R. (1995). Implicit Social Cognition: Attitudes, Self-Esteem, and Stereotypes. Psychological Review, 102, 4-27.

Greenwald, A.G., McGhee D., & Schwartz J.L.K. (1998). Measuring Individual Differences in Implicit Cognition: the Implicit Association Test. Journal of Personality and Social Psychology, 6, 1464-1480.

La Troobe, H.L. & Acott, T.G. (2000). A Modified NEP/DSP Environmenal Attitudes Scale. Journal of Environmental Education, 32, 12-21.

Petty, R.E., & Cacioppo, J.T. (1986). The Elaboration Likelihood Model of persuasion. Advances in Experimental and Social Psychology, 19, 123-205.

Roozen, I.T.M. (1999). Milieu en Consument, Determinanten van Milieuvriendelijk Gedrag van Consumenten (Environment and Consumer, Determinants of Environmentally Friendly Consumer Behavior). Unpublished dissertation, Marketing Department, University of Antwerp: Antwerp, Belgium.

Roozen, I.T.M. & De Pelsmacker, P. (1998). Attributes of Environmentally Friendly Consumer Behavior. Journal of International Consumer Marketing, 10, 21-41.

Wilson, T.D., Lindsey, S. & Schooler, T.Y. (2000). A Model of Dual Attitudes. Psychological Review, 107, 101-126.

Yam-Tang, E. P.Y. & Chan, R.Y.K. (1998). Purchasing Behaviours and Perceptions of Environmentally Harmful Products. Marketing Intelligence and Planning, 16, 356-362.

For further information contact: Delphine Vantomme, Ghent University, Faculty of Economics and Business Administration, Department of, Marketing, Hoveniersberg 24, B-9000 Ghent, Belgium, Tel: +32 (0)9 264 35 23, Fax: +32 (0)9 264 42 79 Email: Delphine.Vantomme@UGent.be

132

6.2.2 ADOLESCENTS' INFLUENCE TACTICS AND PARENTAL YIELDING

Aviv Shoham, University of Haifa Gregory M. Rose, University of Washington - Tacoma

Vassilis Dalakas, Berry College Maja Makovec Brenčič, University of Ljubljana

ABSTRACT This paper builds upon the research by Palan and Wilkes (1997) on adolescents’ influence tactics. We expand on their research in four ways: a) using Israeli parents, b) using a quantitative study, c) using a product-specific study, and d) assessing empirically whether demographic variables impact tactics’ frequency and yielding.

Our questionnaire included three parts. The first included items on adolescents’ influence tactics for two products, cereal and athletic shoes. The second part included measures assessing parental yielding, and the third section addressed standard demographic questions. We used a convenience sample of 136 urban, middle-class families in Israel. Only parents whose kids were 10-18 years old were included in the sample. Either the father or the mother completed the questionnaire. Our results indicated that the least used and least effective influence tactics were “shouting, anger, and getting mad” and “guilt trips.” On the other hand, the most heavily used and most effective were “making a direct request without reasons” and “negotiation and providing what he/she considers to be logical and practical arguments.” Consistent with Palan and Wilkes (1997) we found adolescents’ gender to be of moderate importance. In regard to adolescents’ age, our findings indicate that parental yielding to “deal” tactics tends to increase with adolescents’ age. A parent’s education was found to be important; the higher the parent’s education, the lower the use of “deals,” “begging and whining,” and “direct requests” for cereal, and “direct requests” for shoes. Moreover, the higher a parent’s education, the less he/she yields to “begging and whining” for both products. Finally, we found parental income to be of importance. The higher the family’s income, the higher the use of “deals” (for both products) and “negotiations” (for shoes) and the lower the use of “guilt trips” (for both products). Also, higher income was associated with higher yielding for “deals,” “negotiations,” “begging and whining,”

“examples from friends,” and “direct requests” (for both products). For further information contact: Vassilis Dalakas, Campbell School of Business, Box 5024 – Berry College, Mount Berry, GA 30149-5024

133

6.2.3 A TEST OF THE ROBUSTNESS OF CAUSE-RELATED MARKETING

Debra Z. Basil, University of Lethbridge Michael D. Basil, University of Lethbridge

ABSTRACT This study examines whether Cause-Related Marketing is more effective in low involvement situations. An experiment manipulated involvement via price and scenario. Neither involvement manipulation moderated the effectiveness of CRM, suggesting that CRM does not operate as a peripheral cue and does not appear to be a low involvement phenomenon. INTRODUCTION

Cause-Related Marketing (CRM) is a strategy where a for-profit company donates a portion of its sales or profits to a specified non-profit organization (Smith and Alcorn, 1991). Cause-related marketing (CRM) has become very popular in North America (Lavack and Kropp, 1997) and is “the fastest growing form of marketing today” (Smith, 1994). By 1991 nearly 63% of the largest U.S. corporations and charities were familiar with cause-related marketing strategies and 52% even had some prior involvement with it (Barnes, 1991). By 2000, 65% of non-profit organizations and 85% of companies stated that they regularly participate in some form of cause-related marketing (PMA/ Gable Group, 2000).

But what do consumers think about this strategy? According to the Cone Report (1999), consumers are pleased when companies take action in support of causes. As companies seek new methods to address social issues beyond traditional giving, cause programs have flourished and have been well-received by consumers. This issue has shown interest in both academic (Barone, Miyazaki & Taylor, 2000; Osterhus, 1997; Ross, Patterson & Stutts, 1992; Strahilevitz & Myers, 1998; Webb and Mohr, 1998) and practitioner (e.g., Caesar, 1987; Direct Marketing, 1997; Smith, 1994; Smith & Stodghill, 1994) literature sources. Generally, the literature demonstrates that consumers like cause-related marketing practices (Varadarajan and Menon, 1988).

Several previous studies have shown that CRM appears to increase attitudes or purchase intentions for products (Barone, et al., 2000; Berger, Cunningham, & Kozinets, 1996; Sen & Bhattacharya; Strahilevitz & Myers, 1998). These findings suggest that attitude toward the company

and purchase intentions are likely outcomes shaping the effectiveness of CRM. The Elaboration Likelihood Model (ELM)

But when are CRM appeals effective? Petty and Cacioppo (1986) proposed that the process of persuasion is fundamentally different when consumers elaborate on stimuli than when they do not. According to the ELM, when consumers have greater motivation, they are more likely to process via the central route (elaborate processing). That is, they engage in "deep," "controlled," "systematic" analysis of information stimuli (Petty and Cacioppo, 1986).

The ELM predicts that a consumer’s level of involvement affects what type of information is attended to. In a high involvement situation, consumers will devote attention to information, generate thoughts and connections with the product, scrutinize product information, and exercise care in brand selection (Petty and Cacioppo, 1986). In low involvement situations, however, consumers are less motivated to scrutinize product information or engage in elaborate cognitive activity (Petty and Cacioppo, 1986). Processing is characterized as "automatic," "shallow," "heuristic," or "mindless” (Tversky, Slovic and Kahneman, 1990; Johnson and Eagly, 1989; Petty and Cacioppo, 1986). As a result, peripheral aspects of a message, such as positive or negative cues are believed to be more persuasive. In the case of CRM, we wonder whether consumers rely on CRM as a peripheral cue, therefore limiting the effects of CRM.

Importantly, most research to date testing CRM has involved the study of lower involvement situations, and usually lower-priced products (e.g., Barone et al., 2000; Berger et al., 1996; Strahilevitz & Myers, 1998). The focus on lower involvement scenarios does not tell us whether the effects of CRM would hold up for higher priced products. Therefore, this study provides an important test on the robustness of CRM techniques to higher priced products and in higher involvement situations. We expect that the greater scrutiny given in high involvement situations will lead consumers to question the strategy. To increase the generalizability of this study, two manipulations of involvement will be used – price and situational involvement.

134

Hypothesis #1: There will be an interaction between price and CRM such that:

(a) For low priced items, advertisements with CRM will lead to more positive attitudes than ads without CRM; for high priced items, ads without CRM will lead to more positive attitudes than ads with CRM.

(b) For low priced items, advertisements with CRM cues will lead to higher purchase intentions than ads without CRM; for high priced items, ads without CRM will lead to higher purchase intentions than ads with CRM.

Hypothesis #2: There will be an interaction between situational involvement and CRM:

(a) Under low situational involvement, ads with CRM will lead to more positive attitudes than ads without CRM; under high involvement, advertisements without CRM will lead to more positive attitudes than ads with CRM.

(b) Under low situational involvement, CRM ads will lead to higher purchase intentions than ads without CRM; under high involvement, ads without CRM will lead to higher purchase intentions than ads with CRM.

METHOD

To select products and brands for the study, two pre-tests were conducted. The first was used to identify products for which subjects had comparable attitudes. Twenty-eight (28) university students rated their attitudes and involvement with 17 different products. Subjects rated the product category on three 5-point scales measuring personal attitudes toward the product. The attitude scales were anchored with “Unfavourable/Favourable”, “Negative/ Positive” and “Bad/ Good” on either side. The attitude scale was found reliable (alpha = .98).

Subjects showed moderate attitudes toward six products: PDA’s (Pocket PC’s), Cordless Razors, Toothbrushes (battery-operated), In-Line Skates, Condoms and Volleyballs. Subjects’ attitude ratings were comparable for all six products (F < 1) -- PDA’s (M=4.31), Cordless Razor (M=4.46), Toothbrushes (M=4.26), In-Line Skates (M=4.65), Condoms (M=4.79) and Volleyballs (M=4.43). Because we wanted to manipulate situational involvement, the six products with moderate attitudes were selected.

A second pre-test was used to identify brands with comparable attitudes. A different sample of 46 university students rated their attitudes towards the four most successful brands for the product categories selected in pre-test 1. These brands were selected from the Simmons Study of Media and Markets (Simmons Market

Research Bureau, 2000). The scale was based on five-points and was anchored by “Unfavourable/ Favourable”, “Negative/ Positive” and “Bad/ Good” on either side. The reliability analysis for the attitude scale was 0.97.

Results. In the process of selecting similar brands, two of the previous product categories that were initially selected did not have comparable brands and, therefore, were eliminated. We were then left with four product categories and eight brands that showed comparable attitudes. For PDAs, this left Palm Pilot (M = 4.89) and Compaq (M=4.75) (F = 0.198, p>0.05). For toothbrushes Crest (M = 5.29) and Braun Oral B (M = 5.37) were selected (F = 0.091, p>0.05). For In-Line Skates, Rollerblade (M = 5.30) and CCM (M = 5.44) were selected (F = 0.275, p>0.05). For condoms, Trojan (M = 5.23) and Lifestyles (M = 5.26) were the selected brands (F = 0.893, p>0.05). ANOVAs demonstrated that the brands did not differ significantly from one another (all Fs < 1). The Main Experiment

Subjects. One hundred and sixty (160) undergraduate students from a university in Canada were selected for this research. The subjects were drawn from several university summer courses across a variety of business disciplines. Participation in this study was not a class requirement and was strictly voluntary. The data were collected over a one-week period. None of the subjects in the experimental study had participated in the previous pre-tests. The entire procedure took approximately 25 minutes. Subjects ranged from 18 to 50 years of age, 64% were male and 36% were female.

Design. The experimental study used a 2 (price -- high or low) X X 2 (situational involvement -- high or low) X 2 (CRM -- yes or no) factorial design to test the hypotheses. Price, situational involvement, and CRM information were within-subjects factors. The between- subject factors were non-theoretical -- the pairing of the high and low priced products, which of the brands contained the CRM cue, and four different orderings. Price. For this study, price was manipulated by sampling from a number of products, categorized as below or over one-hundred dollars. The low price products retail for around $10, while the high price products retail for about $200.

Situational Involvement. The independent variable “situational involvement” was manipulated as high involvement by describing to the subjects that they should “imagine they were going to purchase” a specific product from the “mock

135

magazine”. In the low involvement state, the manipulation took place through the lack of such an involvement statement; that is, they were not instructed to think about these other products. The situational involvement manipulations can be seen in Appendix A.

CRM. Some previous research has made use of both verbal situations (e.g., “Imagine…”; e.g., Barone et al., 2000). Other research, however, has made use of advertisements to deliver information on the CRM partnership (e.g., Berger et al., 1996). Typically, CRM information is given as a logo with a “donation” statement. For this study, the operationalization of CRM “as the presence of cause-related marketing cues” was manipulated by creating two versions of the ad, one with a CRM logo and one without. The CRM logo was small and placed in the lower right hand corner of the advertisement. The CRM logo also included a statement that “For every purchase, a donation will be made to...”

Manipulation checks. Paper and pencil checks were administered to validate the success of the various manipulations.

Dependent variables. Four questions were used to assess the effects of the ads. Three questions asked respondents to indicate their attitude toward the (1) brand, (2) advertisement, and (3) company on five-point bi-polar adjective scales. A fourth question was also asked to indicate their purchase intentions (i.e. 1=unlikely to purchase, 5=likely to purchase). Factor analysis and reliability analysis suggested that the three attitude measures could be combined as one unified attitude variable (alpha = .90).

Procedure. Mock magazines were used to test the hypotheses. The use of mock magazines is considered a credible approach in the marketing literature as it has been successfully applied in previous research studies (Kirmani and Wright, 1989).

Subjects in the experiment were told they were part of a test on “consumer attitudes toward advertisements.” Subjects were randomly assigned to one of the eight conditions. A different magazine format was created for each of these eight conditions to counterbalance the assignment of CRM cues and particular brand pairings.

Each subject read an introductory instruction letter that explained the study. Then the subjects read the involvement manipulation describing a purchase situation that they were to frame their mind in when examining the mock magazine. (The low involvement situation occurred when ads for non-relevant products were also included in the booklet.) As a result, each

subject was exposed to each of the high and low priced products, each of the two involvement situations, the presence or absence of CRM cues, and each of the two charities.

To enhance external validity, the advertisements used for each brand were based on actual advertisements. Because we manipulated the addition of CRM information, there were no other differences between the advertisements. When the subjects finished the questionnaire, they were debriefed, asked if they had any questions, thanked, and dismissed from the room. RESULTS

Data were analysed with repeated-measures ANOVAs. Price of products, situational involvement, and the presence or absence of CRM cues were within-subjects factors.

Manipulation Checks. The first manipulation was the price manipulation. Subjects were exposed to both high and low priced products, therefore subjects were asked to indicate which of the two product categories were higher in price. The responses were then determined to be either correct or incorrect. Because approximately 98% of the subjects correctly identified the higher priced product in the magazine, this suggests that the price difference was large enough to be noticed by subjects spontaneously. Thus, the manipulation of price was deemed effective.

The next manipulation was situational involvement. Subjects were asked if they remembered the advertisements and brands found in the advertisements with the following open-ended questions; “What were the product advertisements promoting?” and “What brands were advertised in the magazine?” Subjects remembered an average of 93% of the advertisements in the high situational involvement scenario, but only 74% for the non-focused products (X2= [1, 320] = 22.3, p < .05). This suggests that situational involvement was successfully manipulated.

A final manipulation check was also taken to indicate whether the CRM logos were noticed. Subjects were asked open-ended questions: “After reviewing the magazine, do you remember seeing any charitable organization logos in any of the advertisements?” and “Which product advertisements had the charity logos in them?” Overall, the CRM logos were recalled 86% of the time. Sixty-one percent of the subjects correctly remembered the exact charity/ company pairings, while 29% remembered only the charity organization itself. These results suggest that a

136

majority of the subjects noted the CRM logos and the manipulation was effective. Hypothesis 1 Hypothesis 1 predicted a two-way interaction effect for price and CRM. Hypothesis 1a predicted that for low priced products, CRM advertisements will lead to more favourable consumer attitudes than ads without CRM; however, for high priced products, advertisements without CRM will lead to more positive attitudes and than ads with CRM. The hypothesis for this interaction between price and CRM was not

supported (F [1,159] = 0.73, p = 0.40). The interaction is shown in Figure 1. Meanwhile, Hypothesis 1b predicted that for low priced products, CRM advertisements will lead to higher purchase intentions than ads without CRM; however, for high priced products, advertisements without CRM will lead to higher purchase intentions and than ads with CRM. The interaction was not supported for purchase intentions (F [1,159] = 1.62, p = 0.20). The interaction is shown in Figure 2.

Hypothesis 2

The prediction in H2a of an interaction between situational involvement and CRM was not supported for attitudes (F [1,159] = 0.64, p = 0.43). This interaction is shown in Figure 3. The prediction in H2b of an interaction between situational involvement and CRM was also not supported for purchase intentions. Although a weak significant interaction was found (F [1,159] = 3.27, p = 0.10), the level did not achieve significance. The interaction is shown in Figure 4.

3.4

3.5

3.6

3.7

3.8

Lo Hi SITUATIONAL INVOLVEMENT

NO CRM

CRM

FIGURE 3: Attitudes FIGURE 4: Purchase Intentions

Lo Hi SITUATIONAL INVOLVEMENT

NO CRM

CRM3.0

3.2

3.4

2.8

FIGURE 1: Attitudes

3.4

3.5

3.6

3.7 3.8

NO CRMCRM

Lo Hi PRICE

FIGURE 2: Purchase Intentions

2.8

3.0

3.2

3.4

Lo Hi PRICE

NO CRM CRM

137

DISCUSSION The study tested the robustness of Cause-

Related Marketing. In general, the results show that CRM remains effective, even in high involvement situations.

First, H1 predicted an interaction between price and CRM. It was anticipated that with higher priced items, CRM claims would be less effective than for lower priced items. The results did not support this prediction. Second, H2 predicted that when consumers were more involved with a product (i.e. paying close attention to product details), CRM claims would be less effective than under low situational involvement. The results demonstrated that this prediction was incorrect.

Overall, the conclusion is that CRM claims have a favourable impact on consumer attitudes and purchase intentions, regardless of the degree of price and situational involvement. These finding suggest that cause-related marketing appears to be a good way to promote products. This was evidenced by favourable attitudes and purchase intentions towards companies using CRM.

Limitations. Several limitations of the study arose as a result of the experimental setting. This artificial environment probably created several unnatural factors that influenced subjects’ levels of situational involvement. First, although there was a significant difference between the two situational involvement conditions, the high level of recall suggests that even in the low situational involvement conditions, people paid attention to the messages. Just being in the experiment and being observed may have activated some feelings of personal relevance and increased subjects’ general feelings of involvement. However, since most previous research has established the effectiveness of CRM in low involvement situations, this does not appear to be a fatal flaw as it does extend our knowledge to high-involvement situations.

Second, the advertisements were presented within an artificial context void of the editorial and story content of a real magazine. This probably also increased subjects’ attention to the ads beyond that expected to occur in a natural exposure environment. Third, participants were not actual consumers purchasing a product. The nature of the causes (to benefit disaster relief and heart and stroke patients) may have made it difficult for some respondents to express any negative feelings toward the product, brand, company or advertisement. Although these factors may limit the external

validity of the effects, the do not threaten the internal validity of the design or measures.

The other limitations of this study arise from sampling. First, while these subjects represent a good mix of students, validity would be enhanced by conducting the study with a sample of non-university adults. Second, the focus on only four product categories may not represent the entire universe of decisions. Assessing consumer attitudes and purchase intentions across a wider range of product categories will enable researchers to examine consumer perceptions and decision processes towards CRM claims.

Future research. Since most subjects were favourably motivated toward the donation, later research may want to frame the donation as either the company’s money (a portion of the profits) or the consumer’s (a portion of the selling price). This is expected to increase the scrutiny with which customers evaluate CRM claims. We expect that people are more critical of CRM donations in high price situations and when told the money comes out of their own pocket. REFERENCES Barnes, N.G. Joint venture marketing: A strategy

for the 1990’s. Health Marketing Quarterly, 1991; 9 (1-2): 23-36.

Barone, M., Miyazaki, A.D., & Taylor, K.A. The

influence of cause-related marketing on consumer choice: Does one good turn deserve another? Journal of the Academy of Marketing Science, 2000; 28(2): 248-262.

Berger, I. E., Cunningham, P., & Kozinets, R.V.

The processing of cause-related advertising: Cues, arguments, biases or motivators. Kingston, Ontario, 1996.

Caesar, P. Cause-related marketing: The new face

of corporate philanthropy. Nonprofit World, 1987; 5(4): 21-26.

Cone/ Roper Study. A benchmark survey of

consumer awareness and attitudes towards cause-related marketing. Cone Communications. Boston, Massachusetts, 1999.

138

Direct Marketing. Cause related marketing makes new strides in the United Kingdom. Direct Marketing, 1997; 60(4): 7-8.

Johnson, B.T., & Eagly, A.H. Effects of

involvement on persuasion: A meta-analysis. Psychological Bulletin, 1989; 106(2): 290-301.

Lavack, A.M., & Kropp, F. Impact of values on

consumer attitudes toward cause-related marketing: A cross-cultural perspective. Proceedings of the Sixth Symposium on Cross-Cultural Consumer and Business Studies, 1997; 280-285.

Osterhus, T. Pro-social consumer influence

strategies: When and how do they work? Journal of Marketing, 1997; 61(3): 16-29.

Petty, R., & Cacioppo, J. The elaboration likelihood

model of persuasion. Advances in Experimental Social Psychology, 1986; 19: 124-181.

PMA & Gable Group. Survey of Cause Marketing.

Retrieved July 3, 2002 from http://www.pmalink.org/causemarketing2000/index.html. New York, NY, 2000.

Ross, J.K. III, Patterson, L.T., & Stutts, M.A.

Consumer perceptions of organizations that use cause-related marketing. Journal of Academy of Marketing Science, 1992; 20(1): 93-97.

Sen, S. & Bhattachrya, C. B. Does doing good

always lead to doing better? Consumer reactions to corporate social responsibility. Journal of Marketing Research, 2001; 38: 225-243.

Simmons Market Research Bureau. Simmons 2000

Study of Media and Markets. New York: Simmons Market Research Bureau, 2000.

Smith, C. The new corporate philanthropy. Harvard

Business Review, (May-June) 1994. Smith, S.M., & Alcorn, D.S. Cause marketing: A

new direction in the marketing of corporate responsibility. Journal of Consumer Marketing, 1991; 8(3): 19-35.

Smith, G., & Stodghill, R. Are good causes good

marketing? Business Week, (March) 1994; 21: 64-66.

Strahilevitz, M., & Myers, J. Donations to charity as purchase incentives: How well they work may depend on what you are trying to sell. Journal of Consumer Research, 1998; 24: 434-446.

Varadarajan, P. R., & Menon, A. Cause-related

marketing: A co-alignment of marketing strategy and corporate philanthropy. Journal of Marketing, 1988; 52: 58-74.

Webb, D., & Mohr, L.A. A typology of consumer

responses to cause-related marketing: From sceptics to socially concerned. Journal of Public Policy & Marketing, 1998; 17(2): 226-238.

For more information contact: Dr. Debra Basil, Faculty of Management University of Lethbridge, 4401 University Drive, Lethbridge, AB TIK 3M4 Canada Phone: (403)329-2164; Fax (403)329-2038 Email: debra.basil@uleth.ca

139

6.3.1 FEELING THE URGE TO SPLURGE: THE EFFECT OF VALENCE AND AROUSAL OF CONSUMER MOOD STATES ON IMPULSIVE CONSUMPTION

Alexander (Sasha) Fedorikhin, University of Southern California

Vanessa Patrick, University of Southern California ABSTRACT Feeling states like mood and emotions have been shown to influence a wide variety of internal processes and observable behaviors (Gardner, 1985; Isen, 1999, Pham, 1998). However, one domain that has been relatively unexplored in which the volatility of affective states liaises with consumer behavior is that of impulsive consumption (see Rook and Gardner, 1993 for an exception). In this paper, we bring together two largely independent streams of literature, mood and impulsive consumption, to examine the influence of consumer mood states on impulsive consumption and choice. Impulsivity has been gaining interest among the consumer behavior researchers over the last two decades. Rook (1987) conceptualized psychological impulses as strong, sometimes irresistible urges triggered by confrontations with certain stimuli. He also found that impulsive decisions are normally associated with the struggle of willpower and desire, as they involve a trade-off between short-term gratification and long-term negative consequences. We extend prior research by examining both impulsive choice and impulsive consumption. We examine not only the self-control involved in consumer choice but also the regulation of consumption of the impulse-triggering product following the “impulsive choice”. Thus, we look at impulsivity not merely as a lapse of self-control at a particular point in time but as a process that is subject to self-regulation. We also examine the role of both the valence and arousal components of consumers’ mood states in impulsive consumption and aim to discern the processes that underlie the impact of different mood states on impulsive consumption. The extant literature presents conflicting views on the role that mood might play in impulsive choice and consumption. The literature on the motivational influence of mood on impulsive consumption would suggest that people in a positive mood would tend to engage in impulsive consumption leading to immediate self-gratification in order to maintain or enhance their positive moods (Isen, 1984, Singer and Salovey, 1988; Thayer, 1978, 1989) while

people in a negative mood would engage in impulsive consumption in order to repair or improve their mood. On the other hand, the delay of gratification literature suggests that positive (vs. neutral) mood increases the ability to delay gratification (Fry 1975, Moore, Clyburn and Underwood 1976, Seeman and Schwarz 1974). We expect arousal, the other widely recognize dimension of affect, to moderate the effect of positive mood on impulsive consumption. Drawing on the extant literature, we would expect that high arousal moods increase cognitive load (Paulhus and Lim, 1994) and deplete people’s cognitive resources (Baumeister et al. 1998; Muraven et al. 1998) as compared to the low arousal moods. People really need their cognitive resources to be able to resist the immediate temptation (Shiv and Fedorikhin, 1999; 2002). Hence, we would expect consumers in high arousal moods (with both positive and negative valence) to be more impulsive than those in the neutral mood. In the first of a series of planned studies we examined the impact of the valence component of mood (positive vs. negative vs. neutral) on impulsive consumption, keeping arousal constant. In this study, video clips were used to manipulate positive, negative and neutral mood states. Participants were offered a snack, which was a choice between M&M candies and grapes, and their choice was recorded. These snacks were pretested to confirm that the M&Ms and grapes were clearly differentiated in terms of their impulsive and “sinful” nature. Participants were allowed to eat as much as they wished during the experimental session but were not allowed to take away any leftovers. The quantity of grapes and M&Ms left behind was also recorded. The results of this study reveal that participants in a positive mood were more likely to make a less impulsive choice (grapes) than those in the neutral and negative conditions (M&Ms). We also found that participants in the positive mood condition consumed fewer M&Ms than those in the negative and neutral mood conditions. In the next set of studies, we intend to look at both the valence and arousal components of mood to

140

examine the complete model. In addition, we intend to examine the role of consumer impulsivity, an individual difference variable, and tap deeper into the processes underlying the impact of different mood states on impulsive choice and consumption. In conclusion, this paper demonstrates yet another aspect of consumer behavior that is susceptible to consumer mood states and contributes to the impulsive consumption literature by exploring the importance of ambient affect (mood) that predisposes a consumer to purchase impulsively. For further information contact: Alexander (Sasha) Fedorikhin, Marshall School of Business, University of Southern California, Los Angeles, CA 90089, email: afedorikhin@marshall.usc.edu

141

6.3.2 THE INFLUENCE OF INCIDENTAL AFFECT ON CONSUMERS’ FOOD INTAKE

Nitika Garg, University of Pittsburgh Brian Wansink, University of Illinois at Urbana-Champaign

Jeff Inman, University of Pittsburgh ABSTRACT Lay thinking relates behavior and consumption with the affective state that a person is experiencing. Indeed, incidental affect – the affect consumers imbue from their environment in isolation to the task at hand – has been shown to influence in-store shopping (Woodruffe 1997) and in-home food choice. Recent work on “comfort foods” has suggested that affect influences the choice of these foods because of the temporary psychological comfort some people claim these foods provide them (Wansink and Sangerman 2000). With the nation increasingly concerned with overeating and obesity, understanding how incidental affect influences food intake and whether it can be moderated through warnings or nutritional labeling is an important topic. The implications are of interest to public policy and health professionals as well as to consumers who are interested in controlling their food intake. While some researchers have focused on how marketing-related variables such as stockpiling, promotions, or advertising influence consumption (e.g., Ailawadi and Neslin 1998; Chandon and Wansink 2002; Wansink and Ray 2000) and others have examined how nutritional information for a product influences its choice or consumption (e.g., Baltas 2001; Hackleman 1981; Miller et al. 1998), little research has focused on how contextual factors such as affect (e.g., sadness vs. happiness) influence consumption. In this research, we build on the hedonic contingency hypothesis (Wegener and Petty 1994; Wegener, Petty, and Smith 1995) to examine how different discrete affective states (e.g., sadness, happiness) influence consumption and whether this relationship can be attenuated by nutritional information. The objective of this research then, was to examine the influence of incidental affect on consumption. Specifically, we used a field study and two lab studies to examine the differential effect of sadness and happiness on the amount of food consumed by individuals and the extent to which it is influenced by exposure to nutritional information. Sadness is an affective state that is associated with a sense of loss and research has found that subjects in this

state try and compensate for this by seeking rewards and manage their mood (Wegener and Petty 1994; Raghunathan and Pham 1999;). Happy subjects on the other hand have no need for mood repair but they do concern themselves with mantaining their affective state so that they continue in the happy state (Wegener and Petty 1994). Thus, we hypothesized that subjects in the sadness condition will consume more as compared to those in the happy condition. In Study 1, we examined the affect-consumption relationship within a movie. This design afforded us the opportunity to examine this relationship in a common, innocuous, and externally valid situation. Because of the rapid pace of the movie and because it left most people in a generally favorable mood, we can generalize the findings only to neutral and positive states. The results supported our hypothesis that happy individuals consumed less than those in a relatively less positive state. Specifically, respondents exhibited a five gram decrease in consumption for each one-point increase in our 18-point affect scale. While Study 1 supports our hypotheses and illustrates that the phenomenon exists in a field study, Study 2 gives us additional confidence in the underlying theory that we believe explains this relationship. In Study 2 we used across-movie differences to see if a relationship between affect and consumption emerged in a within-subject context. We found that the same individuals consumed almost 30 percent more when they watched a sad movie versus a happy movie. Over ¾ of our respondents exhibited an increase in consumption in the sad film versus the happy film, with ½ of the respondents exhibiting a substantial (over 20 grams difference) increase. The final study not only establishes the robustness of our results by replicating the phenomenon using a second method of affect induction (a written induction) with the three affective states at the same time – happy, sad and neutral – but it also establishes nutritional information as a moderating factor for the influence of affect on consumption. It also contributes in showing how this effect can be attenuated. When nutritional information was

142

absent, sad individuals acted in a way consistent with what would be expected if people wanted to reward themselves and to try and overcome their “sense of loss” (Lazarus 1991) by consuming more than happy individuals. However, when information was provided about the negative nutrients of the product, consumption levels of individuals in all three affective states dropped to comparable levels. Another interesting finding is that whereas sad and neutral individuals showed a substantial decline in their consumption, happy individuals seemed to be uninfluenced by nutritional information. Thus, it seems that happy individuals are already mindful of their consumption levels in order to maintain their positive state and the presence of nutritional information cannot drive their consumption any lower, as suggested by the hedonic contingency hypothesis (Wegener and Petty 1994). For further information contact: Nitika Garg, Katz Graduate School of Business, University of Pittsburgh, Pittsburgh, PA 15260, nitika@katz.pitt.edu

143

6.3.3 DO PRODUCTS HAVE COOTIES? THE LAW OF CONTAGION IN PRODUCT EVALUATIONS

Andrea C. Morales, University of Southern California

Gavan J. Fitzsimons, Duke University

ABSTRACT Although Americans pride themselves on being a modern and well-developed nation, this research demonstrates that they sometimes behave in ways that are consistent with magical ideas characteristic of more primitive cultures as opposed to scientific based reasoning. Specifically, we show that people make consumption decisions in ways that are consistent with the “law of contagion,” which reasons that when two objects come into contact with one another they permanently exchange properties. We test this idea in the context of consumer packaged goods, so that true microbial contamination is not a possibility and changes in consumption patterns would provide evidence of this so-called “magical thinking” and have no real, rational basis. In study 1, we show that when a product that elicits feelings of disgust (feminine hygiene products) touches a clear package containing a consumable product (cookies), the desire to consumer the cookies is lower than when the disgusting product is simply present in the same shopping basket. Since the contamination effects do not hold when the disgusting product is present, but not touching the consumable product’s package, the decrease in desired consumption cannot be attributed to negative affect induced by the disgusting product’s presence in the basket. Rather, it suggests that consumers perceive the disgusting product as contaminating the consumable product upon contact of the two packages, thereby making it less desirable. Despite predictions that the contamination effects would be stronger for men than women in this product category, the results indicate no interaction with gender. In addition, there was no difference in desired consumption under normal versus constrained cognitive capacity, suggesting that there may be a non-conscious component to these effects. Study 2 provides further evidence of consumers following the “law of contagion” by showing that the disgusting and consumable products’ packages actually have to be in contact and not just positioned next to one another, in order to produce a decrease in desired consumption. In study one,

the disgusting product was either touching the consumable product or there were two other products positioned in between them. Thus, it was unclear whether the decrease in desired consumption was the result of contamination from the product packages touching or simply a negative association from a disgusting product being positioned in proximity to a consumable product. The results of study 2, however, indicate that it is contamination driving these effects and not association, since desired consumption decreases only when the two product packages are in direct contact with one another. In addition, study 2 also tests whether people who feel disgust more often and more strongly than others demonstrate stronger contamination effects between products. Using a disgust scale adapted from Rozin, Millman, and Nemeroff (1986), we show that people who report being less affected by disgust actually lower their desired consumption of products that are touching disgusting products more than those who report being more affected by disgust. Again this suggests that people may not be fully aware that their consumption patterns are being influenced by contagion. In study 3, we investigate the moderating role of package type on contamination effects. The results show that desired consumption of a consumable product only decrease upon contact with a disgusting product when the consumable product is displayed in a clear, as opposed to a solid, package. Since the solid package does not allow consumers to visualize the disgusting product actually touching the consumable product, it inhibits the perception that the two products are exchanging properties. As a result, consumption of a product displayed in a solid package is not affected by contact with a disgusting product. In addition to replicating the moderating role of package type, study 4 shows that contamination effects are not limited to overall product evaluations, but can also occur at the attribute level. Results demonstrate that when a fattening product (lard) is displayed in contact with a low-fat product (rice cakes), the low-fat product is not only perceived as less desirable, but is thought to be more fattening as well. This finding indicates that

144

consumers are indeed evaluating products as though they believe physical contact results in a transfer of specific properties from one product to another. In this case, the lard is thought to transfer its fat content to the rice cakes it is touching. Consistent with previous research on contagion, these results suggest an asymmetry in contamination effects for negative versus positive contagion. Specifically, they indicate that negative contagion, where contact with another object devalues the original object, is more powerful than positive contagion, where contact with another object enhances the value of the original object. In the current context, this suggests that products with negative perceptions can lower consumption of products with positive perceptions when they come in contact, but products with positive perceptions are not able to increase consumption of products with negative perceptions simply by touching. Assuming that low-fat is a more positive attribute, this means that a fattening product can make a low-fat product seem less desirable and less healthy, but a low-fat product cannot make a fattening product seem more desirable and more healthy. Implications of contamination effects for shelf displays and packaging are discussed. For further information contact: Andrea C. Morales, Marshall School of Business, University of Southern California, Los Angeles, CA 90089-0443, acmorale@usc.edu

145

6.4.1 EXCITATION TRANSFER THEORY AND ADVERTISING EFFECTS ATTRIBUTABLE TO AD PLACEMENT ORDER, SUSPENSEFUL PROGRAMMING, AND AD EXECUTION STYLE

Colleen Bee, University of Oregon

Robert Madrigal, University of Oregon ABSTRACT

Advertising messages occur in a variety of program contexts. Involvement with these programs varies greatly and therefore consumer response to advertising embedded within these different programs could also vary. Several studies have examined the effect of viewing contexts on consumer reaction to advertising (Coulter & Punj, 1999; Goldberg & Gorn, 1987; Prasad & Smith, 1994). However, few studies have considered consumer reaction to ads embedded within suspenseful drama. The current study examines the effect of suspenseful programming, ad placement order, and ad execution style on affective and evaluative responses to advertising. One model for understanding the effect of placement order on viewers’ reactions to embedded advertising during suspenseful drama may be derived from excitation transfer theory (Zillmann, 1991). Excitation theory suggests that viewers exposed to suspenseful drama: (a) experience an elevation of sympathetic excitation and (b) appraise the liked characters’ actions as dysphoric. The intensity of these dysphoric reactions influences the intensity of sympathetic activity. As the intensity increases, so too does the magnitude of residual excitation. Because sympathetic activity decays at a relatively slow rate, unnoticed residual arousal is transferred to subsequent reactions and judgments. As a result, viewers may misattribute some of the residual arousal to other stimuli existing concurrently in the environment and may be more polarized in their evaluation of those stimuli. In the context of this study, we propose that exposure to suspenseful drama will yield higher levels of arousal than exposure to nonsuspenseful drama for which excitement is nonexistent. The residual arousal will then be transferred to viewers’ reactions (e.g., emotional responses and evaluative judgments) to advertising embedded around the suspenseful drama. The magnitude of viewers’ responses is likely to be more pronounced when the ad is embedded immediately after exposure to a suspenseful drama. Thus, placement order is viewed as an important predictor of excitation transfer.

Previous research has examined consumer reactions to advertising within the excitation transfer paradigm. For example, Gorn, Pham, and Sin (2001) manipulated level of arousal through music and examined consumer response to either positively or negatively toned ads. They found that arousal polarized consumers’ response to ads in the direction of the ad’s affective tone and that this polarization effect was more pronounced for self-referential items than for object-referential items. In a somewhat different approach, Soldow and Principe (1981) examined the effectiveness of advertising that was embedded in suspenseful and comedic programs. They found that program environments with higher involvement levels cause ads to be less effective than ads embedded in lower involvement programs. These results are contrary to the predictions using excitation transfer theory and the current study. It is worth noting, however, that Soldow and Principe did not explicitly consider ad placement order at the conclusion of the program or the application of excitation transfer theory to predict and explain their results.

The current study exposed subjects to dramatic program content that was either suspenseful or nonsuspenseful. Consistent with past research examining the effects of suspenseful drama on viewers’ reactions (Gan et al. 1997), the context used in this study was a competitive sporting event (i.e. men’s basketball). In the suspenseful game, the point total was extremely close throughout with the home team winning in the last five seconds; whereas in the nonsuspenseful game the home team jumped out to a quick lead in the first half and increased through the second half. The ad programming featured four ads – two presented at the midway point of the program and two at the end. The execution style of the ads was either suspenseful or nonsuspenseful with the order of the four ads varied across respondents. It is proposed that residues of excitation from a suspenseful drama can intensify subsequent affective and evaluative responses to advertising. Furthermore, the residual excitation will interact with ad placement order and ad execution style. A mixed design was used to test the effects of ad placement order, ad execution style and level of

146

program suspense on the relationship between viewer involvement (arousal) with the game and postconsumption reactions to four advertisements. The four ads were embedded in the content as follows: two ads were shown following the first game half of the game and two ads were shown immediately following the end of the game. After viewing each ad, respondents indicated their (a) enjoyment with the ad, (b) emotional responses to the ad, and (c) attitude toward the ad. The results support our propositions. First, ad placement order had a significant effect on viewer involvement and ad reaction measures. More specifically, the correlation between viewer involvement and ad reaction measures (i.e., ad enjoyment, emotional response, and attitude toward the ad) was significantly greater for the ad inserted immediately after the conclusion of the suspenseful drama (placement order 3). Furthermore, the level of program suspense moderated this relationship. For the suspenseful game, the correlations between involvement and each ad reaction (ad enjoyment, emotional response, and attitude toward the ad) at placement order three is significant; however, for the nonsuspenseful game, these correlations at placement order three were not significant. Further analysis suggests a match between program context (high suspense) and ad execution style (suspenseful ad). Significant positive correlations were found between involvement and each ad reaction for suspenseful ads in the suspenseful game condition. Surprisingly, significant negative correlations were found between involvement and ad enjoyment and attitude toward the ad for suspenseful ads in the nonsuspenseful game condition. As expected, nonsignificant correlations were found for nonsuspenseful ads in both game conditions. Based on excitation transfer theory (Zillmann, 1971), the current study indicates that arousal elicited during a highly suspenseful drama intensifies subsequent affective and evaluative responses to advertising. Arousal interacts with ad execution style and ad placement order to influence consumer response to ads. Findings from this study suggest that advertising in suspenseful programs could enhance the communication effectiveness of ads that immediately follow the conclusion of the. Excitation transfer theory helps explain this response to advertising following exposure to highly involving suspenseful drama. The high arousal evoked while viewing suspenseful program

content carried over to the subsequent reactions to advertising embedded in the program. The current study examined viewer reaction to advertising immediately following exposure; further research in this area should also examine the lasting effects of excitation transfer on ad and brand recall, as well as the current ad reaction measures. The context for this study was the manipulation of levels of suspenseful drama. Additional research might also examine how different suspenseful contexts or the outcome of suspenseful drama influences excitation transfer and subsequently, viewer reaction to embedded advertising. Furthermore, this line of research should also examine other explanations for these effects, such as a mood transfer paradigm. For further information contact: Colleen Bee, Charles H. Lundquist College of Business, 1208 University of Oregon, Eugene, OR 97403, (541)346-4179, cbee@darkwing.uoregon.edu

147

6.4.2 FEAR APPEALS IN TRAFFIC SAFETY ADVERTISING - THE MODERATING ROLE OF CONTEXT, ADVICE, DUALITY AND ANXIETY

Wim Janssens, University of Antwerp

Patrick De Pelsmacker, University of Antwerp Liesbeth Bulkmans, University of Antwerp

ABSTRACT

RESEARCH QUESTIONS AND HYPOTHESES The study investigates the responses to three different levels of fear appeals, combined with the presence or absence of an advice or recommendation, urging people to reduce speed in traffic. Based on models explaining the responses to various levels and types of fear appeals and the findings of earlier empirical work on fear appeals (Keller, 1999; Stephenson & Witte, 2001), context effects (Yi, 1993), duality (Williams & Aaker, 2002), and level of anxiety (Gallacher & Krieger, 1994), the following research hypotheses are advanced:

H1: A high fear appeal, as opposed to a moderate or a low fear appeal, leads to the highest attention level, while a moderate fear appeal, as opposed to a low or a high fear appeal, leads to the most positive attitude towards reduced speeding.

H2: A fear appeal with advice, as opposed to a fear appeal without advice, leads to a more positive attitude towards reduced speeding.

H3: A fear appeal in a good news context, contrary to a bad news context, will lead to a more positive attitude towards the ad, and a more positive attitude towards reduced speeding for those individuals who have a higher proclivity to accept duality (contrast effect). A fear appeal in a bad news context, contrary to a good news context, will lead to a more positive attitude towards the ad, and a more positive attitude and behavioral intention towards reduced speeding for those individuals who have a lower proclivity to accept duality (priming effect).

H4: For individuals with a high general level of anxiety, low or medium levels of fear appeals, or a high level of fear appeal combined with a recommendation to control danger, lead to more positive attitudes and behavioral intentions towards reduced speeding. For individuals with a low general level of anxiety, the level of

the fear appeal and whether there is a danger control recommendation or not, does not lead to differences in the attitude and behavioral intention towards reduced speeding.

DATA COLLECTION A twenty-person jury selected three pictures to be used to compose advertisements, a low, medium and high fear level picture. On the basis of these three pictures, 12 advertising stimuli were composed, based on a between-subjects 3x2x2 full factorial design in which three factors were manipulated. The first factor was the level of fear in de ad, i.e. a low, medium or high fear picture. The second factor was the presence or absence of an advice or recommendation in the ad (‘drive more slowly’). The third factor was the type of context, i.e. good news articles or bad news articles. Manipulation checks of the type of ad and the type of context indicate significant differences between the stimuli in the expected direction. Participants were undergraduate students recruited in the last year of high school and the first year of university. Each respondent randomly received one of the 12 possible combinations and a questionnaire. They were told that they were involved in a test for a new newspaper, and they were instructed to look at the newspaper pages just as they would look at an ordinary newspaper. After data cleaning for incomplete and inconsistent answer patterns, 34 respondents were removed and a total useful sample of 589 respondents was obtained. In the questionnaire, 17 Attitude towards the ad items and 12 attitude towards speeding items were presented. Based on exploratory and confirmatory factor analysis three underlying dimensions of the attitude towards the ad (Aad) were found: Aad_information (e.g. ‘this spot gives useful information’), Aad_understanding (e.g. ‘this ad is confusing’ and Aad_attention (e.g. ‘this ad catches my attention’), as well as one ‘attitude towards speeding’ dimension. The summated scales of the items loading exclusively on each dimension are used as dependent variables. Five independent variables are measured. Besides the manipulated

148

factors fear appeal (3 levels), advice (2 levels), and type of context (2 levels), also the level of duality and anxiety is measured. Duality is measured using the scale of Ryan (1984). The anxiety level is measured by the Trait component of the STAI (State and Trait Anxiety Index) developed by Spielberger (1983). Respondents have to rate 20 4-point-items and these scores are summated to obtain the Anxiety Trait index (thus ranging from 20 to 80). RESULTS Three different models are estimated to test the various hypotheses. In the first model, the basic (3x2x2) factorial design is tested. In the second model the level of anxiety is added as a covariate. In the third model the levels of anxiety and duality are defined as two-level factors, and the resulting (3x2x2x2x2) model is tested, including three-factor interaction effects. One of the most consistent results of this study is the superiority of high fear appeals. The importance of moderating personality traits such as the level of anxiety and duality of an individual was also clearly demonstrated. For the attitude towards speeding, on the one hand the personality trait ‘anxiety’ appears to play a moderating role in combination with advice, and on the other hand, it appears that with regard to context effects, duality appears to play a moderating role. Duality plays an important role when a fear appeal is combined with a specific context. For individuals with a low degree of duality, i.e. people who can more easily process stimuli with mixed emotions and appeals, a fear appeal in a contrasting (positive) context appears to be the most effective. High duality individuals react more positively upon fear appeals in a congruent context. For low and high anxiety individuals, the picture that emerges in terms of the effect of different types of fear appeals is quite subtle. Low anxiety individuals seem to react most positively to low fear appeals without advice. This lends support to the idea that especially these people do not want to feel manipulated and patronized. High anxiety individuals to a certain extent seem to react more positively upon recommendations, but mixed results appears whether this reaction is most positive in a low or a high fear appeal. REFERENCES

GALLAGHER, F. & KLIEGER, D.M. (1994) Sex Role Orientation and Fear. The Journal of Psychology, 129 (l), 41-49.

KELLER, P.A. (1999) Converting the Unconverted: The Effect of Inclination and Opportunity to Discount Health-Related Fear Appeals. Journal of Applied Psychology, 84 (3), 403-415.

RYAN, M. (1984) Monitoring Text Comprehension: Individual differences is epistemological standards. Journal of Educational Psychology, 76(2), 248-258.

SPIELBERGER, C.D. (1983) Manual for the State-Trait Anxiety Inventory (STAI-form Y). Palo Alto, CA: Consulting Psychologists Press.

STEPHENSON, M.T. & Witte, K. (2001) Creating Fear in a Risky World, in: Public Communication Campaigns, Newbury Park, CA: Sage Publications, 88-102.

WILLIAMS, P. & AAKER, J.L. (2002) Can mixed emotions peacefully coexist? Journal of Consumer Research, 28, 636-649.

YI, Y. (1993) Contextual Priming Effects in Print Advertisements: The Moderating Role of Prior Knowledge, Journal of Advertising, 22(l), 1-10.

For further information contact: Patrick De Pelsmacker, Universiteit Antwerpen Management School, St.Jacobsmarkt 13, B-2000 Antwerpen, Belgium. E-mail: patrick.depelsmacker@ua.ac.be

149

6.4.3 THE IMPACT OF CONTEXT INDUCED FEELINGS ON EFFECTIVENESS OF AD APPEALS

EVOKING THE SAME FEELING OR A DIFFERENT FEELING OF THE SAME VALENCE

Tine Faseur, Ghent University ABSTRACT

This study concentrates on the impact of context induced feelings on the effectiveness of embedded ads. Quite some contradictory results can be found in previous research concerning affective context effects. Although most authors observed that people evaluate ads more positively after being exposed to a funny or happy program (e.g. De Pelsmacker et al. 2002, Goldberg and Gorn 1987, Kamins and Marks 1991, Yi 1990), there is less consensus concerning the effect of a negative feeling inducing program. In his article Forgas (1995) reviewed existing research concerning the role of affective states in social judgments. He developed an integrative theory, “the Affect Infusion Model”, which gives a comprehensive explanation of how affect can influence our thinking and judgments based on four alternative processing strategies. When, on the one hand, ‘direct access’ or motivated processing is used, the information search will be predetermined and directed, and little generative and constructive processing will be required. This limits the scope of affect infusion effects. As a result, when these strategies are employed, judgmental outcomes are uninfluenced by or are incongruent with the prevailing affective state (cfr. the “Consistency effect model” supported by Kamins and Marks (1991)). When, on the other hand, ‘heuristic’ or ‘substantive’ processing is used some degree of constructive thinking is involved and affect will have a mood-congruent effect. This will occur through two alternative mechanisms. According to the “affect as information principle” (Sinclair et al. 1994) feelings can directly inform judgments during fast heuristic processing as judges use their affective state as a shortcut to infer their evaluative reactions to a target. According to the second mechanism, the “affect priming principle” (or the “mood congruency model”) (Bower 1981), affect may indirectly influence judgments during substantive processing, i.e. through its selective influence on attention, encoding, retrieval and associative processes. This model, states that when a positive (negative) feeling is induced, positive (negative) material in memory will become more accessible and respondents will generate more (less) favorable reactions towards the target stimuli.

One practical remark is that almost all previous models are valence-based, meaning that they are based on the assumption that feelings and emotions can be placed on a continuum going from “positive” to “negative” affect. Implicit in this assumption is that all positive feelings and all negative feelings are seen as essentially equivalent. This assumption has for long been generally accepted and is consistent with the conviction of many researchers (e.g. Russel et al. 1989, Reisenzein 1994, Feldman 1995). Recently, however, quite some researchers (e.g. Babin et al. 1998, Raghunathan and Pham 1999, Mitchell and Brown 2001) have questioned this view and argue that all feelings should be seen as unique and discrete. For example, Raghunathan and Pham (1999) and Lerner and Keltner (2000) found that different negative emotions can lead to differential effects on judgment and choice (related to risk perception). Relating to context effects, Mitchell and Brown (2001) found that anger and sadness had different effects on ad evaluation and information processing. However, no studies have yet been conducted concerning the possible differential effects of different positive feelings. In light of the discussion of the bipolarity versus the uniqueness of feelings, the objective of the current study is to investigate the impact of a positive (warm) and negative (fear) print context on embedded ads evoking 1. a feeling identical to the context (respectively a warm and fear appeal) and 2. a different, but valence-congruent feeling (respectively an exciting and sad appeal). In this experiment we assume that respondents use heuristic or substantive processing when forming an attitude toward the target ad. Therefore, we expect that (H1) in general the warm (positive) context will lead to higher attitudes toward the ad and the fear (negative) context will lead to lower attitudes toward the ad than in the condition where no context is shown. Furthermore, following the growing evidence for the uniqueness of feelings (Raghunathan and Pham (1999), Lerner and Keltner (2000), Mitchell and Brown (2001)), we would expect that (H2) the warm context article will have a different effect on attitudes toward the advertisements evoking a warm as compared to an exciting feeling. Similarly, we would expect that

150

(H3) the fearful magazine article will have a different effect on attitudes toward the advertisements evoking fear versus sadness. To test the hypotheses two experiments were conducted in a magazine context. In both experiments a two (no context versus a warm/fearful context article) by two (print ad reflecting the same feeling as the context, i.e. a warm/fear ad, versus print ad reflecting a different feeling from the context but similar in valence, i.e. an exciting/sad ad) between subjects design was constructed. All ads were promoting a home cinema installation. The feelings were manipulated by means of the large screen showing a shot from a movie scene. Anova-results for the first experiment show that the warm context induced significantly more positive ad and brand attitudes and purchase intentions as compared to when no context article was shown. No significant effect of ad type (warm versus exciting), nor a significant interaction effect between the context and the ad appeared. For the second experiment, it is found that a fear as compared to no context induced a significantly more negative attitude towards the ad and a marginally significantly less positive purchase intention. The effect of the context on attitude towards the brand was insignificant. Furthermore, as was the case for the first experiment, neither significant main effects of ad type (fear versus sad appeal), nor significant interaction effects between context and ad type were found. So the results of this study do not seem to suggest that in an advertising context valence-congruent feelings have to be treated as really different. However, more research is needed to see whether these conclusions can be confirmed. Future studies should focus on, amongst other, different emotional contexts, different and stronger emotional advertising appeals and other media than magazines. REFERENCES Babin, B. J., Darden, W. R. & Babin, L. A. (1998).

Negative Emotions in Marketing Research: Affect or Artifact? Journal of Business Research, 42, 271-285.

Bower, G. H. (1981). Mood and Memory. American Psychologist, 36, 129-148.

De Pelsmacker, P., Geuens M., & Anckaert P. (2002). Media context and advertising effectiveness: the role of context style, context quality and context-ad similarity. Journal of Advertising, 31, 49-61.

Feldman, L. A. (1995). Variations in the Circumplex Structure of Mood. Personality and Social Psychology, 74, 967-984.

Forgas, J. P. (1995). Mood and Judgment: The Affect Infusion Model (AIM). Psychological Bulletin, 117, 39-66.

Goldberg, M. E. & Gorn, G. J. (1987). Happy and sad TV programs: how they affect reactions to commercials. Journal of Consumer Research, 14, 387-403.

Kamins, M. A., Marks, L. J., & Skinner, D. (1991). Television Commercial Evaluation in the Context of Program Induced Mood: Congruency versus Consistency Effects. Journal of Advertising, 20, 1-14.

Lerner, J. S., & Keltner, D. (2000). Beyond Valence: Toward a Model of Emotion-specific Influences on Judgment an Choice. Cognition and Emotion, 14, 473-493.

Mitchell, M. M., & Brown, K. M. (2001). The Effects of Anger, Sadness and Happiness on Persuasive Message Processing: A Test of the Negative State Relief Model. Communication Monographs, 68, 347-359.

Raghunathan, R., & Pham, M. T. (1999). All Negative Moods are not Equal: Motivational Influences of Anxiety and Sadness on Decision Making. Organizational Behavior and Human Decision Processes, 79, 56-77.

Reisenzein, R. (1994). Pleasure-Arousal theory and the intensity of emotions. Journal of Personality and Social Psychology, 67, 525-539.

Russell, J. A. and Carroll, J. M. (1999). On the bipolarity of Positive and Negative Affect. Psychological Bulletin, 125, 3-30.

Russell, J. A., Weiss, A., & Mendelsohn, G. A. (1989). Affect Grid: A single Item Scale of Pleasure and Arousal. Journal of Personality and Social Psychology, 37, 493-502.

Sinclair, R. C., Mark, M. M., & Clore, G. L. (1994). Mood-Related Persuasion depends on (Mis)Attributions. Social Cognition, 12, 309-326.

Yi, Y. (1990). Cognitive and Affective Priming Effects of the Context for Print Advertisements. Journal of Advertising, 19, 40-48.

For further information contact : Tineke Faseur, Ghent University, Hoveniersberg 24, 9000 Ghent, Belgium, Tel. 32 9 264 35 25, fax. 32 9 264 42 79, E-mail: Tineke.Faseur@UGent.be

151

7.1.1 SHOULD IT BE ABOUT THEM OR ME? EMOTIONAL UNDERPINNINGS OF THE PERSUASIVENESS OF HEALTH MESSAGES

Nidhi Agarwal, New York University Jennifer Aaker, Stanford University Geeta Menon, New York University

ABSTRACT

Consider an advertisement for the Breast Cancer Research Foundation depicting a picture of spouse and children that states, “Breast cancer doesn’t just affect women…Sadly they won’t be the only casualties. Consider the family and friends who survive them.” Such advertisements focus on the family and ask the reader to think about the consequences of one’s illness on other family members. Consider, in contrast, advertisements that focus on the individual: “Five million Americans have hepatitis. Do you?” Thus, health messages can focus on the consequences of the illness for either of these referent groups. We draw on health-related and social psychological research to explore when and why health messages that focus on self versus family may be differentially effective. We make a case that when messages that are focused on the consequences of an illness for self (vs. family) are compatible with the accessible independent (vs. interdependent) self view, increased perceptions of risk result. More importantly, we examine the emotional processes that might underlie the effectiveness of compatibility between the self view and message focus.

Every individual has at least two types of self view (or self-construal): an independent and an interdependent self view (Markus and Kitayama 1991). An independent self view defines a person in terms of unique attributes that distinguish him or her from others, whereas an interdependent self view defines a person by characteristics that relate him or her to others. Appeals that are compatible with one’s accessible self view tend to be more persuasive (Aaker and Lee 2001). We suggest that when the message referent (in this case, whether the consequences depicted are for oneself or close others) is compatible with this self view, message effectiveness should increase. Therefore, self-focused appeals should be more persuasive for individuals with an independent self view, and family-focused appeals will be more persuasive for individuals with an interdependent self view. Moreover, we suggest that the effectiveness of health messages may relate to their potential to change specific emotions.

Recent research in the area of referencing and framing suggests that framing of messages may elicit specific emotions. The literature on regulatory focus shows that individuals with an accessible independent versus interdependent self tend to experience emotions along different dimensions (Lee, Aaker, and Gardner 2000), namely the happiness-dejection versus the acquiescence-agitation dimensions, respectively (Higgins, Shah, and Friedman 1997). In this research, we contend that when message referent is compatible with the self view, it threatens emotions associated with that self view. Specifically, if a self-focused appeal threatens the chronic emotion of happiness and increases feelings of sadness associated with an independent self view, then processing a self- (vs. family-) focused message should lead to a greater sensitivity to happiness-dejection related emotions (i.e., a decrease in happiness and an increase in sadness). If a family- (vs. self-) focused appeal intensifies the experience of the chronic emotions along the peacefulness-agitation dimension associated with an interdependent self view, it should manifest itself in a greater decrease in peacefulness and a greater increase in agitation.

Markus and Kitayama (1991) have suggested that when individuals experience a particular emotion, they anticipate the social interaction triggered by that emotion. Hence, priming a specific emotional state (e.g., happiness or peacefulness) should make individuals sensitive to appeals that impact that emotion (e.g., self-focused or family-focused). When we directly prime happiness and peacefulness, participants would be more sensitive to appeals that change the primed emotions. Also, since positive emotions seem to facilitate processing and acceptance of relevant negative information (Keller, Lipkus, and Rimer 2003), a compatible appeal should be more persuasive. Specifically, we suggest that for a happy person, a self-(vs. family-) focused appeal should be more persuasive. For a peaceful person, a family- (vs. self-) focused appeal should be more effective.

Negative emotions might also impact the

effectiveness of compatible appeals. Under negative

152

mood, people are driven to improve their mood and they tend to avoid negative information. The compatible (vs. incompatible) appeal is relevant negative information and if processed, more likely to dampen a current mood state. Hence participants feeling negative emotions are more likely to process and be persuaded by an incompatible (vs. compatible) appeal. Hence, inducing the negative emotions of dejection or agitation should trigger a mood maintenance motive and a lower likelihood of elaborating on a relevant negative (i.e., compatible) appeal, making compatible (vs. incompatible) appeals less effective.

The results of an experiment manipulating message referent (self vs. family) and self view support compatibility effects on persuasion and emotions. Changes in specific emotions (happiness, peacefulness, and dejection and agitated), as well as risk estimates, concern about the disease, and intentions to get tested were measured. As expected, when the self view was compatible with the message referent (e.g., independent and self-focus), the message was more effective. Further, when the focus of the appeal (e.g., self-focus) was compatible with the self view (e.g., independent), the health appeal elicited a greater sensitivity to emotions consistent with the self view (happiness-dejection).

In a second experiment, we primed

happiness or peacefulness before exposure to appeals differing in message referents. The results provided corroborative evidence of the compatibility effect and its underlying mechanism. Participants experiencing happiness had higher self-risk estimates when they were exposed to a self- versus family-focused appeal, whereas participants experiencing peacefulness had higher risk estimates when exposed to a family- versus self-focused appeal. In another experiment, consistent with our theorizing about negative emotions, priming dejection or agitation led to a reversal of these compatibility effects.

Thus, our studies identify conditions when compatibility between message referent and self view can increase message effectiveness and the changes in specific emotions associated with self view that drive compatibility. These findings have implications for the literature on emotions as well as health. These findings add to the emotions literature by highlighting the importance of the role of specific emotion types in understanding the mechanisms underlying compatibility effects. By investigating the effects of the consequences of a

disease on self or others, we also extend health-related research. For further information contact: Nidhi Agrawal at the Marketing Department, Stern School of Business, New York University, 44 West 4th Street, New York, New York, 10012. Email: nagrawal@stern.nyu.edu. Phone: 212-998-0523

153

7.1.2 PUBLIC COMMITMENT: A KEY MOTIVATOR IN LONG-TERM CUSTOMER COMPLIANCE

Stephanie Dellande, Chapman University

Prashanth Nyer, Chapman University

ABSTRACT

This paper examines the nature of motivation in gaining customer compliance in compliance dependent services (CDS) - services in which the customer co-creates with the provider while within the service facility, and continues to perform on his/her own once away from the provider to ensure desired outcomes and customer satisfaction (Dellande and Gilly 1998). Examples include education (traditional and online), weight loss programs, debt management programs and retirement savings programs. Dellande (1999) found that among customer attributes, motivation was more important than ability or role clarity in gaining customer compliance in CDS.

In this study we empirically examine the role of public commitment as a motivator of paper conservation in the context of an on-line course. We find that individuals who make a public commitment to a target behavior are more likely to exhibit long-term compliance. Further, we find that subjects making a public commitment to a target behavior are more likely to evaluate the behavior as being important. Individuals who are susceptible to interpersonal influences are more likely to engage in long-term compliance following a public commitment to a target behavior.

Long-term Compliance and Public Commitment

Kiesler (1971) posits that a key factor that determines the magnitude of commitment to a position is the publicness with which the individual declares his/her commitment. Public commitment to an action contributes to an individual’s motivation to act due to anticipated personal and social disapproval and penalties for failure to follow through with promised activities (Parrott et al. 1998). People have a strong desire to appear consistent and rational in the eyes of other people (Tedeschi 1981).

H1: Individuals who make a very public commitment to a target behavior (and to a lesser extent those who make a moderately public commitment), will engage in greater levels of long term compliance with the target behavior compared to

individuals who do not make a commitment.

Public commitment makes ones attitudes more salient thereby providing an anchor for later responses to communication and appeals to action (Halverson and Pallak 1978).

H2: Individuals who make a public commitment to a target behavior will evaluate the target behavior as being more important compared to individuals who do not make a commitment. Further, the effect of public commitment on compliance behavior will be partly mediated by the evaluation of the importance of the target behavior.

Individuals who are high in susceptibility to interpersonal influence (SII) are more willing to conform to the expectations of others (Bearden, Netemeyer and Teel, 1989) and are expected to be more likely to engage in long term compliance following a public commitment compared to individuals who are low in SII.

H3: The effect of public commitment on long term compliance will be moderated by susceptibility to interpersonal influence (SII), such that individuals high in SII will exhibit stronger long term compliance behavior compared to those lower in SII.

Methodology

We investigated the effect of public commitment on paper conservation among 96 students of an online introductory marketing course. The class roster was not made available to the students. The course introduction included a brief essay on the importance of minimizing the wastage of paper. Students were randomly assigned to three conditions.

Subjects in the ‘low public commitment’ condition were sent a personalized email message asking the student to commit to reducing the use of paper by reading course lectures online. In addition to the

154

above, subjects in the ‘high public commitment’ condition were also informed that the instructor would acknowledge the students who had agreed to conserve paper by listing their names in an email to be sent to the whole class. Subjects in the high public commitment condition were made to believe that their commitment to reading lectures online was made known to many others in the class. Subjects in the ‘no commitment’ condition were not asked to make a commitment to conserving paper.

A few days after the manipulation of the various commitment conditions, all subjects were sent an email message containing an online survey which included measures of importance of paper conservation and susceptibility to interpersonal influence (adapted from Bearden, Netemeyer and Teel, 1989).

The amount of time spent by each student viewing the on-line lectures was monitored, and this measure of time spent online was used as a proxy for compliance with the target behavior. Thus students who read the online lectures online would spend more time viewing the online lectures, while students who printed the lectures would spend less time viewing the online lectures.

Analysis

A simple analysis of variance indicated that the level of commitment had a significant impact on TIME, the total time spent online on the lecture material. Subjects in the high public commitment condition spent significantly more time (385 minutes) than subjects in the low public commitment condition (335 minutes) or those in the no-commitment condition (205 minutes).

Subjects’ evaluation of the importance of paper conservation (IMP) was expected to mediate the effect of the experimental manipulation of public commitment on the time spent online (TIME). To verify this, first a simple MANOVA was performed with TIME and IMP as dependent variables and the commitment conditions as the factor. The public commitment conditions had a significant effect on TIME and IMP. Next an ANCOVA was performed with IMP as a covariate, while TIME continued as the dependent variable. IMP had a significant effect on TIME, and the introduction of IMP as a covariate led to a reduction in the impact of the commitment manipulation on TIME, thus supporting the second hypothesis.

Regression analyses showed that as hypothesized susceptibility to interpersonal influence (SII) had a

significant effect on compliance behavior (TIME) under conditions of high and low public commitment , but not under the no-commitment condition, thus providing support for the third hypothesis.

References

Bearden, William O., Richard G. Netemeyer and Jesse E. Teel (1989), “Measurement of Consumer Susceptibility to Interpersonal Influence,” Journal of Consumer Research, 15 (March), 473-481.

Dellande, Stephanie (1999), “Gaining Customer Compliance in Services,” Dissertation. University of California at Irvine.

_________ and Mary C. Gilly (1998), “Gaining Customer Compliance in Services,” Advances in Services Marketing and Management, 7, 265-292.

Halverson, Richard R. and Michael S. Pallak (1978), “Commitment, Ego-Involvement, and Resistance to Attack,” Journal of Experimental Social Psychology, 14, 1-12.

Kiesler, Charles A. (1971), The Psychology of Commitment: Experiments Linking Behavior to Belief, San Diego, CA: Academic Press.

Parrott, Roxanne, Jennifer Monahan, Stuart Ainsworth, and Carol Steiner (1998), “Communicating to Farmers About Skin Cancer,” Human Communication Research, 24 (March), 386-409.

Rorer, Brett, Carolyn M. Tucker, and Hattie Blake (1988), "Long-Term Nurse-Patient Interactions: Factors in Patient Compliance or Noncompliance to the Dietary Regimen," Health Psychology, 7, 35-46.

Tedeschi, James (1981), Impression Management: Theory and Social Psychological Research, New York, NY, Academic Press.

For further information contact:

Stephanie Dellande, Ph.D., Assistant Professor of Marketing, The George L. Argyros School of Business & Economics, Chapman University, One University Drive, Orange, California 92866 , Phone: (714) 628-7347

155

7.1.3 NURTURANCE, RELIGIOSITY AND APPEARANCES: THE EFFECT OF PERSONALITY TRAITS ON RESPONSES TO CORPORATE PHILANTHROPY

Debra Z. Basil, University of Lethbridge

Michael D. Basil, University of Lethbridge Deanne Weber, Porter Novelli International

ABSTRACT

Previous research has attempted to identify “altruistic” personality traits relating to interpersonal acts. Initially this research stream focused primarily on interpersonal helping behavior (e.g. Turner, 1948), but later research expanded inquiry into the realm of charitable donations (e.g. Harvey, 1990). The present research seeks to continue the progression of this stream of research, moving forward to an examination of the next generation of philanthropy. Specifically, this research seeks to identify consumer personality types that are supportive of corporate-based philanthropy.

A variety of personality traits have been correlated with traditional forms of altruism (see Krebs, 1970), however three general categories of traits seem to emerge as most predictive of altruistic behavior. These are: the nurturing individual, the religious/spiritual individual, and the individual concerned with appearances. The present research seeks to determine whether these personality traits will influence individuals’ support for corporate philanthropy as well.

Although similar behaviours supporting corporate philanthropy are expected from individuals possessing high levels of these personality traits, the underlying motivation is expected to differ. Only the nurturing and religious personality trait individuals are expected to hold normative expectations of corporate philanthropy. In other words, nurturing and religious personality types are expected to place greater importance on corporate philanthropy, whereas the individual concerned with appearances is not necessarily expected to hold a similar view.

This research used secondary data from Porter Novelli’s “ConsumerStyles 2002” survey in order to assess the impact of personality traits on response to and perceptions of corporate philanthropy. The data were collected by Synovate, Inc. A stratified random sample was selected from a mail panel containing 550,000 potential respondents. A total of 10,000 surveys were mailed, and 6,065 responses were received, for a response rate of 61%. Each respondent received a

small gift for participation, and was entered into a drawing for $1,000.

The independent variables of interest sought personality trait information, based on responses to personality adjectives. Five adjectives were identified as nurturing tendencies (caring, down-to-earth, generous, loyal, and family oriented). Three adjectives were identified as concern with appearances (stylish, trendsetter, sophisticated). Finally, two adjectives were identified as representing religiosity (religious, spiritual). Demographic information regarding respondents’ age and gender was also collected and used for control purposes in the analyses.

Two dependent variables were of interest for the present research. These variables assessed personal behavior relating to corporate philanthropy and views of the importance of corporate philanthropy. Personal behavior was evaluated through responses to the statement “When I see a company doing charitable work, it makes me more likely to try their products.” The importance of corporate philanthropy was evaluated by having respondents indicate their perception of how important it is for public companies to participate in five different charitable behaviors, including monetary donations to local charities, giving products to those in need, helping employees and their families, protecting the environment, and participating in community projects.

This research demonstrates that personality disposition is an effective predictor of response to corporate philanthropy. A nurturing disposition, a religious disposition, and concern with appearances all lead to more support of companies that perform philanthropic behavior. These results parallel findings in previous literature for interpersonal helping behavior and charitable donations, suggesting that this relatively new form of charity, support for corporate philanthropy, elicits consumer responses similar to more traditional forms of charity.

Individuals appear to map their own behavior toward corporate philanthropy in a manner similar to their responses to other forms of charity. However, perceptions of the importance of corporate philanthropy differ somewhat from

156

individuals' reports of their own responses to corporate philanthropy. More specifically, some individuals perform altruistic behavior because they are concerned with appearances, rather than out of a pure desire to help others. These individuals report that they respond positively to corporate philanthropy in their own personal behavior, but they do not feel that corporate philanthropy is particularly important. This may be because supporting corporate philanthropy is simply a means by which they can maintain appearances. The helping aspect itself becomes secondary and is of little concern.

The present findings hold important implications for managers. Situational factors will always be vitally important in determining consumer response to corporate philanthropy, but the present results suggest that individuals' dispositions can play a role as well. Perhaps creating synergy between dispositional and situational factors will prove to be an effective means of leveraging value for corporate philanthropy. Previous research has demonstrated that reminding individuals of traits they possess makes the impact of those traits stronger in a given situation (Kraut, 1973). Similarly, reminding individuals of their nurturing or religious dispositions may lead to stronger responses toward corporate philanthropy. Future research examining actual, rather than self-reported, behavior would be a valuable extension to this line of inquiry.

REFERENCES Harvey, J. (1990). Benefit segmentation for

fundraisers. Journal of the Academy of Marketing Science, 18(1), 77-86.

Kraut, R. E. (1973). Effects of social labeling on giving to charity. Journal of Experimental Social Psychology, 9, 551-562.

Krebs, D. L. (1970). Altruism - An examination of the concept and a review of the literature. Psychological Bulletin, 73(4), 258-302.

Turner, W. D. (1948). Altruism and its measurement in children. Journal of Abnormal and Social Psychology, 43, 502-516.

For more information contact: Dr. Debra Basil, Faculty of Management University of Lethbridge, 4401 University Drive, Lethbridge, AB TIK 3M4 Canada Phone: (403)329-2164; Fax (403)329-2038 Email: debra.basil@uleth.ca

157

7.2.1 RENEGOTIATING WHO I AM: THE EFFECTS OF CHANGING CULTURAL IDEOLOGIES

AND INTERPERSONAL INFLUENCE ONTHE SELF

Camelia Micu, University of Connecticut Robin A. Coulter, University of Connecticut

Linda L. Price, University of Nebraska

ABSTRACT

There is a rich history of research concerning self, self-concept, multiple selves, selves in transition, and the social nature of the self (Arnould & Price, 2000; Byrne, 2002; Folkes & Keisler, 1991; Geertz, 1986; Kleine, Kleine, & Kernan, 1993; Markus, 1977; Markus & Narius, 1986; Markus & Wurf, 1987; McAlexander & Schouten, 1989; Shouten, 1991). Additional research has examined and acknowledged the role of culture in defining the self and recently, Arnett (2002, p. 774) argued that globalization has its “primary psychological influence on issues of identity.” He introduces and defines two constructs particularly relevant to context of transitioning economies: bicultural identity and identity confusion. Arnett describes bicultural identity such that “in addition to their local identity, young people develop a global identity that gives them a sense of belonging to a worldwide culture, and includes an awareness of the events, practices, styles, and information that are a part of the global culture.” He suggests some face identity confusion, “for some people, adapting to the rapid changes taking place in their culture is more difficult. The image, values and opportunities they perceive as being part of the global culture undermine their belief in the value of local cultural practices.”

In this paper we consider bicultural

identity and identify confusion in the context of Romanian women’s perspectives on and consumption of appearance-related products (i.e., cosmetics, jewelry, and clothing) as the country transitions from a communist system with a centrally planned economy to a free market economy. Our focus was to examine the complexity of self-identity and renegotiation in the context of changing cultural (communist vs. consumerist) ideologies. We conducted depth interviews with 28 Romanian women whose ages ranged from 19 to 54 and who varied on socioeconomic status.

Romania is an interesting research venue

because under communism and consistent with Western feminist discourse, cosmetics were

depicted as oppositional to gender equality. Looking feminine or sexy was discouraged by social norms. Within the past decade, Romania and other Central European countries have witnessed a dramatic infusion of cosmetic products and brands from multi-nationals such as Avon, P&G, and L’Oreal, increased exposure to commercial, persuasive media, and many more shopping alternatives (“Ready to Shop Until They Drop,” 1998; Euromonitor, 2000). Concurrently, “beauty” became a subject of wide debate: “A rapid change […] is the imagery for sexualized femininity […] apparent in magazine and job advertisements throughout the region. Employers seek women who are properly groomed, with made-up faces, stylish clothes, and shaved legs” (Gal & Kligman, 2000, p. 112). Many acknowledge that in post-socialist Central Europe, a woman’s feminine appearance is becoming increasingly important (Coulter, Price, & Feick, 2003; Gal & Kligman, 2000).

Our data provide evidence of both

bicultural identity and identity confusion, and suggest that these two constructs are sometimes inextricably linked. Collectively, our informants report on the general societal pressure for Romanian women to be more attentive to their appearance – a pressure that they associate with globalization, i.e., the availability of an unprecedented number of appearance-related (clothing and cosmetics) products and brands, as well as Western media with beautiful models promoting these goods. Thus among our informants, women’s global identity is grounded in looking feminine and sexy, and many are in the process of renegotiating their self-identity. In general, younger women express a desire to belong to a worldwide culture, and are more likely to embrace a global identity. Older informants, who lived through communist rule, and experienced the expectation of being a breadwinner, the homemaker and the childbearer, seem more perplexed with assimilating a global identity associated with personal appearance. Many struggle with developing a balance between global expectations and local beliefs, attempting to balance aspects in a hybrid identity.

158

We observe identity confusion in both

older and younger informants, although it does seem more pronounced among older women. Our data suggest that identity confusion is often exacerbated by women’s relations with influential others. For example, several of our informants’ reported that they needed to attend to their appearance to either maintain or get a job, but that their husbands/boyfriends reject their cosmetics use. These women struggle to not bother their significant others and not compromise their jobs. Thus, confusion comes from the idea that wearing cosmetics to work is consistent with their roles as employees in a global world, but clashes with their local roles as wives/girlfriends. Some of our young informants are the daughters of men who hold to a communist ideology, and do not want their daughters wearing make-up and sexy clothing. Although many of these young women are focused on establishing a global identity, they also experience identity confusion as they attempt to break free of their father’s communist appearance-related ideologies. Our data also suggest that identity confusion is precipitated by interactions with co-workers. As co-workers rethink their ideological stance and interest in taking on a global identity, they seek information about appearance-related products from one another, but at the same time have to deal the “competition” of focusing on appearance and achieving balance in their personal identity. Additionally, our younger informants discussed how an increased attention to appearance (in the context of taking on a more global identity) has altered their roles and relationships with their mothers. Our younger informants often buy cosmetics for their mothers who are more comfortable with a local identity and not interested in wearing makeup. Daughters endeavor to convince their mothers of the health and personal (youthfulness) benefits of using cosmetics, only to witness them struggling with this local/global paradox of female identity.

In sum, we have considered bicultural identity and identity confusion in the context of Romanian women, particularly with regard to their appearance. We see that one’s definition of self, and therefore necessarily one’s redefinition of self is contingent upon personal perspectives that are shaped by cultural and social forces, and only by examining consumers’ social context and important relationships can we hope to understand identity dynamic

159

7.2.2 PERSONAL VALUES FOR CUSTOMER SERVICE

Nicole Hafey, Central Queensland University Tony Ward, Central Queensland University

ABSTRACT

This classification paper identifies personal values that have a direct impact on customer service performance of customer service providers in a retail context. The benefits of the research are to provide customer service managers with a set of personal values that should be considered during recruitment and for training/retraining purposes. INTRODUCTION

Much has been written about the impact of personal values in business contexts, in particular in searching for predictors of job performance (Nathanson & Becker 1973; Barrick & Mount 1991; Tett, Jackson & Rothstein 1991). In marketing, the role and importance of the personal values of consumers has been widely recognised (Vinson, Munson & Nakanishi 1977). ‘Philosophers have known about the importance of values at least since the ancient Greeks, and social scientists have known about how important values are for at least 50 years’, (Kahle & Kennedy 1989, p. 5). Most of the research based on personal values can be sourced back to the seminal work of Rokeach (1979), and more recently, Schwartz (1992). This search for personal values relevant to marketing personnel is complicated by the large number of such values found in the literature, (Gordon in Prein & Botwin 1966; Rokeach 1973; Super 1973; Vinson, Scott & Lamont 1977; Kramer 1984; Prakash 1984; Munson 1984; Kahle & Kennedy 1989; Crosby, Bitner & Gill 1990; Brathwaite 1994; Feather 1994; Schwartz & Bilsky 1997; Super & Nevitt in Aiken 1999; Johansson & Webber in Aiken 1999; Schwartz & Sagie 2000), some 213 being found by the authors of this paper. There are many divergent tasks required of marketers, ranging from selling to customer service, to promotions, to research and planning, and each of these areas require people with varying characteristics and skills. For this research the authors chose to focus on values that would directly influence the performance of customer service providers, a segment of great importance as they generally form the main interface between an organization and its customers. This research therefore sought to identify all of the values in the literature, eliminate equivalents and duplications,

retain only those positive values that were directly applicable to customer service providers in a retail context, and then to classify them. Thus, the focus of this research was to identify the most relevant positive personal values that would directly influence the performance of retail customer service providers, and to classify them in a meaningful manner. The benefits of undertaking this research are: - by identifying the values found to be relevant to the performance of customer service providers, a focus for future research can be identified; - a separate review of the relationship marketing and personality literature revealed that there were no empirical investigations examining the effects of personal values of individuals in a relationship marketing situation which is so important in a customer service environment (Ward & Acutt 2000); - a taxonomy of personal values relevant to the performance of customer service providers would aid the focusing of future research; - it is anticipated that the identification of relevant personal values would facilitate the recruitment and retraining of customer service staff. Definitions

The term ‘personal values’ is used in this research to indicate personality types, traits and values of individuals, and is defined as, ‘identifiable characteristics of a person which indicates a preference to perceive the world in a certain manner’. There are three elements of this definition that are of particular importance: a) identifiable characteristics – in order to be of any use in marketing research, personal values must be identifiable. It is acknowledged that there may be values that we cannot identify, and such (hypothetical) values are expressly excluded from this research; b) a preference to– individuals do not perceive the world in a consistent manner, but vary their behaviour and/or attitudes according to mood, often with variations on very short time frames. There are no absolutes in personal perception, and consequent behaviour, and we all exhibit inconsistencies in this regard. This definition acknowledges such inconsistency and indicates that the presence or

160

otherwise of a particular value is only an indicator of how a person will tend to perceive the world in general, while acknowledging there will often be variations; c) perceive the world – our personal values act primarily to influence our attitudes and the way we perceive the world. These attitudes and perceptions in turn direct much of our behaviour, and it is this combination of attitudes and perceptions and behaviour that affect how we are perceived by others, and for this research specifically how customers perceive the performance of customer service providers. A service providers personal values are thus an integral part of their attitudes and perceptions and will have a significant influence over how they behave, and in turn how they are perceived by customers. Personal values As previously mentioned, research has identified that values both directly and indirectly impact upon consumer buying behaviour. A number of studies in the 1970’s, researched value systems and their influence on consumption-related behaviour (Henry 1976; Vinson 1976), marketing segmentation (Vinson & Munson 1976), perceptions of product attribute importance (Scott & Lamont 1974) and brand awareness and preferences (Vinson, Munson and Nakanishi 1977). Further studies have identified the role values play in relation to consumer behaviour constructs such as brand attitudes, purchase behaviour, patronage behaviour and media usage (Sherrell, Hair & Bush 1984). Kahle and Kennedy (1989) identified that consumer behaviour based research has confirmed that ‘values directly influence interests, time-use activities, and roles which in turn influence consumer behaviours’ (p. 6). Furthermore, they highlight the importance of values within the realm of marketing research by stating, ‘the function of marketing is to help consumers fulfil their values’ (p. 6). Similar findings have been identified in the relationship marketing, sales and customer service literature. Moreover, most of the research which has been conducted on personal values in consumer behaviour settings (Vinson, Munson and Nakanishi 1977; Novak & MacEvoy 1990; Rogers, Ross & Williams 1992) has been grounded in the Rokeach value system. Rokeach’s seminal works (1968, 1973), identified 36 basic human values (Brathwaite 1994) which were divided into two dimensions, ‘terminal values’ and ‘instrumental values’ (Rokeach 1973). Terminal values were defined as ‘preferred end states’ (Rogers, Ross &

Williams 1992, p. 82). Alternatively, instrumental values were defined as ‘preferred modes of conduct’ (Rogers, Ross & Williams 1992, p. 82). Rokeach went on to further sub-divide terminal values into personal values (for example, end states such as salvation) and social values (for example, end states such as world peace), instrumental values were also further sub-divided into moral values (values associated with interpersonal modes of conduct) and competence values (values associated with intrapersonal modes of conduct) (Aiken 1999). It has been acknowledged that instrumental values are reflected in human behaviour and that more global factors such as culture and social are attributed to the development of terminal values (Kramer 1984). Kramer further stated that ‘a person’s behaviour, knowledge, intellectual and temperamental characteristics and moral stability are expressed on the instrumental value level’ (p. 244). McNeese & Crook (2003) recognised that ‘individuals or groups are identified by the values they adopt and even more by the values that are evident in their daily activities, choices and decisions’ (p. 261). Pitts & Woodside (1984) also commented on Rokeach’s contention that values are standards for either developing or maintaining attitudes toward objects or situations (p. 55). These statements identify how values (and more specifically, instrumental values) make a significant contribution to individuals behaviour, and in this instance, how they behave, or regulate their behaviour (Vinson, Scott & Lamont 1977). As we are looking towards identifying a personal values based profile of customer service providers in a retail setting, the emphasis needs to be on the classification of instrumental values which are more directly related to behaviour in a marketing situation. METHOD

This section commences with a brief review of the research philosophy that guided the study followed by a description of the classification technique. This research is essentially exploratory in nature as insufficient information was available about personal values in a customer service context to engage in an explanatory study. Classification research aims to ‘categorize research units into groups, demonstrate differences, and to clarify social events and relationships’ (adapted from Sarantakos 1996, p. 7). For qualitative research the logic of theory building is usually inductive, and moves from reality through data generation, and on

161

to verification (Sarantakos). In this research the final step of verification was not addressed. The researchers used a reflective thinking technique (Cooper & Emory 1995) where, having identified a gap in the literature that needed to be addressed, they proceeded with an analysis of observable constructs (Sarantakos). Comprehensive taxonomies and the majority of literature on taxonomic theory have been developed for the sciences, and in particular for the natural sciences (Pankhurst 1978). Although there was a significant quantity of literature relating to taxonomy and classification (for example, Hunt 1983; Bowen 1986), no work was found which specifically addressed a comprehensive review of the theory of taxonomy in a marketing context, even though the use of the technique is well established within the marketing discipline (for example, Mills & Margulies 1980; Lovelock 1983). A review of the methods of classification was undertaken, which showed that arrangement procedures could be divided into two major categories: hierarchical and non-hierarchical (Mezzich & Solomon 1980). Non-hierarchical methods produce configurations that do not present rankings in which lower ranking taxa become members of larger more inclusive taxons, while hierarchical methods produce tree-like taxonomic systems, comprising taxa arranged into a number of ranks. A detailed review of the personal values relevant to customer service showed that some hierarchical structure was immediately evident and thus was chosen. The focus in this research was to describe personal values as they are at present, thus a phenetic basis for establishing relationships between entities, and between taxa, was adopted (Sneath and Sokal 1973). The method of determining the similarity or dissimilarity of personal values was then addressed. The basis of this research was to identify: a) relevant personal values; b) both similarities and dissimilarities of values from a number of sources and their potential impact in a consumer service situations. Thus, single dissimilarities that were judged to fundamentally affect customer service were classified as different taxa, an approach termed monothetic (Sneath & Sokal 1973). In summary, the classification of personal values for customer service used traditional taxonomic methods to formulate a hierarchical structure, using a phenetic, monothetic technique to identify dissimilarities in order to place personal values into appropriate taxa. This method has satisfactory validity, as:

- content validity is provided by the clear nature of differences identified, and no specific difficulties were encountered in this regard; - internal validity is provided by consistency of classification, following the method outlined above and the application of strict classification ‘rules’; - external validity is provided by the extensiveness of the literature search for values; and, - construct validity is provided as similarities and dissimilarities were clearly identified.

The classification was undertaken by two researchers, both with extensive experience in the customer service literature, and as practitioners.

RESULTS

Using the general techniques identified above, four steps were identified to systematically address the task, namely, 1) list all values from the literature, 2) a rationalization process to identify and combine similar values to reduce the number of duplications, 3) identify and remove those values that did not directly impact on customer service, and 4) to construct a taxonomy of the remaining values. Step 1 – List of values. The first task was to undertake an extensive review of relevant literature to search for values previously identified by many authors. General discipline areas searched included psychology, social psychology, marketing psychology, cross-cultural studies and consumer behaviour. This total listing of values is much too extensive to include here. Step 2 – Rationalization. This step was the most difficult as it sought to address four attributes of the initial list. Firstly, a number of values that had either the same or very similar characteristics had been given different names by the originating authors, for example, ‘benevolence’ (Prein & Botwin 1966; Schwartz & Sagiv 1995) and ‘concern for the welfare of others’ (Rokeach 1979; Brathwaite 1994). These situations were addressed by assigning these values with names that were widely used and accepted, for example, ‘benevolence’. Secondly, a characteristic of the listing is that not every value will be applicable in every customer service situation, a necessary feature of a comprehensive evaluation. A third attribute of values is that they are a mixture of attitudes/perceptions (for example, ambition), and consequent behaviour(s) (to act ambitiously, or to be seen to be ambitious). Thus, the underlying attitude/perception manifests itself in a person’s behaviour. This ‘divide’ is seen in many of the tests of personal values (for example Gordon 1961 as

162

referenced in Prein & Botwin 1966), which are comprised of a mixture of attitudinal/perceptual and behavioural elements. In a customer service context there are many values that manifest themselves in behaviours that are most relevant to customer service performance (for example, understanding a customers perspective), while there are other values that also appear to have a direct impact on customer service performance, but which would not normally manifest themselves in a behavioural form (for example, fairness and self-respect). Finally, as previously discussed, the Rokeach classification divides values into instrumental and terminal values. For the purpose of customer service performance, only the instrumental values are applicable, as it is these values that manifest themselves as behaviours that impact on customers (rather than on oneself) that are applicable. Thus, all terminal values were removed. This rationalization was undertaken using a reiterative process of continuous refinement using a number of colleagues familiar with the ‘values’ literature. The output from this step was the set of 115 values shown at Table 1. Insert Table 1 Step 3 – Removal of non applicable values. This third step required the refinement of the list of values in table 1 in order to remove ‘non-applicable’ values in a customer service context. This process was informed by extensive experience in, and research of, the role of values in customer service performance by the authors, together with the extensive review of customer service literature. Particular attention was paid to the range of customer service tasks across many industries, including different forms of contact, such as face-to-face and telephone. It is acknowledged that a case can be made that every component of an individuals persona is applicable in any interpersonal encounter, as we are all affected (even if only in the smallest manner) by such fundamental elements of our being. The criteria used was thus to eliminate values where no apparent direct influence on customer service performance was evident. In addition, each of the remaining values was inspected to ensure that they had direct applicability to a customer service encounter. This process thus provided an internal validity check on the remaining values, that is, no case could be found for rejection, while a case could be found for retention. In addition, a number of values that would have a negative effect, such as Power and Selfishness, were also deleted. These ‘negative’ values were omitted to avoid duplication as they had a positive opposite that was already included,

for example, the opposite of selfish in embodied in friendly, caring, benevolent, loyal and working for the welfare of others. As shown in table 2, 54 values remained in the list as having some positive direct effect on customer service performance. Insert Table 2

Step 4 – Classification of values for customer service providers. Many authors have provided hierarchal structures to their values schemes, so many highly ranked terms that had both similarities and dissimilarities with different names and lower order classifications were identified. The situation with highly ranked terms was the most difficult to address as fundamental decisions had to be made regarding the overall hierarchal structure of the values ‘tree’. A basic decision was made to adopt the basic structure of Rokeach’s instrumental values, making a number of adaptations to suit the particular requirements of this research. It was also decided to restrict the ranks to just two, as a more quantitative methodology would be required to provide additional ranks with validity. In this classification some minor changes were made to terminology in order to provide a consistent perspective, especially to first ranked values. This process was further supported by an apperception study in which 100 respondents were asked to provide antecedents and insights of their perception of values. The result of this classification process is shown at table 3, where 14 first ranked values are identified, with a total of 40 second ranked values. Insert Table 3

DISCUSSION

It is anticipated that the taxonomy shown in table 3 will provide a source of personal values that customer service managers can use for recruitment and training/retraining of customer service staff. The provision of such a structure should also be of use in directing the evaluation of customer service performance by marketers. Empirical testing in a number of customer service situations should lead to further refinement for specific applications. Such testing should focus on further reducing the number of personal values and assigning weightings to each value. Until shown otherwise, it should be assumed that pertinent values and their relative weightings will vary across applications. REFERENCE LIST

Aiken, L (1999) Personality Assessment: Methods & Practices, 3rd ed. Hogrefe & Huber Publishers, Seattle.

163

Barrick, M & Mount, M (1991) ‘The Big Five Personality Dimensions and Job Performance: A Meta-Analysis’. Personnel Psychology, 44, 1-26.

Bowen, J (1986) Development of a Taxonomy of Services to gain Strategic Marketing Insights, unpublished dissertation, Texas A & M University.Brathwaite, V (1994) Beyond Rokeach’s Equality-Freedom Model: Two Dimensional Values in a One-Dimensional World. Journal of Social Issues, 50, 4, 67-94.

Cooper, D & Emory, C (1995) ‘Business Research Methods, 5th edn, Irwin, Chicago.

Crosby, L, Bitner, M & Gill, J (1990) Organisational structure of values. Journal of Business Research, 20, 123-134.

Feather, N (1994) Human Values and Their Relation to Justice. Journal of Social Issues, 50, 4, 129-151.

Henry, W (1976) ‘Cultural Values do Correlate with Consumer Behaviour’, Journal of Marketing Research, 13, 121-127.

Hunt, S, (1983) Marketing Theory, Irwin, Homewood, Illinois.

Kahle, L & Kennedy, P (1989) ‘Using the List of Values (LOV) to Understand Consumers’. The Journal of Consumer Marketing, 6, 3, 5-12.

Kramer, H. E., (1984) ‘The Value of Higher Education and it’s impact on Value Formation’. In Pitts, R E & Woodside, A G, Personal Values and Consumer Psychology, Lexington Books, Toronto.

Lovelock, C (1983) 'Classifying services to gain strategic marketing insights,' Journal of Marketing, 47, 3, Summer, 9-20.

McNeese-Smith, D & Crook, M (2003) ‘Nursing Values and a Changing Workforce’. JONA, 33, 5, 260-270.

Mezzich, J & Solomon, H (1980) Taxonomy and Behavioral Science, Academic Press, London.

Mills, P & Margulies, N (1980) ‘Toward a core typology of service organizations’, Academy of Management Review, 5, 2, 255-65.

Munson, J (1984) Personal Values: Consideration on Their Measurement and Application to Five Areas of Research Enquiry. In Pitts, R. E. & Woodside, A. G. Personal Values and Consumer Psychology, Lexington Books, Toronto.

Nathanson, C & Becker, M (1973) ‘Job Satisfaction and Job Performance: An Empirical Test of some Theoretical Propositions’, Organisational Behaviour and Human Performance, 9, 2, 246.

Novak, T & MacEvoy, B (1990) ‘On Comparing Alternative Segmentation Schemes: The List of Values (LOV) and Values and Lifestyles

(VALS)’. Journal of Consumer Research, 17, 1, 105-109.

Pankhurst, R (1978) Biological Identification, Edward Arnold, Bristol.

Pitts, R & Woodside, A (1984) ‘Personal Values and Market Segmentation: Applying the Value Construct’. In Pitts, R E & Woodside, A G, Personal Values and Consumer Psychology, Lexington Books, Toronto.

Prien, E & Botwin, D (1966) ‘The Reliability and Correlates of an Achievement Index. Educational and Psychological Measurement, 26, 1047-1052.

Prakash, V (1984) Personal Values and Product Expectations. In Pitts, R E & Woodside, A G, Personal Values and Consumer Psychology, Lexington Books, Toronto.

Rodgers, J, Ross, S & Williams, T (1992) ‘Personal Values and Purchase Dissatisfaction Response’. Journal of Consumer Satisfaction, Dissatisfaction and Complaining Behaviour, 5, 81-92.

Rokeach, M (1968) Beliefs, Attitudes and Values. Jossey Bass, San Francisco.

Rokeach, M (1973) The Nature of Human Values. The Free Press, New York.

Rokeach, M (1979) Understanding Human Values, Individual and Societal. MacMillian Publishers, New York.

Sarantakos, S (1996) ‘Social Research’, MacMillian, Melbourne.

Schwartz, S (1992) ‘Universals in the content and structure of values: Theoretical advances and empirical tests in 20 countries’. Advances in Experimental Social Psychology, 25, 1-65

Schwartz, S. H. & Bilsky, W (1997) 'Toward a Universal Psychological Structure of Human Values'. Journal of Personality and Social Psychology, 53, 3, 550-562.

Schwartz, S & Sagie, G (2000) Value Consensus and Importance, A Cross-National Study. Journal of Cross-Cultural Psychology, 31, 4, July 2000, 465-497.

Schwartz, S. H. & Sagiv, L (1995) Identifying Culture-Specifics in the Content and Structure of Values. Journal of Cross-Cultural Psychology. 26, 1 95-116.

Scott, J & Lamont, L (1974) ‘Relating Consumer Values to Consumer Behaviour: A Model and Method for Investigation’, In Greer, T (ed), Increasing Marketing Productivity, American Marketing Association, Chicago.

Sherrell, D, Hair, J & Bush, R (1984) ‘The Influence of Personal Values on Measures of Advertising Effectiveness: Interactions with Audience Involvement’. In Pitts, R E &

164

Woodside, A G, Personal Values and Consumer Psychology, Lexington Books, Toronto.

Sneath, P & Sokal, R (1973) Numerical Taxonomy, Freeman, San Francisco.

Super, D (1973) ‘The Work Values Inventory’. In Zytowski (ed.) Contempory Approaches to Interest Measurement. University of Minnesota Press, Minneapolis.

Tett, R, Jackson, D & Rothstein, M (1991) ‘Personality Measures as Predictors of Job Performance: An Analytic Review’, Personnel Psychology, 44, 703-742.

Vinson, D (1976) ‘Human Values and the Marketing Function’. In Ericson, R (ed), Avoiding Social Catastrophies and Maximising Social Opportunities: The General Systems Challenge. Society for General Systems Research, Washington.

Vinson, D & Munson, J (1976) ‘Personal Values: An Approach to Market Segmentation’. In Bernhardt, K (ed), Marketing: 1776-1976 and Beyond, 3rd ed. American Marketing Association, Chicage.

Vinson, D, Munson, J & Nakanishi, M (1977) ‘An Investigation of the Rokeach Value Survey for Consumer Research Applications’, Advances in Consumer Research, 4, 247-252.

Vinson, D, Scott, J. & Lamont, L. (1977) The Role of Personal Values in Marketing and Consumer Behaviour. Journal of Marketing, 41, April 1977, 44-50.

Ward, T. & Acutt, B. (2000) Intrinsic personal characteristics of service providers, Central Queensland University, Rockhampton.

165

TABLE 1

Primary list of personal values (alphabetical order)

Ability utilization Acting autonomously Advancement Affectionate Altruism Ambitious Aspiring Authoritarian Being well respected Belonging Benevolent Broadminded Capable Caring Cautiousness Cheerful Choosing own goals Clean Competent Conforming Conscientious Considerate Consistent Courageous Courteous Creative Curious Daring Dependable Desire to lead Dutiful Effectiveness Enthusiastic Ethical Excitement Experienced Fairness Forgiving Friendly

Fun Generous Goodness Hard working Helpful Honest Hospitable Humorous Idealism Imaginative Impartial Independent Influential Informal Intellectual Intelligent Interested Jovial Joyful Knowledgeable Light-hearted Likeable Logical Loving Loyal Morality Neat Obedient Open-minded Original thought Personal growth Personal orientation Personal relations Polite Power Preserving public image Prestige Prompt Pro social Prudent

Pursuit of knowledge Realistic Recognised by community Recognition Refined Reflective Reliable Respectful Responsible Restrained Risk Self-controlled Self-development Self-direction Self-disciplined Self-improvement Selfish Self-reliant Self-sufficient Sensitive Sincere Social interaction Standing up for beliefs Strict Supportive Tender Tidy Tolerant Truthful Understanding Variety seeking Virtuousness Warm relationships Well-mannered Willing to pardon others Witty Working for the welfare of others

166

TABLE 2

Personal Values of Customer Service Providers (alphabetical order)

Ability utilization Belonging Benevolent Capable Caring Cheerful Clean Competent Conforming Conscientious Considerate Consistent Courteous Dependable Dutiful Effectiveness Enthusiastic Ethical

Forgiving Friendly Helpful Honest Hospitable Independent Interested Light-hearted Likeable Logical Loyal Neat Obedient Personal relations Polite Prompt Pro social Reliable

Respectful Responsible Restrained Self-controlled Self-disciplined Self-reliant Self-sufficient Sensitive Sincere Social interaction Supportive Tidy Tolerant Truthful Understanding Warm relationships Well-mannered Working for the welfare of others

167

TABLE 3 Hierarchy of Customer Service Provider Personal Values

First rank values Second rank values

Benevolent

Caring, Considerate, Friendly, Hospitable, Personal relations, Pro social, Social interaction, Warm relationships, Sensitive

Capable

Ability utilization, Competent, Effectiveness, Intelligent

Cheerful Likeable, Light-hearted

Clean Neat, Tidy

Enthusiastic Interested

Forgiving Tolerant, Understanding

Helpful

Supportive, Working for the welfare of others

Honest Ethical, Sincere, Truthful

Independent Self-reliant, Self-sufficient

Logical Consistent

Obedient Conforming, Dutiful, Loyal, Respectful

Polite Courteous, Well-mannered

Responsible

Conscientious, Dependable, Prompt, Reliable

Self-controlled Restrained, Self-disciplined

168

7.2.3 PSYCHOLOGICAL ANTECEDENTS OF IMPULSIVE AND COMPULSIVE BUYING: A HIERARCHICAL PERSPECTIVE

Tao Sun, Emerson College

Guohua Mark Wu, San Jose State University Seounmi Youn, Emerson College

ABSTRACT

This study developed a structural equation

model that investigates the hierarchical relationships among Big Five factors, impulsive buying and compulsive buying. It identified positive relationships between impulsive buying and compulsive buying, between emotional instability and the two traits, and between impulsive buying and two other personality traits (openness and extraversion).

Introduction

Quality research has been conducted in the

areas of impulsive buying (e.g., Rook & Fisher, 1995) and compulsive buying (e.g., O’Guinn & Faber, 1989) separately. Arguments and propositions have been made on the relationship between the two purchase-related traits. Although Mowen (2000) found that impulsiveness (not impulsive buying) was a strong predictor of compulsive buying, no formal publication has examined the relationship between impulsive buying and compulsive buying at the same time on the same population (or sampling) base. This study represents an effort to examine the purchase-related scales concurrently. More importantly, it seeks to uncover the hierarchical relationships among Big Five personality traits, impulsive buying and compulsive buying. This is a significant area to study, since detailed analyses of the relationships between scales representing different forms of purchase behavior will offer insights into the structure of purchase-related traits, while at the same time provide evidence about the usefulness of the proposed typology. Also, consumers’ standing on the purchase-related traits could be related to more basic personality dimensions (e.g., Big Five factors) so as to find out to what extent different purchase behaviors have their roots in more basic tendencies of the individual consumer (Baumgartner, 2002).

Hypotheses and Research Question

Impulse buying tendency is defined as the

“degree to which an individual is likely to make unintended,

immediate, and unreflective purchases” (Weun et al., 1997, p. 306). Impulsive buyers tend to purchase things spontaneously and kinetically (Rook & Fisher, 1995). Considered a common consumer behavior, impulsive purchases are usually made in a store that are different from those a consumer planned to make entering the store. High impulse buyers are more likely to have more open shopping lists, are more likely to be receptive to sudden and unexpected buying ideas, and are more apt to respond to spontaneous buying stimuli (Bearden & Netemeyer, 1999). Many stores rely on impulse spenders for profit and arrange their products to make an impulse purchase more likely. Verplanken and Herabadi (2001) developed a 20-item scale to measure general impulse buying tendency, which included cognitive aspects (e.g., lack of planning) and affective aspects (e.g., feeling of excitement). Then they correlated their scale with the five-factor personality inventory (extraversion, agreeableness, conscientiousness, emotional stability and autonomy) (Hendriks et al., 1999). They found that the complete scale correlated positively with extraversion, negatively with conscientiousness and autonomy. Emotional stability and agreeableness were unrelated to impulse buying tendency.

Compulsive shopping is different from impulsive buying in that the trait, also known as oniomania, has been classified by the American Psychological Association as an under-recognized and hard- to- treat disorder. Compulsive buying disorder is characterized by excessive or poorly controlled preoccupations, urges, or behaviors regarding shopping and spending (Black, 2001). As an abnormal or aberrant form of consumer behavior, compulsive consumption is “typified by chronic buying episodes of a somewhat stereotyped fashion in which the consumer feels unable to stop or significantly moderate the behavior(s)” (Bearden & Netemeyer, 1999, p. 52). Psychiatric research indicated that compulsive buying has the earmarks of a psychiatric disorder (Black, 1996). Desarbo and Edwards (1996) linked compulsive consumption to such traits as dependence, materialism, perfectionism, denial, depression, approval seeking, low self-esteem, excitement seeking, and lack of impulse control. Valence et al. (1988) identified positive and

169

significant correlations between compulsive buying scale on the one hand and depression, anxiety and nervousness on the other. Black et al. (1998) also found that compulsive buyers were more likely than comparison subjects to develop lifetime mood disorders (especially major depression).

Mowen and Spears (1999) developed a hierarchical model in which the Big Five personality factors were employed as cardinal traits (independent variables), the needs for arousal and for materialism as central traits (meditating variables), and compulsive buying as the surface traits (dependent variables). Mowen and Spears (1999) employed the traits of extraversion, agreeability, conscientiousness, emotional stability, and openness to experience that were found in the Big Five model developed by Goldberg (1992) and refined by Saucier (1994). Among other things, Mowen and Spears (1999) found positive relationships between compulsive buying and neuroticism (emotional instability), and between compulsive buying and agreeableness. A negative relationship was found between compulsive buying and conscientiousness. No relationships were established between compulsive buying and the other two Big Five factors, extraversion and openness (or creativity). By modeling the cardinal, central and surface traits at the same time, Mowen and Spears (1999) demonstrated a means for understanding the psychological antecedents of certain surface traits. Another study by Mowen (2000) investigated the mediating role of impulsivity. In his diagram, Mowen included the Five-Factor traits, as well as need for material resources (materialism), need for body resources, and need for arousal, as elemental traits (independent variables), with impulsiveness and competitiveness as mediating variables, and compulsive buying as dependent variable. The results revealed that emotional stability, introversion, and openness to experiencewere negatively related to a measure of impulsiveness. Materialism was positively related to impulsiveness. The model accounted for 30 percent of its variance of impulsiveness. In turn, a positive relationship was found between impulsiveness and compulsive buying. Other significant predictors of compulsive buying were emotional instability (negative relation), conscientiousness (negative relation), and materialism (positive relation). The model accounted for 31 percent of the variance in compulsive buying.

One of Mowen’s above findings confirmed a well-documented relationship between impulsivity and compulsive buying (e.g., DeSarbo and Edwards, 1996; Lejoyeux et al., 1997).

Nonetheless, the relationship between impulsive buying (a purchase-related trait) and compulsive buying has been suggested by many, but not been proven empirically. For example, as O’Guinn and Faber (1989) argued, impulsive buyers undergo an acute loss of impulse control while they are shopping, whereas compulsive buyers suffer from a chronic loss of impulse control that develops into a repetitive pattern marked by much more dire consequences than that experienced by the impulsive shopper. Thus they suggested that compulsive buying is a much more serious phenomenon than impulsive buying. According to them, impulsive consumers are not as extreme in their behavior as compulsive consumers, and they merely qualify as being materialistic. In other words, impulsive buying is a behavior pattern that happens to ordinary consumers.

Both of these two purchase-related constructs seem to be related to a loss of impulse control. The difference tends to lie in the degree of the loss of impulse control, which we believe can be measured by the same impulsiveness scale developed by Puri (1996) and used by Mowen (2000). It is also our belief that impulsiveness has the same psychological valence as emotional instability. Other sources even suggested that impulsive buying could lead to compulsive buying behavior (Zanella, 2001). In sum, our inspection of previous literature seems to suggest that both impulsive buying and compulsive buying can be traced to the same cardinal psychological traits. It is also possible that impulsive buying as a central trait, along with cardinal traits (Big Five factors), will help predict compulsive buying as a surface trait. These propositions and previous literature review lead to the development of two hypotheses and three research questions (RQ) for this study.

H1: A positive relationship will be found between impulsive buying and compulsive buying. H2: Both impulsive buying and compulsive buying will be positively related to neuroticism (emotional instability). RQ1: Can previously reported correlates of impulsive buying and compulsive buying be replicated in this study? Such as positive relationship between extraversion and impulsive buying, positive relationship between agreeableness and compulsive buying, and negative relationships between conscientiousness and each of the two purchase-related traits.

170

RQ2: How much variance in impulsive buying can be accounted for by Big Five factors? RQ3: How much variance in compulsive buying can be accounted for by Big Five factors and impulsive buying? Method

The objective of this study is not to find diagnosable compulsive and impulsive consumers among a sample of subjects. Instead, it aims at examining the hierarchical relationships among Big Five personality traits, impulsive and compulsive buying tendencies that exist in the same sample of subjects.

Using three previously established scales, the study developed a questionnaire with three major sections. The first section contained questions designed to measure a person’s level of compulsive buying. Participants were instructed to read a series of statements regarding their consumption tendencies and rated how strongly they personally agree or disagree with these statements, where “–3” indicates “strong disagreement,” “0” means “neither agreement nor disagreement” and “3” signifies “strong agreement.” There are a total of fourteen questions, which were taken verbatim from Faber and O’Guinn’s compulsive consumption scale (1989).

The second section contained questions designed to measure the impulsive buying. It is a combination of buying impulsiveness scale developed by Rook and Fisher (1995) and impulse buying tendency scale by Weun et al. (1997). These two scales were put together in our survey because the impulse buying tendency scale by Weun et al. (1997) took account of a new dimension of impulsive buying, that is, the hedonic (or emotional) dimension of engaging in impulsive buying (e.g., It is fun to buy spontaneously). However, this dimension was not covered in the buying impulsiveness scale (Rook & Fisher, 1995). It is hoped that a combined scale of impulsive buying reflects both the cognitive and emotional aspect of the impulsive buying. Thus this study combined the scales by Rook and Fisher (1995) and Weun et al. (1997) and created a 14-item scale, which contains two items with reverse scoring. Like the compulsive buying scale, the combined impulsive buying scale was measured on a 7-point Likert scale from “-3” to “3”.

The third section was designed to examine the participant’s personality traits. The 50-item personality questions were adapted from the 40-item Big Five personality scale developed by

Saucier (1994), with additional question items taken by Mowen and Spears (1999) from Trapnell and Wiggins (1990), and Duijsens and Diekstra (1995). The confirmatory factor analysis conducted by Mowen and Spears on Saucier’s 40-item scale failed to confirm the structure of Five-Factor model. We were concerned that we would not be able to successfully confirm the structure either by simply adopting Mowen and Spears’ refined scale. So we decided to combine both of them in the hope of developing a satisfactory Five-Factor model out of these combined items by using confirmatory factor analysis. The personality items, which ask respondents how accurately they describe themselves, were measured on a 9-point scale, from “1” (extremely accurate) to “9” (extremely inaccurate).

Two hundred and twenty-four college students (at both graduate and undergraduate levels) volunteered to participate in the study. These subjects mainly attended a state university on the west coast and the rest of them went to a private college on the east coast. Once the surveys were collected, questions and answers were coded and entered into the SPSS for data mining and analyses. Recodings were done on reverse scored questions. Separate reliability tests for both compulsive and impulsive questions were conducted. The Cronbach’s alpha was 0.79 for 14 compulsive questions, close to the alpha level (alpha = .82) originally reported by Faber and O’Guinn (1989) in their 14-item initial scale development. The 14 impulsive questions produced an alpha of 0.90, higher than the alpha of .82 reported by Rook and Fisher (1995) for their original 9-item buying impulsiveness scale (the alpha level for the 5-item impulsive buying tendency scale by Weun et al. was not reported). Impulsive questions were thus averaged to form an overall score that measures impulsive buying. Compulsive questions were averaged to create an overall composite score that measures compulsive buying.

An exploratory factor analysis with the expectation of extracting 5 factors was first conducted on the 50-item personality questions. Items with factor loadings lower than .50 were deleted from the original list (because they tended to spread across several factors). There were 33 items left with 5 items measuring conscientiousness, 7 items for extraversion, 7 items for neuroticism (emotional instability), 7 items for agreeableness, and 7 items for openness. Using Amos 4.01, we performed a confirmatory factor analysis on the 33 items. The results revealed a satisfactory fit, with χ2 = 1149.4, df = 495, ρ < .00, CFI = .97, NFI = .95, TLI = .96, RFI = .94,

171

RMSEA = .08. The paths between the five latent variables (Big Five factors) and their corresponding items were all significant at ρ< .00. As a result, new scales that measure Big Five personality traits were created out of the 33 items. All coefficient alphas were higher than .81 (Table 1).

With the help of Amos 4.01, a structural equation model was established, with the Big Five personality traits serving as exogenous variables (cardinal traits), impulsive buying scale as the mediating variable (central traits), and compulsive buying scale as the dependent variable (surface trait). The modification indexes were examined to eliminate or add path as necessary. Figure 1 provides the results of this model.

Results and Discussion

The fit indexes for the hierarchical model were satisfactory, with χ2 = 14.4, df = 6, ρ < .05, CFI = .99, NFI = .99, TLI = .99, RFI = .98, RMSEA = .08 (Figure 1). A path from impulsive buying to compulsive buying was significant at ρ < .00, supporting the first hypothesis that these two purchase-related traits are significantly related. The paths from neuroticism (emotional instability) to compulsive buying, and from neuroticism to impulsive buying, were both significant at ρ < .01, thus supporting the second hypothesis that both consumption traits were positively related to the psychological trait of emotional instability. As was also indicated in Figure 1, the path from extraversion to impulsive buying was also marginally significant at ρ < .10, further replicating a previous finding by Verplanken and Herabadi (2001). However, this model was not able to replicate previous reports about the positive relationship between agreeableness and compulsive buying (Mowen & Spears, 1999), negative relationship between conscientiousness and impulsive buying (Verplanken & Herabadi, 2001), negative relationship between conscientiousness and compulsive buying (Mowen & Spears, 1999). It shows the necessity of more extensive future research on investigating correlates of impulsive buying and compulsive buying. The hierarchical model also identified an additional significant direct path to impulsive buying that was not originally hypothesized or reported elsewhere. The path from openness to impulsive buying was significant (at ρ < .05), a seemingly surprising finding, since a negative relationship was previously identified between impulsiveness (though not impulsive buying) and openness to experience by Mowen (2000). However, we believe the positive path from openness to experience to impulsive buying is understandable, because

unplanned and impulsive purchase is associated with tendencies to be creative, original and imaginative. Despite the significant direct paths from neuroticism, extraversion and openness to impulsive buying, only 6% of the variance in impulsive buying was accounted for by the Big Five personality traits. As a comparison, impulsive buying, along with the Big Five factors, accounted for 44% of the variance in compulsive buying (which also demonstrates the effectiveness of this model). The discrepancy in terms of the variances explained for impulsive buying and compulsive buying provides empirical evidence for the argument that impulsive buying occurs as a response to external stimuli (e.g., situational factors such as point-of-purchase displays), while compulsive buying results from an internal and psychological trigger, such as emotional instability, anxiety and impulsivity (DeSarbo & Edwards, 1996), and in this study, impulsive buying.

Conclusion

This study developed a structural equation model that investigates the hierarchical relationships among Big Five personality traits, impulsive buying and compulsive buying. It identified positive relationships between impulsive buying and compulsive buying, between emotional instability and the two purchase-related traits, and between impulsive buying and two other personality traits (openness and extraversion). While the study replicated previous findings on the characteristics of impulsive buying and compulsive buying, it provided partial (if not indirect) evidence for previous suggestions that impulsive buying behavior tends to be motivated by immediate gratification and are more likely to be prompted by the physical proximity of a desired product (Rook & Fisher, 1995), and less likely by inner (cardinal) personality traits. On the other hand, the study demonstrated that a significant amount of variance in compulsive buying can be accounted for by cardinal and central traits, especially emotional instability and impulsive buying, further supporting the argument that compulsive purchases are activated by psychological tension (Valence et al., 1988) and respond to negative feelings or events (Faber & O’Guinn, 1992). As previous studies suggested (O’Guinn & Faber, 1989), impulsive buying might be derived from an occasional loss or lack of impulse control. When the occasional lack of impulse control accumulates to a certain point where a chronic loss of impulse control occurs, impulsive buying might develop into compulsive buying.

172

Table 1. Items Included to Measure Big Five Personality Traits Note: R indicates that item is reverse scored.

A. Neuroticism (Emotional Instability) (α = .81) 1. Moody more than others 2. Emotions go way up and down 3. Testy more than others 4. Temperamental 5. Fretful 6. Jealous 7. Touchy

B. Conscientiousness (α = .82) 1. Organized 2. Disorganized R 3. Orderly 4. Efficient 5. Sloppy R

C. Agreeableness (α = .81) 1. Sympathetic 2. Kind to others 3. Warm 4. Tenderhearted with others 5. Cold to others R 6. Unsympathetic R 7. Rude with others R

D. Extraversion (α = .84) 1. Quiet when with people R 2. Shy R 3. Talkative when with others 4. Withdrawn from others R 5. Feel uncomfortable in a group of people R 6. Feel bashful more than others R 7. Extroverted when with people

E. Openness (α = .82) 1. More original than others 2. Frequently feel highly creative 3. Appreciate art 4. Enjoy beauty more than others 5. Find novel solutions 6. Imaginative 7. Uncreative R

173

Figure 1. Path Model for Big Five Personality Traits, Impulsive Buying and Compulsive Buying

Note: All the above paths represented by solid lines were statistically significant at least at ρ < .10. The paths represented by dotted lines were not significant. *Compulsive = Compulsive Buying; **Impulsive = Impulsive Buying. REFERENCES Baumgartner, H. (2002). Toward a personology of

the consumer. Journal of Consumer Research. 29, 286-292.

Bearden, W.O., & Netemeyer, R.G. (1999). (Eds.). Handbook of marketing scales: Multi-item measures for marketing and consumer behavior research. Thousand Oaks, CA: Sage Publications.

Black, D.W. (1996). Compulsive buying: A review. Journal of Clinical Psychiatry, 57(8), 50-55.

Black, D.W. (2001). Compulsive buying disorder: Definition, assessment, epidemiology and clinical management. CNS Drugs. 15(1). Special Issue: (pp. 17-27). New Zealand: Adis International Ltd.

Black, D.W., Repertinger, S.G., Gaffney, G.R., & Gabel, J. (1998). Family history and psychiatric comorbidity in persons with

compulsive buying: Preliminary findings. American Journal of Psychiatry. 155 (7), 960-963.

DeSarbo,W., & Edwards, E. (1996). Typologies of compulsive buying behavior: A constrained clusterwise regression approach. Journal of Consumer Psychology. 5, 231-262.

Duijsens, I.J., & Diekstra, R.F.W. (1995). Typologies of compulsive buying behavior: A constrained clusterwise regression approach. Journal of Consumer Psychology, 5, 231-262.

Faber, R.J., & O’Guinn, T.C. (1989). Classifying compulsive consumers” Advances in the development of a diagnostic tool. In Thomas K. Scrull (Eds.), Advances in Consumer Research, V. 16, 738-744.

Faber, R., & O’Guinn, T. (1992). A clinical screener for compulsive buying. Journal of Consumer Research. 19, 459-470.

Fisher, R., & Rook, D. (1995). Normative influences on impulsive buying behavior.

.17

Neurotic

Conscientious

Agreeable

Extravert

Openness

Impulsive**

.19

.14

R2 = .06

.12

-.05 -.06

.01

.07

.04

.04

R2 = .44

.59 Compulsive*

174

Journal of Consumer Research. 22, 305-314.

Goldberg, L.R. (1992). The development of markers for the big-five factor structure. Psychological Assessment, 4, 26-42.

Hendriks, A.A. J. , Hofstee, W.K.B., & De Raad, B. (1999). The five-factor personality inventory (FFPI). Personality and Individual Differences, 27, 307-325.

Lejoyeux, M., Tassain, V., Solomon, J., & Ades, J. (1997). Study of compulsive buying in depressed patients. Journal of Clinical Psychiatry. 58(4), 169-173.

Mowen, J., & Spears, N. (1999). Understanding compulsive buying among college students: A hierarchical approach. Journal of Consumer Psychology, 8, 407-425.

Mowen, J. (2000). The 3M model of motivation and personality: Theory and empirical applications to consumer behavior. Boston: Kluwer Academic press.

O’Guinn, T., & Faber, R. (1989). Compulsive buying: A phenomenological exploration. Journal of Consumer Research, 16, 147-171.

Puri, Radhika. (1996). Measuring and modifying consumer impulsiveness: A cost-benefit accessibility framework. Journal of Consumer Psychology, 5 (2), 87-113.

Saucier, G. (1994). Mini-markers: A brief version of Goldberg’s unipolar big-five markers. Journal of Personality Assessment, 63, 506-516.

Trapnell, P.D., & Wiggins, J.S. (1990). Extension of the interpersonal adjective scales to include the big five dimensions of personality. Journal of Personality and Social Psychology, 59, 781-790.

Valence, G., d’Astous, A., & Fortier, L. (1988). Compulsive buying: Concept and measurement. Journal of Consumer Policy, 11, 419-433.

Verplanken, B., & Herabadi, A. (2001). Individual differences in impulse buying tendency: Feeling and no Thinking. European Journal of personality, 15, S71-S83.

Weun, S., Jones, M.A., & Beatty, S.E. (1997). A parsimonious scale to measure impulse buying tendency. In W.M. Pride and G.T. Hult (Eds.), AMA Educators’ Proceedings: Enhancing knowledge development in marketing (pp. 306-3-7). Chicago: American Marketing Association.

Zanella, V. (2001). How to curb impulse buying tendencies. PageWise Inc. Retrieved on October 16, 2001 from http://utut.essortment.com/impulsebuying_rhtb.htm.

175

7.3.1 TAKE AVERSION VS. THE JOY OF GIVING: A CASE OF GUILT-BASED ASYMMETRY

Michal Strahilevitz, University of Arizona

ABSTRACT Giving to others (e.g., donating to charity) and acquiring for oneself (e.g. buying a new TV) can both be viewed as positive outcomes. Indeed, in pursuit of the positive emotions that helping others and acquiring can each create, people are often willing to sacrifice both money and time. Kahneman and Tversky (1979) demonstrated the phenomenon known as loss aversion – that losses loom larger than gains. In this research, a series of experiments will be presented that demonstrate a phenomenon that can best be described as “take aversion” -- that taking away looms larger than not giving. In the first study, each subject was randomly assigned to one of four reward conditions. All subjects were given a fifteen-minute questionnaire, followed by a reward for having completed it. In the first condition, “cash endowment,” the reward was $1 in cash. In the second condition, “charity-choice endowment,” the reward was a $1 donation to their choice of three charities. In the third condition, “full-choice endowment,” subjects chose whether they preferred to get $1 cash, or donate that $1 to their choice of three charities. In the fourth condition, “charity-without-choice endowment,” subjects were endowed with a donation to one of the three charities (with the charity being randomly assigned instead of chosen by the participant). The same three charities were used in the second, third, and fourth conditions, whereas no charity was used in the first condition. In the last stage of this study, all subjects were given the choice between keeping their original endowment and switching to one of the other options that had been offered in the other conditions. In other words, those who were endowed with money had the option of switching to one of the three charities or keeping the money, whereas those who were endowed with their choice of donations were given the option of leaving the donation with that charity, switching to a new charity, or taking $1 in cash rather than having the donation made. Similarly, in the condition where subjects were endowed with a donation (but not given a choice regarding which charity to donate the money to), subjects had the option of switching to a new charity, switching to cash, or keeping the

donation where it was. Finally, subjects who had initially been endowed with all four options were given the option of keeping their initial choice, or switching to one of the other options. Dependent measures included initial choice for the two conditions where there were two consecutive choice tasks, and final choice for all four conditions. Of those who had been endowed with cash (condition 1), the vast majority preferred to keep the money when given the option of switching to any one of three charities. Of those who had been endowed with their choice of three charities (condition 2), the majority chose not to switch to the cash option when it was presented to them (which was after they had selected a charity), and none chose to switch charities. Most of the participants who were given all four options (condition 3) chose cash initially. Interestingly, none of the participants who had started with all four options chose to switch from charity to cash or from one charity to another, however several did switch from cash to charity. Of those endowed with a donation without a choice of charity (condition 4), almost a third of the subjects switched to another charity, and slightly more than a third stuck to the charity they had been endowed with (suggesting that take aversion operates to some extent even when taking from one charity and giving to another). However, in this “endowed with one charity” condition, more subjects switched to a new charity option than to a cash option (suggesting that one donation to charity is a better substitute for a donation to a different charity than is cash). Verbal protocols and justifications suggested that guilt was a major driver of aversion to undoing a donation. A second study will be presented which distinguishes take aversion from status quo bias and looks at some of the underlying variables (i.e., guilt, self-perception, and personal ethics) that might be responsible for take aversion. Implications for understanding what take aversion and loss aversion share in common and how they differ will be discussed. The discussion will also include an examination of how the guilt of taking (from a good cause), the pain of losing (one’s own wealth), the joy of giving (to a good cause), and the fun of gaining or acquiring (for oneself) differ from

176

one another in terms of both affect and consequences. For further information contact: Michal Strahilevitz, Marketing Department, Eller School of Business, University of Arizona, 320 McClelland Hall, Tucson, AZ 85721, email: consumerbehavior@yahoo.com

177

7.3.2 A LITTLE SOMETHING FOR ME, AND MAYBE FOR YOU TOO: PROMOTIONS THAT

RELIEVE GUILT

Sooyeon Nikki Lee, New York University Priya Raghubir, New York University

Kim P. Corfman, New York University

ABSTRACT Depending on the nature and the intended user of a purchase, consumers may experience feelings of guilt that discourage purchase. The intensity of the guilt is highest when consumers anticipate purchasing hedonic or pleasurable products and services for themselves (vs. someone else), which are more likely than utilitarian or practical purchases to be considered unnecessary. The literature on guilt defines the construct as mainly interpersonal and often resulting from a sense of advantageous inequity – getting more than one deserves compared to someone else or compared to an absolute standard. An obvious way to alleviate feelings of guilt is to restore equity by redistributing outcomes, as has been demonstrated in the social justice literature (Walster, Walster, and Berscheid 1979). The guilt associated with the purchase of a hedonic product for oneself may result from a comparison with someone else – someone else’s frugal behavior may induce feelings of guilt in someone contemplating the purchase of a luxury item – or it may be more budget-based or related to values of frugality or practicality. In the latter case, when a consumer contemplating a hedonic purchase feels guilt, it is due to a perceived imbalance between the amount of frugal or practical behavior and the amount of self-indulgent behavior implied by the purchase. Consumers clearly vary in their feelings about the optimal balance between these kinds of behaviors and in their tolerance for engaging in more self-indulgent behavior than those around them. However, we propose that the feelings of guilt that do arise can be reduced by adding behaviors to the practical side of one’s own scale or increasing someone else’s consumption behavior. Marketers can help consumers make those equity-restoring, guilt-reducing adjustments by offering free gifts with purchase. We predict and demonstrate that a free gift offered with the purchase of a hedonic product reduces feelings of guilt and that the degree to which guilt is reduced depends on the intended user of the free gift and whether it is, itself, hedonic or utilitarian. First, we hypothesize that when the free gift is intended for use by the purchaser, guilt

reduction is greater when the gift is utilitarian than when it is hedonic. This is because it helps consumers restore the balance between self-indulgent and practical behaviors. Second, we hypothesize that guilt reduction is greater when the free gift is intended for use by someone else than by the purchasers themselves (e.g., the purchase of a bottle of women’s perfume that comes with a tube of men’s shaving cream). Giving something to someone else as well as to oneself improves interpersonal equity. Finally, we hypothesize that when the free gift is intended for someone else, guilt reduction is greater when the gift is hedonic than when it is utilitarian (e.g., the bottle of women’s perfume comes with men’s cologne). Not only is the purchaser increasing someone else’s consumption level, the weight is all being placed on the hedonic side of the other party’s scale. In several pilot studies, we have confirmed the association of stronger feelings of guilt with hedonic (vs. utilitarian) purchases for oneself and with free gifts intended for oneself (vs. for others). In the studies, we focus on the situation in which there is the most guilt to alleviate – when the anticipated purchase is hedonic and look across a variety of product and services types and price levels. In Study 1, we show that when a free gift is included, feelings of guilt decrease using pictorial ads as stimuli. Going further within the free gift condition, we investigate the intended user and nature of the free gift, in a 2 (for the purchaser vs. for someone else) X 2 (hedonic vs. utilitarian) between-subject design in Study 2. The guilt as the driving force in altruistic behavior has been examined, but not in the context of consumer purchases of hedonic products and services or promotions. Thus, this research provides insight into the ways in which consumers balance their consumption behaviors to reduce feelings of guilt and how promotions can be used to influence the feelings of guilt experienced by those who anticipate making hedonic purchases for themselves. For further information contact: Priya Raghubir, Haas School of Business, University of California, Berkeley, CA 94720-1900, email: raghubir@haas.berkeley.edu

178

7.3.3 FORM VS. FUNCTION: EMOTIONAL AND BEHAVIORAL CONSEQUENCES OF MAKING HEDONIC VS. FUNCTIONAL TRADEOFFS

Ravi Chitturi, University of Texas at Austin

Raj Raghunathan, University of Texas at Austin Vijay Mahajan, University of Texas at Austin

ABSTRACT Most important decisions, including consumption decisions, involve making trade-offs. In the process of making difficult trade-offs, consumers experience significant negative affect (e.g., Luce 1998). For example, decisions involving a trade-off between safety of a car (high vs. low) and its price (low vs. high) may induce negative emotions that, in turn, influence subsequent behavior (e.g., Luce, Payne and Bettman 1999). In general, decisions involving difficult trade-offs have been found to induce decision avoidance—the tendency to postpone purchase—presumably because consumers feel uneasy about taking a decision without first having resolved the negative emotions evoked by that decision-situation (Luce 1998). Although decision avoidance is one way to resolve, at least temporarily, the conflict induced by difficult trade-offs, not all purchase situations may permit avoidant behavior, because of time or other constraints. It is, therefore, important to ascertain the nature of consumption decisions under these circumstances. Previous research (e.g., Luce, Bettman and Payne 2000) has found that, in decisions involving price-quality trade-offs, people tend to choose the option higher in quality. To formulate a more general theory, however, it is important to incorporate dimensions other than price or quality, since consumers often face decisions involving a trade-off between two equally priced alternatives of equal overall quality (e.g., a Volvo and a Lexus) that differ on other attributes. The focus of this research is on the emotions induced by decisions involving trade-offs between utilitarian vs. hedonic attributes—which refer to the functional and aesthetic aspects of products, respectively—and the effect of these emotions on purchase behavior. Marketing literature provides conceptual support for the idea that the utilitarian and hedonic dimensions capture important aspects of product differences (e.g., Hirshman and Holbrook 1982; Strahilevitz and Myers 1998). However, little research pertains directly to the emotional and behavioral impact of making decisions involving trade-offs between utilitarian vs. hedonic

attributes. For example, it is unclear, from previous research, whether and to what extent negative emotions are induced in decisions involving a choice between, say, a highly functional computer that does not look good and one that is not so functional, but is aesthetically pleasing. In this research, we report findings from three experiments that were conducted to test for emotional and behavioral consequences of making such trade-offs. In brief, we find that decisions involving trade-offs between utilitarian and hedonic attributes evoke feelings of guilt and sadness. Specifically, subjects report feeling guilty when considering the purchase of the hedonically superior (and functionally inferior) alternative, and report feeling sad when considering the functionally superior (and hedonically inferior) choice. The experience of these negative emotions, in turn, leads to asymmetric effects on the purchase decision. Specifically, whereas the experience of guilt leads to an increased likelihood of purchasing the functionally superior alternative, the experience of sadness does not result in a similar increase in the likelihood of purchasing the hedonically superior alternative, controlling for the intensity of these emotions. Reminiscent of preference reversal effects documented in previous decision-research (e.g., Lichtenstein and Slovic 1973), however, our subjects are willing to pay more for products that are hedonically superior. That is, although subjects prefer to choose the option that is functionally superior, they are willing to pay more for the option that is aesthetically more pleasing. Together, these results suggest that marketers should focus first on providing a satisfactory level of functional performance, and thereafter focus on enhancing the hedonic appeal of their offerings. For further information contact: Raj Raghunathan, University of Texas at Austin, Marketing Department, CBA 7.202, 1 University Station, #B6700, Austin, TX 78712, email: raj.raghunathan@mccombs.utexas.edu

179

8.1.1 RETAIL ASSORTMENT VARIETY: HOW FOCUS ON “ASSORTMENT” VERSUS THE “ITEM” CAN AFFECT JUSTIFICATION STRATEGIES

Barbara E. Kahn, University of Pennsylvania

Mary Frances Luce, University of Pennsylvania Seshan Ramaswami, Singapore Management University

ABSTRACT

In most shopping experiences there is an inherent tension: The more options and attributes that are carefully considered, the more likely the consumer will make an optimal choice; however, the more factors considered the more likely the consumer will face increased choice conflict. If preferences are not strong and stable, consumers may rely on “justification heuristics,” which can yield sub-optimal decisions but are easy to justify and thus can lessen the perceived conflict.

In a retail setting, consumer decisions are frequently made in a sequential or hierarchical fashion where an initial choice is made first among a set of assortment of options and then a single item is made from the specific assortment selected. How the consumer partitions the shopping task between the store decision and the final item decision is likely to influence the nature of the perceived decision conflict, and thus also influence the ultimate conflict-reduction strategies employed.

A common strategy that is often used in retail environments when preferences are uncertain is a reliance on brand name. By focusing on brand name as a signal for quality, consumers can avoid making the difficult trade-offs necessary to choose an optimal selection. However, we hypothesize the way branding may operate as a justification heuristic will depend upon whether the consumer is focusing on the “assortment” or on the “item.” If the consumer is focused on the “assortment” level, then we predict consumers will opt for a “flexibility heuristic” and choose the assortment that offers a large variety of brands. Wider assortments are easier to justify because they do not require as many trade-offs to be made at the assortment choice level. On the other, if the consumer is focused on the “item” level then we predict consumers will opt for a “social proof heuristic,” and would choose the item that would be easiest to justify to others. When the consumer is focused on the item level, wider assortments make it more difficult to justify any individual item, and thus

consumers will tend to gravitate towards stores with smaller assortments, where it is easier to justify the choice of a specific item. Therefore, we predict that depending upon whether the consumer is focused on the assortment or item level, choice can be systematically varied.

We test this basic hypothesis in a series of experiments that varies the focus on whether consumers are focused on choosing among assortments or choosing among items. Specifically, we have designed an experimental paradigm in which we systematically compare choices made when participants are first asked to choose between two assortments, and then asked to indicate their final choice within the chosen assortment (the assortment condition) versus when participants are asked to choose their favorite item within each of the two assortments, and then choose the most preferred of those two items (the item condition). We also include as a control, an unconstrained condition where items are not first grouped into assortments and participants are just asked to indicate their final item choice. In all cases the identical items are included in the total choice set; the difference is just whether participants are focused on the first-stage assortment level, or on the final-stage item level.

The results of three experiments support the hypothesis that focus on an assortment level yields a preference for flexibility (i.e., a larger variety of brands) while focus on an item level yields a preference for choices that are easiest to justify to others. In particular, whereas participants preferred assortments with more diverse brands in the former condition, participants preferred assortments with status quo brands or market share leaders in the second condition. For further information contact: Barbara E. Kahn, Marketing Department, 700 Jon M. Huntsman Hall, 3730 Walnut Street, University of Pennsylvania, Philadelphia, PA 19104-6340, email: kahn@wharton.upenn.edu

180

8.1.2 PROJECTING ONE’S OWN VARIETY-SEEKING TRAIT ON TO DECISIONS MADE FOR OTHERS

Tilottama G. Chowdhury, University of Connecticut S. Ratneshwar, University of Missouri – Columbia

Kalpesh K. Desai, State University of New York, Buffalo

ABSTRACT

Is consumer variety-seeking behavior limited to one’s own purchases, or does it also extend to purchases made for others? Prior research has provided considerable support for the former (e.g., Ratner & Kahn 2002; Ratner and Kahneman 1999). Researchers have further distinguished between variety seeking as an internal trait versus variety seeking triggered by external stimuli such as price promotions (Kahn and Louie 1990; Van Trijp et al. 1996). We focus here on variety seeking as a trait-like, individual-difference variable (Baumgartner and Steenkamp 1996; Van Trijp et al. 1996).

Several researchers have studied the variety-seeking trait in regard to individuals’ own purchase behaviors. For example, Van Trijp et al. (1996) find that the “intrinsic” variety-seeking trait interacts with product-level characteristics (e.g., hedonic vs. non-hedonic attributes) in determining whether consumers switch between brands or repeat purchase. However, prior research has not explored whether a person’s own variety-seeking trait might actually be projected on to decisions made products for others. We explore this issue by investigating the composition of consumers’ consideration sets in regard to gift purchase decisions (Desai and Hoyer 2000; Kardes et al. 1993; Roberts and Lattin 1991). Specifically, the key dependent variable in the present research is the degree of heterogeneity in the consideration set formed by a consumer for buying a gift, i.e., the extent to which the choice alternatives in a gift consideration set span diverse product categories (Johnson 1984, 1989; Park and Smith 1989; Ratneshwar et al. 1996).

Why might individuals project their variety-seeking tendencies even on to purchases made for others? Optimal stimulation level (OSL) theory suggests that as stimulation (complexity, arousal, etc) plunges below the ideal level, an individual will try to improve it, mainly through exploration and novelty seeking behavior. A high (vs. low) variety-seeking individual is likely to be more motivated to improve his/her stimulation level by seeking novelty in the decision-making process

(Van Trijp et al. 1996). Further, this same individual may try to achieve an improved stimulation level even while buying gifts for others and thus might look for more variety in choice alternatives for stimulation. Even if a high variety-seeking individual starts the search process with a specific product category in mind, the need for variety and increased stimulation should eventually cause this person to consider other categories. As a result, a high (vs. low) variety-seeking individual will be more likely to browse diverse product categories, which in turn is more likely to produce a heterogeneous consideration set.

This prediction was tested in an experiment where participants (N = 135) shopped for gifts after being provided with gift-giving scenarios. Male and female participants were provided with separate assortments of available products that were tailored to their respective genders. In each case, the assortment consisted of eight different product categories (e.g., wristwatches, bath & beauty products), with each category having four product alternatives (i.e., 32 choice alternatives in all). The products were organized by category and depicted in full color, as if in the pages of a catalog. Participants were asked to indicate the alternatives “they would seriously consider” for buying the gift. Individual differences in the variety-seeking trait were measured later after participants had completed some unrelated filler tasks.

The heterogeneity of participants’ consideration sets was assessed by three different measures (Ratneshwar et al. 1996). The results of the experiment supported the prediction. This fall, two further experiments are planned to identify moderating variables and boundary conditions for the observed effect. In sum, there is now good evidence to suggest that the variety-seeking trait influences purchase behaviors beyond one’s own self-consumption. The present research also extends literature in the area of two-stage choice models (Hauser & Wernerfelt 1990; Roberts & Lattin 1991) by demonstrating the role of individual differences in variety seeking in engendering consideration sets with heterogeneous alternatives.

181

For further information contact: S. (Ratti) Ratneshwar, Department of Marketing, College of Business, 403C Cornell Hall, University of Missouri – Columbia, Columbia, MO 65211, email: ratneshwars@missouri.edu .

182

8.1.3 PRODUCT ASSORTMENT AND INDIVIDUAL DECISION PROCESSES

Alexander Chernev, Northwestern University ABSTRACT

The assumption that more variety is always better is not only intuitively appealing but is also supported by numerous findings in social psychology, decision making, and economics (Payne, Bettman, and Johnson 1993; Simonson 1999). A recently advanced alternative viewpoint suggests that larger assortments (e.g., those with diverse options or more variety) do not always benefit individuals (Iyengar and Lepper 2000). It has been theorized that increasing the size of the choice set may have adverse consequences because it also increases the demand on individuals’ cognitive resources, potentially leading to a cognitive overload (Greenleaf and Lehmann 1995; Huffman and Kahn 1998).

Most of this recent research, however, has focused on documenting the negative effect of large assortments on choice without investigating factors that moderate its impact. This paper goes beyond simply demonstrating the effect and identifies factors that determine when large product assortments will have a positive impact on consumer preferences, as predicted by traditional economics literature, and when large assortments will have an adverse impact on preferences, as suggested by recent findings in the decision literature. More importantly, this research offers insights into the psychological mechanisms and individual decision processes underlying the impact of assortment on choice.

The four experiments presented in this research identify the degree to which individuals have articulated preferences as a factor moderating the impact of assortment on choice. Specifically, it is proposed that individuals without articulated preferences are less likely to choose from larger assortments compared to individuals with an articulated ideal attribute combination. It is further argued that preference articulation moderates the impact of assortment on individuals’ decision processes. It is proposed that, relative to individuals without an articulated ideal point, individuals with articulated preferences (1) are more selective in processing the available information, (2) are more likely to rely on alternative-based rather than attribute-based processing, and (3) are more likely to evaluate the available alternatives in a

confirmatory manner, using their ideal attribute combination as a reference point. In contrast, individuals without articulated preferences are expected to be (1) more comprehensive in evaluating the available alternatives, (2) more likely to rely on attribute-based rather than alternative-based processing, and (3) more likely to evaluate the available alternatives in a comparative fashion, using the other options’ performance as a reference point.

Experiments presented in this research use complementary methodologies to support the above theoretical propositions. Experiment 1 demonstrates that individuals without articulated preferences are less likely to choose from larger assortments relative to individuals with an articulated ideal point. Experiment 2 relies on a computer-based simulation to examine information search patterns, and experiment 3 uses a reason-based analysis to investigate how preference articulation moderates the impact of assortment on decision strategy. Finally, experiment 4 adopts recall-based process measures to provide further support for the research hypotheses.

The four experiments reported in this research yield converging evidence that the impact of assortment on individual decision processes and choice is moderated by the degree to which individuals have articulated attribute preferences. The data show that individuals with an articulated ideal point are more likely to choose from larger assortments than individuals without a readily available ideal attribute combination. Furthermore, these experiments offer insights into the impact of preference articulation and assortment on individuals’ decision strategy. Specifically, preference articulation is shown to be associated with more selective, alternative-based, and confirmatory processing for individuals with an articulated ideal attribute combination and more comprehensive, attribute-based, and comparative processing for those without articulated preferences. For further information contact: Alexander Chernev, Northwestern University, Kellogg School of Management, 2001 Sheridan Road, Evanston, IL 60208, email: ach@kellogg.northwestern.edu

183

8.2.1 SELF-MONITORING AND ATTITUDE ACCESSIBILITY REVISITED

Sandor Czellar, University of Lausanne

ABSTRACT

Research shows that low self-monitors form judgments and intentions and display behavior that are more consistent with their attitudes than high self-monitors (Ajzen, Timko & White, 1982; DeBono & Omoto, 1993; Kraus, 1995). This observation about the rather high attitude-behavior consistency of low self-monitors has led several scholars to investigate the relation between self-monitoring and attitude accessibility (DeBono, Green, Shair & Benson, 1995; DeBono & Snyder, 1995; Kardes, Sanbonmatsu, Voss & Fazio, 1986; Mellema & Bassili, 1995).

However, research on self-monitoring and attitude accessibility has remained inconclusive about two alternative hypotheses. In a pioneering study, Kardes, Sanbonmatsu, Voss and Fazio (1986) reported shorter response latencies for low self-monitors than for high self-monitors in implicit cognition tasks. A first explanation of this result is that low self-monitors could have more accessible attitudes than high self-monitors. An alternative explanation is that the attitudes of both low- and high self-monitors may be equally accessible, but that – out of a concern for self-presentation - high self-monitors deliberate more in certain attitude response tasks, even if these are implicit in the form of response latencies. Our purpose is to investigate whether the attitude accessibility hypothesis or the alternative self-presentation hypothesis explains better the difference between low- and high self-monitors in response latency tasks.

Our investigation relies on the methodological framework of the Implicit Association Test (IAT) (Greenwald, McGhee & Schwartz, 1998). We systematically manipulated the environmental conditions: For a group of participants, IATs were conducted in a setting with low motivation for self-presentation; in another group, we set the IATs with high motivation for self-presentation. The attitude accessibility hypothesis predicts longer response latencies for high self-monitors than low self-monitors in any experimental condition (that is, we can expect a main effect for self-monitoring). The alternative self-presentation hypothesis suggests that high self-monitors will only have longer response latencies in the high self-presentation condition (meaning an interaction effect between self-monitoring and experimental condition).

To increase motivation for self-presentation, we chose a socially sensitive topic: knowledge structures about status. Category concepts and their associated word stimuli were selected on the basis of an exploratory study with sixty participants. The main study was conducted with sixty-six participants, who completed five status-related IATs. A first group of thirty-three students participated in group sessions on a voluntary basis. Upon arrival into a computer room equipped with workstations, they were instructed that they would be taking part in a categorization study as part of a research project. They were instructed to answer as quickly as possible in the tasks and were assured that making mistakes was not a problem as long as they answered quickly. They were also assured that their answers were anonymous and that data would be treated in an aggregate analysis. The purpose of these instructions was to minimize the motivation for self-presentation in the course of the experiment. Participants then completed the five IATs. A second group of thirty-three volunteer students participated in individual sessions; personal appointments were fixed with each if them. Just as before, they were instructed to answer as quickly as possible to a computer-assisted categorization exercise. In addition to this requirement, each participant was informed that she/he would be debriefed after the experiment; that her/his results would be compared with other people’s results and that these analyses could be sent to them if they wanted to. The purpose of these instructions was to increase motivation for self-presentation in the course of the experiment. Participants then completed the same five IATs as people in the group session.

We found no main effect for self-monitoring in response latencies, either in compatible or incompatible tasks. However, we found an overall robust interaction effect between experimental condition and self-monitoring, demonstrating that in high-motivation conditions only did self-monitors have longer response latencies than low self-monitors. In Kardes et al.’s (1986) research, the experimental conditions were comparable to ours in the high motivation condition, allowing for considerable self-presentation opportunity. In contrast to this, DeBono et al. (1995) used group sessions, which minimized self-presentational concerns and obtained no difference between low- and high self-monitors with respect to response latencies. We found similar results in the group

184

conditions. Overall, these results paint a consistent picture and provide a support for the explanation that high self-monitors hesitate and deliberate more in high-motivation conditions than in low-motivation conditions. If motivation for self-presentation increases, low self-monitors carry out attitude tasks in much the same way as previously; whereas high self-monitors, although their automatic associations are equally accessible, tend to hesitate more to “push the right button”.

The evidence reported here strongly suggests that implicit cognition tasks – just as explicit ones – may be sensitive to self-presentation effects, but only in special experimental conditions and for people who do exert expressive control over their behavior. We believe the reliability and validity of applied research using implicit cognition measures would certainly be increased by using experimental conditions that minimize the opportunity for self-presentation. REFERENCES Ajzen, I., Timko, C., & White, J. B. (1982). Self-monitoring and the Attitude - Behavior Relation. Journal of Personality and Social Psychology, 42(3), 426-435. DeBono, K. G., & Omoto, A. M. (1993). Individual Differences in Predicting Behavioral Intentions from Attitude and Subjective Norm. Journal of Social Psychology, 133(6), 825-831. DeBono, K. G., & Snyder, M. (1995). Acting on One's Attitudes: The Role of a History of Choosing Situations. Personality and Social Psychology Bulletin, 21(6), 629-636. DeBono, K. G., Green, S., Shair, J., & Benson, M. (1995). Attitude Accessibility and Biased Information Processing: The Moderating Role of Self-monitoring. Motivation and Emotion, 19(4), 269-277. Greenwald, A. G., McGhee, D. E., & Schwartz, J. L. K. (1998). Measuring Individual Differences in Implicit Cognition: The Implicit Association Test. Journal of Personality and Social Psychology, 74, 1464-1480. Kardes, F. R., Sanbonmatsu, D. M., Voss, R. T., & Fazio, R. H. (1986). Self-monitoring and Attitude Accessibility. Personality and Social Psychology Bulletin, 12(4), 468-474.

Kraus, S. J. (1995). Attitudes and the Prediction of Behavior: A Meta-analysis of the Empirical Literature. Personality and Social Psychology Bulletin, 21(1), 58-75. Mellema, A., & Bassili, J. N. (1995). On the Relationship Between Attitudes and Values: Exploring the Moderating Effects of Self-monitoring and Self-monitoring Schematicitiy. Personality and Social Psychology Bulletin, 21(9), 885-892. For further information contact: Sandor Czellar, University of Lausanne, BFSH 1 1015 Lausanne-Dorigny, Switzerland

185

8.2.2 THAT'S MY OWN CREATION! EXPERIENCING ATTACHMENT TO SELF-ALTERED PRODUCTS

Ruth Mugge, Delft University of Technology, Netherlands

Hendrik N.J. Schifferstein, Delft University of Technology, Netherlands Jan P.L. Schoormans, Delft University of Technology, Netherlands

ABSTRACT

Product attachment is the emotional bond

a consumer experiences with a product during ownership. This definition implies that an emotional tie exists between the owner and his/her object and that this specific durable has an important meaning to the owner. This study investigates how personally altering a product’s appearance contributes to the degree of attachment the owner experiences to this product. The personalization of a product’s appearance has gained increasing interest among companies of consumer durables. Different degrees in personalization can be distinguished: from the redecoration of a cupboard to the interchangeable fronts of NOKIA mobile phones. Until now, research has focused on investigating the effect of personalization on consumers’ preference. This study extends this stream of research by investigating its effect on the experience of product attachment during ownership.

By personally altering the product’s

appearance, the owner invests his/her energy in the product. In addition to the physical energy that is needed to accomplish the actual alteration of the product’s appearance, the owner also invests “psychic energy” in it. Psychic energy is the owner’s mental energy directed to the product. In the literature, it is concluded that the experience of attachment to a product is related to the psychic energy the owner has invested in this product. The more psychic energy a consumer invests, the more meaning gets “layered” on the object and the more important it becomes. To alter the product’s appearance, the owner needs to choose among the possible options for alteration (e.g., design and colors), which requires this psychic energy. As a result, the product is personalized and the product’s appearance becomes self-expressive: It reflects the owner’s self to oneself and to others. Past studies confirmed that by expressing the self, the product gains symbolical meaning, due to which the owner can become attached. In conclusion, we hypothesize that by altering the product’s appearance, the owner invests psychic energy, and the product is used to express the owner’s self. In

turn, self-expression positively affects the degree of attachment to this product.

In our study, we compared students who

had altered their bicycle by painting it in several (conspicuous) colors (n = 58) with students who had not altered their bicycle (n = 91). Bicycles were chosen as the category under study, because Dutch students frequently alter these products’ appearances. In general, Dutch students own a bicycle, which they use daily to travel from their home to university. Dutch students are likely to buy a second-hand bicycle or to use an old one, because of money constraints and because bicycles are often stolen. In the Netherlands, it is common to paint old bicycles to make them personal and distinctive, and to prevent theft. The subjects were asked to fill out a questionnaire concerning their bicycle. In this questionnaire, multi-item measures on seven-point Likert scales were obtained for the dependent variables (i.e., psychic energy, self-expression, and product attachment) and for several control variables (i.e., satisfaction, memories concerning a person, memories concerning an event, product utility, product appearance, and financial value). Because subjects filled out the questionnaire for the bicycle they already owned, it was not possible to assign subjects randomly to conditions. Consequently, control variables were included to control for additional differences between the two conditions.

The results revealed no significant differences for the control variables. We conclude that no other influencing factors were present between the two conditions in addition to the alteration of the product’s appearance. In accordance with our hypotheses, significant differences were found for the dependent variables. Subjects in the “alteration” condition indicated to have incorporated more psychic energy (Malteration = 3.70 vs. Mno alteration = 1.60; t(145) = 10.88, p < .001), considered the product more self-expressive (Malteration = 3.52 vs. Mno alteration = 2.43; t(144) = 5.01, p < .001), and experienced more product attachment (Malteration = 4.13 vs. Mno alteration = 3.53; t(145) = 2.36, p < .05) than those in the “no alteration” condition.

186

To test the relationships between the

dependent variables, we estimated a structural model. The model resulted in a good fit (χ2 = 92.08 (df = 75), p = .09; GFI = .92; CFI = .98; RMSEA = .040). Alteration of the product’s appearance had a significant effect on psychic energy (γ = .65, p < .01). Psychic energy had a significant effect (γ = .62, p < .01) on self-expression, which in turn had a significant effect (γ = .68, p < .01) on product attachment. Three other models were estimated to test whether additional indirect relationships existed between the variables next to the hypothesized relationships (e.g., Psychic energy → Product attachment). None of these relationships were supported by the data. These results give support for the proposed relationships. Specifically, we find that if consumers invest psychic energy in a product by altering its appearance, it becomes an expression of the self. Self-expression, in turn, has a positive effect on the degree of attachment to this product.

As discussed, different degrees of

personalization exist. This study merely investigates the post-purchase alteration of a bicycle through painting it. Future research should examine the effects of other types of personalization and other product categories to generalize our findings. Specifically, it is interesting to investigate the effects of personalizing a product’s appearance in pre- and post-purchase settings. Additionally, this study focuses on creative alterations of a product. Future research should extend this study by differentiating between creative and labor-intensive alterations. For further information contact: Ruth Mugge, Delft University of Technology, Faculty of Industrial Design Engineering, Department of Product Innovation and Management, Landbergstraat 15, 2628 CE Delft, the Netherlands, Phone: +31 15 278 3801, Fax: +31 15 278 7662, Email: r.mugge@io.tudelft.nl

187

8.2.3 SELF ESTEEM AND SOCIAL INFLUENCE: THE MODERATING ROLE OF GENDER

Mehdi Mourali, Concordia University Michel Laroche, Concordia University Frank Pons, University of San Diego

ABSTRACT

Consumers frequently influence and are influenced by others (Bearden and Etzel 1982, Feick and Price 1987). In addition, the literature has documented systematic individual differences in both exerting and receiving social influence (e.g., Bearden et al. 1989, Feick and Price 1987). On one hand, consumers vary in their susceptibility to interpersonal influence (McGuire 1968, Bearden et al. 1989). That is, some people are inherently more influenceable than others across a wide range of social situations. On the other hand, consumers also differ in their tendency to exert interpersonal influence, which is captured by their scores on the market mavenism scale (Feick and Price 1987).

A common determinant to both exerting and receiving social influence is the individual’s self-esteem. Self-esteem represents the evaluative component of a person’s self-concept (Blaskovich and Tomaka 1991). Individuals high in self-esteem are thought to be less susceptible to influence from others, and at the same time, to have a greater tendency to exert influence on others. We propose that the effects of self-esteem on both exerting and receiving interpersonal influence may vary in strength across genders.

Gender differences in social behavior have often been attributed to differences in the social roles occupied by men and women (Eagly 1987). Past research has typically offered that, in enacting their social roles, men are guided primarily by agentic goals, whereas women are guided by communal concerns (Bakan 1966, Spence et al. 1975). Agency reflects the individual drive to differentiate from the collective and is manifested in striving for mastery and power, which enhance and protect that differentiation. Communion reflects the participation of the individual in the collective and is manifested in striving for intimacy, union, and solidarity with the collective (Wiggins 1991). Men, therefore, are characterized as being relatively self-focused. Their self-esteem is mainly derived from instrumental actions, such as self-assertion, self-efficacy, and mastery. In contrast, women are viewed as being sensitive to the needs of both self and others. They derive their self-esteem mainly from relationships (Spence et al.

1975, Spence et al. 1985, Schwalbe and Staples 1991).

Women’ s communal orientation, their greater concern for interpersonal harmony and their emphasis on interpersonal sources of self-esteem suggest that their self-esteem is more strongly associated with both receiving and exerting interpersonal influences than men’s. In other words, we predict that the negative relationship between self-esteem and susceptibility to interpersonal influence is stronger for women than for men, and that the positive relationship between self-esteem and market mavenism is stronger for women than for men.

Data were collected from 419 residents of a large North-American city and were subsequently analyzed using Bentler’s (1992) EQS software. After establishing measurement equivalence across the male and female samples, our analysis focused on comparing the path coefficients linking self-esteem to the various dimensions of social influence between the two groups.

Consistent with our predictions, the findings indicated that the negative effect of self-esteem on susceptibility to normative influence (including the utilitarian and the value expressive dimensions) was stronger for women than for men. Also in line with previous predictions, the positive effect of self-esteem on consumers’ market maven tendencies was found stronger for women. Other noticeable results include the absence of a significant relationship between self-esteem and consumers’ susceptibility to informational influence for both genders, and the absence of a relationship between self-esteem and men’s susceptibility to normative influence.

In sum, confirming the importance of self-esteem as a determinant of social influence, this study has also shown that the effects of self-esteem on consumers’ susceptibility to and exertion of influence are not as simple as it has traditionally been thought. Gender was identified as an important moderator but other variables might also influence the effects of self-esteem. Culture is one such variable.

REFERENCES

188

Bakan, D. (1966), The Duality of Human Existence: An Essay on Psychology and religion. Skokie, IL: Rand McNally.

Bearden, W. O. and M. J. Etzel (1982), “Reference group Influence on Product and Brand Purchase Decisions,” Journal of Consumer Research, 9, 183-194.

Bearden, W. O., R. Netemyer and J. E. Teel (1989), “Measurement of Consumer Susceptibility to Interpersonal Influence,” Journal of Consumer Research, 15, 473-481.

Bentler, P. M (1992), EQS: Structural Equations Program Manual, BMDP Statistical Software, Los Angeles, CA.

Blascovich, J. and J. Tomaka (1991), “Measures of Self-Esteem”, in Measures of Personality and Psychological Attitudes, eds. J.P. Robinson, P.H. Shaver and L.S Wrightsman, San Diego: Academic Press, 115-160.

Brown, J.J. and P.H. Reingen (1987), “Social Ties and Word-of-Mouth Referral Behavior,” Journal of Consumer Research, 14, 350-362.

Eagly, A.H. (1987), Sex Differences in Social Behavior: A Social Role Interpretation. Hillsdale, NJ: Erlbaum.

Feick, L.F. and L.L. Price (1987), “The Market Maven: A Diffuser of Marketplace Information,” Journal of Marketing, 51, 83-97.

McGuire, W.J (1968), “Personality and Susceptibility to Social Influence,” in Handbook of Personality Theory and Research, eds. E. F. Borgatta and W.W. Lambert, Chicago: Rand McNally, 1130-1187.

Schwalbe, M.L. and C.L. Staples (1991), “Gender Differences in Sources of Self-Esteem,” Social Psychology Quarterly, 54, 147-159.

Spence, J.T., K. Deaux and R. Helmreich (1985), “Sex Role in Contemporary American Society,” in Handbook of Social Psychology, 3 (2), eds. E. Aronson and G. Lindzey, NY: Random House.

Spence, J.T., R. Helmreich and J. Stapp (1975), “Ratings of Self and Peers on Sex Role Attributes and their Relation to Self-Esteem and Conceptions of Masculinity and Femininity,” Journal of Personality and Social Psychology, 37, 395-412.

Wiggins, J.S. (1991), “Agency and Communion as Conceptual Coordinates for the Understanding and Measurement of Interpersonal Behavior,” in Thinking Clearly about Psychology, eds. W.M. Grove and D. Cicchetti, Minneapolis, University of Minnesota Press.

For further information contact:

Mehdi Mourali, PhD Student in Marketing, John Molson School of Business, Concordia University, 1455 de Maisonneuve West, Montreal, Quebec, Canada, H3G 1M8, Tel: (514) 848-2424 ext. 2738, E-mail: mmourali@jmsb.concordia.ca

189

8.3.1 REVEALED PREFERENCE DETERMINANTS OF MATE SELECTION: EVIDENCE FROM AN EXPERIMENTAL DATING MARKET

Raymond Fisman, Columbia University

Sheena Sethi-Iyengar, Columbia University Itamar Simonson, Stanford University

ABSTRACT We analyze mate selection behavior in an experimental 'Speed Dating' market. Our data allow us to observe all acceptances and rejections of mates, rather than just matches, which allows for a more refined analysis of mate selection. We draw on earlier literatures in social and evolutionary psychology to make two contributions. First, we document the personal characteristics of a pair that generate acceptances in our experimental setting. Second, we examine whether experimental variation unrelated to personal characteristics affect decisions. We find that choices are affected by group size and ordering of meetings, which would be irrelevant in the frictionless matching model. Collectively, our results provide guidance for future work on the theory of mating markets. The literature on mating markets in economics finds its roots in the pioneering work of Becker (1973), which models marriage as a frictionless matching process. It is built on the assumption of a ‘quality threshold’ rule where individuals are willing to match with partners above a quality cutoff. This model’s parsimony makes it a useful descriptive tool, and it generates assortative matches under which individuals match with partners of similar quality. It is incomplete, however, in a couple of ways. First, the model does not attempt to characterize the determinants of ‘quality’ and how these factors may vary across individuals: Further, while theorists have developed more nuanced mating models that extend the basic framework, these models quickly become very complex, suggesting that some empirical guidance would be useful for determining which assumptions are most in need of modification. In this paper, we utilize the results of an experimental ‘Speed-dating’ market to empirically examine determinants of mating choices. Our approach has the advantage of providing a setting where we observe revealed preferences directly, rather than just matches, and where aspects of the mating process may be controlled. This approach allows for direct observation of individual preferences (i.e., Yes/No decisions for each mating candidate), rather than only final matches as in prior studies.

Motivated by the vast literature on mating in psychology, we generate predictions regarding the determinants of matches, including individual and pairwise personal characteristics (e.g., physical attractiveness) and ‘structural’ aspects of our experimental setting (group size and ordering) that are independent of personal attributes.

We observe that the basic frictionless matching model makes strong empirical predictions, independent of any individual’s definition of quality. Hence, selectivity should be invariant to elements of the experimental setting. Earlier work in psychology suggests some potential deviations from this prediction; in particular, we examine the effects of group size and ordering on mate choice. While there is a vast body of empirical research on mate preferences, with few exceptions, prior work has been based on either survey-based evaluation of attribute importance or observational studies. In our studies, we observe mating choices, matches, as well as attribute importance weights and mate ratings on these attributes, which allows for a much richer analysis. Collectively, our results provide guidance for future work on the theory of mating markets. For further information contact: Sheena Sethi-Iyengar, Columbia University, Columbia Business School, Uris Hall, 3022 Broadway, New York, NY 10027 email: ss957@columbia.edu .

190

8.3.2 DATING PREFERENCE AND DATING DECISIONS: THE EFFECT OF REPEATED CHOICES ON CONSUMER PREFERENCES

Sheena Sethi-Iyengar, Columbia University

Itamar Simonson, Stanford University Raymond Fisman, Columbia University

ABSTRACT In this research we investigate the construction of consumers’ tastes when making repeated choices in a category. It is proposed that, although consumers may have preconceived notions regarding their preferences, their actual choices are often determined by other factors of which they are unaware, cannot easily verbalize, or do not wish to admit. However, consumers are expected to learn their preferences when making repeated choices in a category. Such learning, however, may not last if consumers’ prior theories regarding their preferences are more powerful than the temporary insight provided by their choices. Recent field and laboratory experiments by Iyengar and Lepper (2000) show that while the provision of extensive rather than limited choices is a more appealing choicemaking context, choosers proved more likely to make a choice and commit to their choices when confronted by a more restricted choice set. This investigation seeks to examine why despite the greater ability to optimize resulting from the provision of more rather than fewer choices, choosers experience the task of choosing as being confusing rather than clarifying. We propose that whether the act of choosing is about preference matching or preference-formation depends on the number of options included in the choice set. Specifically, the following study tests the hypothesis that in the limited choice condition choosers would be more likely to make choices that were congruent with their preconceived preferences, thus treating the act of choosing as an exercise in identifying the options matching preconceived preferences; by contrast, choosers in the extensive choice condition would be more likely to make choices which reflected preferences more congruent with post-decision preferences, thus treating the act of choosing as an exercise in ascertaining the content of their preferences. We study these questions in a study involving an actual hurry-dating event. In this field study, one hundred and forty volunteers participated in a hurry-dating event, which took place in an on-campus restaurant located at Columbia University.

During the evening of each dating event, either ten males and ten females (limited choice) or twenty males and twenty females (extensive choice) were matched up with each other—such that every male and female participant met with every participant of the opposite sex. During the dating event, each pair met for four minutes and afterwards spent one minute evaluating one another. The members of each pair confidentially rated each other on attractiveness, fun-loving, intelligence, sincerity, ambition, and shared interests, and then provided a yes or no decision regarding their desire for a subsequent meeting. In turn, participants’ decision criteria was calculated by deriving the correlational relationship between the attribute ratings with subsequent yes/no decisions. In contrast, participants’ preference criteria were obtained through their responses to a dating preference questionnaire which was administered on three occasions: the day before the dating event, the day after the dating event, and three weeks following the dating event. Dating preferences were assessed through participants’ importance ratings of the identical six attributes. Results show first that there were a greater percentage of matches in the limited choice condition (25%) as compared with the extensive choice condition (14%). Second, while the decision criteria among limited choice participants was more correlated with their preference criteria previous to the event, the decision criteria of choosers in the extensive choice condition was more correlated with their preference criteria following the dating event—supporting our hypothesis that limited choice participants engage in preference-matching, while extensive choice participants engage in preference-formation. Interestingly, three weeks later, choosers’ preference criteria in the extensive choice condition was no longer correlated with either their decision criteria or their preference criteria following the dating event; instead, extensive choice participants’ preference criteria three weeks later was correlated with their preference criteria previous to the event, suggesting that their preference criteria had reverted back to their preferences before the dating experience.

191

Theoretical and practical implications of these findings will be discussed. For further information contact: Sheena Sethi-Iyengar, Columbia University, Columbia Business School, Uris Hall, 3022 Broadway, New York, NY 10027 email: ss957@columbia.edu .

192

8.3.3 LEARNING AND JUGGLING IN ONLINE DATING

Jeana Frost, Massachusetts Institute of Technology Dan Ariely, Massachusetts Institute of Technology

ABSTRACT Online dating is becoming more and more popular, with many new users joining every day. In the current research we use a large dataset from a major online dating site to look at this emerging market. In the first part of the paper we provide some general description of this market, and point to some of the structural problems with such online systems. For example, the data shows that the vast majority of requests for correspondence are directed at a very small percentage of the individuals who are members of the site. This behavior causes market inefficiency, because the few individuals who get a lot of email are overwhelmed by the amount of correspondence (in fact in many cases their email quote is quickly exceeded and emails bounce back). These highly sought after individuals do not answer most of their emails, which leaves most of the people who wrote them disappointed. At the same time there are a lot of other individuals who get no or almost no email. A second inefficiency comes from deception. Many individuals in online dating portray themselves in more positive light, which helps them to get correspondence with others, and even to get an initial date. However, during the first meeting the expectations that were created are tested first hand which leads many times to disappointments. Thus, the value of lying as a strategy to start a relationship is self-defeating when it comes to later steps in the relationships. A third inefficiency stems from peoples’ short-term horizon when joining these communities. In most cases, individuals provide only very superficial information about themselves, which makes it harder for others to search for them, or to figure out the degree of their fit (this is also true regarding the pictures that individuals post). After analyzing these three market-inefficiencies (no email discrimination, laying, and low investment), we will propose some changes to the online dating sites that could overcome some of these difficulties and improve the marketplace. The second part of the paper will focus on two specific behaviors: discrimination and juggling. To examine discrimination we measure success level of individuals, defined as the probability to get

responses to new email contacts they try to initiate. The main question of concern here is whether individuals learn over time how to discriminate between individuals who are likely to replay to them from the ones that will not, and also whether they learn over time how to discriminate between individuals who are likely to have a long correspondence with them from the ones that will not. Regrettably, the results thus far show that no such learning takes place, or perhaps that the improvement due to such learning is overwhelmed by optimism. The second question – juggling – draws upon work by Shin and Ariely, showing that individuals value options for their own sake and are willing to expand high effort to make sure that options (even ones they do not want) remain available to them as options. This tendency to invest effort in “keeping doors open” seems to be highly related and relevant to online dating. Many individuals try to start a new relationship with many individuals at once in order to increase the probability of a successful dating relationship. Occasionally, these individuals can get responses from multiple potential partners and at that point these individuals are faced with a dilemma: They can pick the most likely candidate and focus all their effort and time on him or her, risking that the lack of attention to the other options will eliminate them as viable dating partners. Alternatively, these daters can try to juggle multiple relationships at the same tike, not risking the immediate elimination of any one of them but also not giving any of the relationships the full conditions (time and attention) for it to blossom. The initial results suggest that people try to juggle multiple relationships at the same time (with no obvious gender difference). Moreover, the results also suggest that this tendency does not diminish over time. For further information contact: Jeana Frost, MIT, Sloan School of Management, 50 Memorial Drive, Cambridge, Massachusetts 02142, email: frost@media.mit.edu

193

PAPER PRESENTED AT THE 2003 SCP SUMMER CONFERENCE

IDENTIFYING PROSPECTS AND REACHING TARGETS: NEGLECTED DISTINCTION WITHIN MARKETING

Geraldine Fennell, Consultant, Columbus, OH and

Joel Saegert, The University of Texas at San Antonio

ABSTRACT Management's strategy for reaching targets with its advertising message is to select media vehicles whose characteristics reflect disproportionately those of prospects and to rely on the content of the message to engage the attention of targets, i.e., those prospects who experience the targeted wants. It is noted as remarkable that the modern discipline of marketing has existed for five decades, while its Principles of Marketing textbooks fail to include an appropriately differentiated discussion of the way management proceeds to identify prospects, and select, serve, and communicate with targets for marketing strategy. The paper distinguishes and discusses the strategic tasks of defining the market for a venture, i.e., defining the relevant universe including specifying a criterion to qualify prospects and, within that universe, selecting and communicating with targets for the venture. Errors of omission and commission in current textbooks are discussed. INTRODUCTION

How does management plan to use funds earmarked for advertising to best advantage? Facing a given naturally-occurring population, how is management to decide whom it should regard as in its market or, to take just one dimension of market definition, whom to regard as prospects? Later, having selected targets from among its prospects, how does management plan to get the right message to its targets? One answer to such questions is contained in management's marketing analysis and the strategy to which that analysis leads. The specifically advertising component of marketing strategy is concerned with getting the right message into the environment of the right people, so that they have a chance to take on board what the brand is claiming, and lodge in memory such claims, linked to the brand's name, to be available when a decision occasion arises.

Remarkably, where one would expect to find such issues discussed in the marketing literature, one finds ambiguity and omission.2 Textbook 2 See recent reviews of textbook treatment of market, market segmentation, target market, relevant universe

authors' presentation of "market," and "market segmentation," fails to treat, or to treat adequately, terms and concepts such as prospects, targets, targeted conditions, relevant universe. Indeed, the oxymoron, "target market," serves as unsatisfactory conceptual stopgap where adequate discussion is missing.

Our plan for this paper is first to discuss some problems that arise concerning the term, "target market." That is followed by a section in which we discuss one theoretical position that is helpful in addressing questions such as we pose at the outset. In subsequent sections, we consider how to implement the corresponding product and, then, advertising strategy. The paper concludes with a brief overview of the respects in which the approach we describe here differs from the received wisdom of mainstream textbooks.

Target Market: Oxymoron

As point of entry, let us ask the reader to reflect on the term, "target market." "Target" refers to the narrowest focus, and "market" to the broadest reach, of strategy for a particular venture. "Target market" is, in fact, a misreading or corruption of "market target," which, unproblematically, favors one of the poles of the oxymoron—it refers to the narrowest strategic focus—while using each word appropriately3. The phrase, "market target" properly expresses that a venture implicates a market—that is, the arena in which management competes for business—and, within a market, the targets that management has selected as the best use of its scarce resources for the venture. Collapsed into the oxymoron, target market, there is the implication that there is only one strategic decision, i.e., to select a market as the narrowest strategic focus for a venture. (Fennell & Saegert, 1998, 2000a, 2000b; Fennell, Saegert, & Hoover, 1999). 3 See definition of market segmentation, "The process of subdividing a market into distinct subsets of customers that behave in the same way or have similar needs. Each subset may conceivably be chosen as a market target to be reached with a distinct marketing strategy..." (American Marketing Association, 1985, emphasis added, pp. 165-6).

194

Running counter to the well known fact that wants within a market are heterogeneous, authors fail to make clear how a market is to be chosen so as to be homogeneous, as the term, target, requires or, alternatively, how management's brand can be fashioned to target, i.e., respond to the wants of, a market, given that wants within a market are heterogeneous. Moreover, if choosing a "target market" is management's strategic goal, what is the role of the reputedly pivotal term, segment?4

At present, considering twenty Principles of Marketing textbooks, none addresses the issues just mentioned. Moreover, there is wide diversity among authors in how they interpret the term, target market. Let us mention two broad categories of such diversity. Given that the term, target market, comprises contradictory terms, it is inevitable that authors tilt5 to one or the other, as illustrated in Figure 1, where an example of target tilt is shown to the left with market tilt to the right. As shown to the left, a product market is segmented and a segment is chosen as strategic focus, becoming the "target market" [7]6, and incurring the absurdity of a segment of a market becoming a market. The sequence here is segmenting before selecting a "target market." Alternatively, as shown to the right, a "target market" is selected, within which market segmentation identifies separate groups (segments) [15]. Here "target market" refers to the entity that management studies to further refine its strategy, i.e., market tilt. The sequence here entails segmenting after selecting a "target market." Considering the definitions of “target market” provided in each of the textbooks examined, Fennell, Saegert, and Hoover (1999) concluded that the authors divided equally into those whose definitions tilt to target and to market, respectively..

Secondly, it is illuminating to consider how authors describe the members of a target market. Wide diversity is again apparent, as illustrated in Figure 2. In fact, across all twenty texts, only two authors use the same terms to refer to individuals in a target market, i.e., "group of people" [4, 11]. Moreover, from that point on, these two authors differ as regards the remainder of their definitions 4 Consider the four textbooks whose authors state that a target market is a market segment, thus seeming to equate the whole with a part (Appendix). 5 Tilt refers only to the sense in which authors use the term, target market. we do not intend to suggest that the authors' view of marketing strategy favors one or the other pole of the oxymoron. Note, in Figure 1, bold face emphasis is added. 6 Authors are shown within square brackets, as listed in the Appendix.

(see Appendix). In addition to these two, three other authors refer to "people" or "individuals"[10, 12, 15]. Considering Figure 2 overall, one quarter of the authors use the term "market" or "market segment" to describe members of a target market (5);7 all but one of the remaining authors (14) refer to a group or set whose members they variously describe as customers (3[7, 8,16]), potential customers (2[3, 9]), buyers (2[13,17]), or consumers (2[1,2]); the remaining author [14] leaves the description at "group" without further specification. Such divergence among authors points to a lack of rigor that exists at this time within the marketing discipline. Even at the level of dictionary definition, words such as people, individuals, buyers, consumers, customers, and potential customers are not interchangeable.8 However, what is at issue here is not casual discourse, but the task of imparting to students a clear sense of the meaning of terms, as systematically used in the discipline that the textbooks are introducing. In a marketing context, it is not, in fact, a matter of indifference which terms an author selects among a set that includes, for example, individuals, customers, potential customers, consumers, group.9 In sum, concerning the collective to which authors apply the term, "target market," we find great diversity among the textbooks, as well as ambiguity and a lack of rigor.

While diversity among textbook authors is not necessarily to be decried, what we see here is lack of respect for the subject matter, rather than the outcome of theoretically divergent positions, responsibly argued. In the remainder of the present paper, we plan to draw out the implications of one theoretical position, so far as such implications relate to identifying prospects and communicating with targets. Two Strategic Tasks: Qualifying Prospects, Selecting Targets

Among issues raised by the anomalous "target market," two are highlighted above. The first 7 Number of authors is shown within round brackets. 8 Moreover, even within a single text, there may not be consistency. For example, target market is defined in the glossary as: "The group of individuals that the marketer wants to satisfy" and, elsewhere in the text, as, "The...customers the organization intends to satisfy." 9 It should also not be a matter of indifference whether or not an author qualifies the collective by a further term such as defined, homogeneous, or particular/specific. Here, authors are about equally divided between those who do and do not further qualify their chosen collective using the words shown in the first column of Figure 2.

195

of these arises from interpretative tilt, i.e., whether "target market" points to the outside reach (market tilt), or to the narrowest focus (target tilt), of management's interest for the focal venture. The second is manifest in the diversity, across authors, concerning how they designate members of a target market. Both problems trace to lack of clarity concerning two strategic tasks: Management's selecting the (1) universe and, (2) within that universe, a focus, for its venture. The first task addresses the question: Within the geographic area in which it does business, does management regard everyone as belonging in the market for the venture? Everyone, i.e., children and adults? People without funds or authority to use funds, or interest in engaging in marketplace exchanges? People who cannot qualify as predisposed to buy/use management’s product category? The second task arises having selected a universe. Does management plan to: serve all wants as found within that universe; challenge a major competitor on the specific wants within the product category that the competitor is targeting; serve wants that the people who experience them believe are well-satisfied by available offerings; wants that are within or outside management's current interest or competence, in its own estimation or as perceived in the marketplace? Such issues are in need of clarification. Elements of a Theoretical Perspective The principle that guides the answer that we present here is that of return on investment (ROI), i.e., getting the most out of scarce resources by putting those at one's disposal to best use. Management implements that principle in a context in which it largely relies on using impersonal media of communications to offer its wares. The offering, then, must speak for itself in a competitive context, i.e., people exposed to its advertising must find that the brand's message engages their attention by virtue of presenting information for which they are already seeking. They must view the claimed attributes to be worth parting with resources to acquire, and retain in memory the claimed attributes associated with the brand name, to be ready when next a decision occasion arises. For brevity here, skipping over issues of actually acquiring the brand, including finding it in stock, the brand in use must deliver the promised attributes at least to a degree that the trier does not exclude the brand from his or her consideration set. Moreover, the preceding events must happen sufficiently often and at sufficient price to yield satisfactory ROI. Product Strategy. In sum, considering product strategy, a producer in a competitive environment has to rely on predispositions of

audience members to select, and act on, information about the offering. Moreover, such predispositions sensitize to certain kinds of sensory and cognitive stimulation, hence the individuals who possess them are already preset to select certain kinds of information and not notice others. Self-interest suggests that management first exclude from consideration those individuals in the population whom it cannot regard as likely to spend resources to acquire or use an offering in its product category. Such individuals are not normally going to allocate their resources to process messages about the product category or, exceptionally, if they do, such messages will not result in a marketplace transaction. First, then, management carves out, from a naturally occurring population, individuals whom it can qualify as likely ready to allocate resources to the focal product category, i.e., its prospects.10 With an eye to ROI, compared with the excluded nonprospects, management's scarce resources are better spent investigating the sensitivities (wants) of prospective users of the product category, with a view to assessing whether or not it can respond to some region of such sensitivities profitably in a competitive environment.

To recap the problem of strategic development as it implicates the issue of the relevant universe: a first task of management is specifying the outer limits of its venture's potential relevance. Consider the earth's geographic area (Figure 3A). Management makes a selection—will its venture be restricted to a city, region, entire country, or beyond (Figure 3B)? Within its chosen geographic area, management finds a general population (Figure 3C). How does it determine who in that general population is worth allocating resources to study their wants concerning the focal venture? A minimal qualification is people with the means, authority, and inclination to spend (Figure 3D). Individuals so qualified must be further screened for predisposition to the domain of management's venture. As no venture is relevant to all tasks/interests that individuals pursue, we move from considering a universe of individuals, to considering a universe of behavior, i.e., the activities in which people engage. In thus identifying prospects within the general population, management states the qualifying criterion in behavioral terms, e.g., buyers/users of its product category, or individuals who engage in an activity or pursue a task or interest, for which some version of the product category could be used (Figure

10 The term, prospects, carries a meaning that is specific to the focal product category, i.e., prospects are individuals qualified as predisposed to buying/using some version of the focal product category.

196

3E). So far, we have considered universes of geography, individuals, behavior. A final dimension to consider is time. When management considers explicitly the time frame for which it is planning, it facilitates taking account of changes in relevant variables that may occur within a day, week, month, season, or year. Making its selections on the preceding dimensions, management has specified the outer boundaries of its universe for the venture in hand.11 To allow for intraindividual variability over time, it is helpful to consider the universe as comprising occasions for the focal activity, i.e., prospects multiplied by the number of occasions per individual for engaging in the focal activity during a calendar year12 (Figure3F). Management now investigates the qualitatively different kinds of want that prompt prospects to engage in the focal activity (Figure 3G). It is often not possible, technologically, or credibly, to fashion a brand that is responsive to the full range of such qualitatively diverse wants within the product category. Management's investigation, then, is designed to assess the likelihood that it can respond to some region of wants13 where it can put its resources to profitable use. Developing its strategy, management is identifying the wants that its offering will target, thus designating as targets the individuals who experience those wants (Figure 3H). In sum, management (1) Defines the a universe for its venture by selecting along many dimensions, including region of the earth, population within that region, behavioral universe of such individuals, and planning time horizon14; (2) Within that universe, it targets one or more kinds of want found among prospects. Advertising Strategy. We turn next to address the

11 Management further refines that universe in specifying its market. An additional most important consideration is the competitive environment that accompanies its choices on dimensions already discussed. If management does not want to engage particular competitors, it can remake its choices accordingly in arriving at the final dimensions of the market in which it plans to compete for a share of the exchanges that prospects are disposed to make. 12 For example, all occasions for feeding the dog, i.e., individuals who feed dogs times occasions, in the US during 2004. 13 Wants have been discussed as arising in qualitatively distinct kinds of demand creating condition (DCC). Seven motivational classes of such conditions have been described (e.g., Fennell 1997). 14 Other dimensions of market definition for a given venture include selecting the kind of communications medium to be used, broad price level (premium, regular, discount), and a method of exchange.

question of how the resulting strategy is communicated via advertising. Within a universe of prospects, management plans to communicate with targets, i.e., the prospects who experience the targeted conditions. How does management plan to get the weight reducing claims of a brand of dog food, which it has formulated for overweight dogs, to people who are concerned to do something about their dog's overweight? Consider all the media vehicles there are to choose among. To simplify the discussion here, assume that the producer of dog food is determined to use TV to the exclusion of radio and print media. The question remains, how does management plan, using TV, to announce to people concerned to control their dog's weight that a brand of dog food is available, specially tailored for overweight dogs15.

How is management to select a media vehicle to reach prospects who experience the conditions it has selected to target? In cases where it is possible directly to ascertain activities engaged in by members of the audience of media vehicles, management chooses vehicles whose audience members disproportionately engage in the focal activity (e.g., caring for a dog). Otherwise, in the more usual case where only demographic information about media vehicle audience members is provided, the best management can do is select media vehicles based on their containing, disproportionately, audience members whose demographic characteristics match those of individuals who engage in the focal activity (care for a dog) or buy/use the focal product category (buy/use dog food). As regards a domain of activity or product use, such as caring for a dog, or using dog food, respectively, contexts within the activity likely are not reported for audiences of media vehicles. Given that there could be at least half a dozen and, more likely, many more, distinct kind of context or condition within activity domain for one product category,16 it is clear that producers of data on the characteristics of media vehicle audiences cannot obtain and report so much information for each audience. In general, the best that management of a specific venture can do is to select a media vehicle whose audience members possess demographic

15 While, for illustrative purpose, we are posing the issue in the form of a specific instance, the example is intended to stand for the general case in which management first chooses a universe from a general population and then selects some region of that universe as the target of its strategy. 16 Recall, further, that the Simmons Markets and Media study (SMM) reports information for some hundreds of product categories.

197

characteristics that reflect, disproportionately, those of individuals who are prospects. It goes without saying that such audience contains individuals who are and are not prospects and, among prospects, individuals who do and do not experience the targeted conditions. The question remains, how does management plan to reach those prospects who experience the wants for which it has fashioned the brand? In regard, then, to ensuring that the brand's message is processed by people who experience the targeted conditions, all management can do is rely on targets to attend to a portrayal, in the ad, of the events or conditions that the targets experience for which the brand's attributes have been chosen. Individuals who are concerned about their dog's weight are likely to select a message that portrays such concern and offers a responsive brand (Fennell 1979). Beyond communicating the brand's message, of course the success of management's strategy depends, further, on targets' retaining the brand's message linked to the brand's name so that the information will be available when the relevant decision occasion arises. In sum, management's strategy for reaching targets with its advertising message is to select media vehicles whose characteristics reflect disproportionately those of prospects and to rely on the content of the message to engage the attention of those prospects who experience the targeted wants. Discussion

In this paper, we have described aspects of the fine grain of the way marketers must proceed in practice as they devise and implement strategy for a venture. In particular, we have focused on the tasks of carving a universe for the venture out of the general population in management's region of geographic space, and then devising a product strategy for competing within that universe. The first task involves specifying the outer limit of potential interest in the venture, and the second is concerned with selecting a particular sub area of wants within the product category where management's resources are put to best use. Similarly, management must use a two-step approach in planning to communicate the brand's message to targets. It reaches prospects, i.e., members of its market, by selecting media vehicles whose demographics disproportionately reflect those of its prospects, and relies on elements in the brand's message to engage the attention of targets, and communicate the brand's name and claims so that they are available to targets for use when the decision occasion arises.

The two-step nature of the task, both as regards developing, and communicating, management's response to wants, is not well represented in marketing textbooks. The deficit is one of omission and commission.

Omission. Defining market, textbook authors acknowledge that management excludes some individuals that it finds in the general population of its geographic area. Across various textbooks, authors mention that individuals without resources, authority to spend resources, and inclination to engage in marketplace exchange are not considered to belong in a market. Textbook authors, however, fail to discuss the remaining key criterion for inclusion in a market, i.e., predisposition toward buying/using some version (brand) of management's product category. Authors do not refer to the managerial practice of specifying the criterion by which population members are qualified as prospects for purposes of marketing planning, research, and analysis (Fennell and Saegert, 1998)17. There is no discussion of the systematic significance in marketing of defining the relevant universe in terms of predisposition to buy/use some version of management's product category (Fennell, Saegert and, Hoover 2002).

A further omission is reference to the different means management uses to reach, respectively, prospects, i.e., via selecting media vehicles to maximize an ad's exposure to prospects, and targets, i.e., by relying on elements within the ad's message selectively to engage the attention of prospects for whose wants the brand—and the ad's message—have been tailored. Commission. It is a widespread practice for textbook authors to use demographic characteristics to illustrate aspects of what they present as marketing analysis. As shown in the Appendix, one third of textbook authors illustrate "target market" with demographic examples, without, however, stating and discussing the basis on which particular demographic categories are selected. Similarly, many authors use demographic groups to illustrate market segments (Fennell and Saegert, 2000). In fact, demographic variables and other broad personal descriptors are only indirectly relevant to a marketing analysis. The role of demographic variables is as already described, i.e., a means of locating media vehicles whose audience members contain product category users disproportionately, so that management increases its chance of exposing its brand message in the presence of prospects. Targets, however, as distinct from prospects, are not well

17 Moreover, especially remarkable, textbooks in marketing research present the notion of a research universe, but illustrate it, unsystematically, sometimes by means of product use, sometimes in terms of demographic examples (Fennell et al., 2002)

198

described by demographic variables. Typically, events within product categories, such as the diverse wants found in the product use situation (Fennell 1978), are independent of demographic variables. Demographic descriptors characterize people too broadly, and lack relevant information, when it comes to the diverse conditions that pertain within domains of activity. Whether one's dog has a poor coat, is overweight, is regarded as the owner's infant, or guardian, or companion, whether the owner's self-image is that of an indulgent, or informed, dog owner—and similarly, for conditions that pertain within other domains of activity such as brushing teeth, doing the laundry—are found to cut across the broad-sweep characterizations, such as age groups, geographic location, income, of demographic classification. While there is a low to moderate relationship between demographic characteristics as independent variables, and domain of activity and product use as dependent variables, wants within a product category and brand use as dependent variables typically show little if any association with such broad descriptors. Accordingly, there is neither logical nor empirical ground for illustrating concepts such as market targets, or market segments, by means of demographic variables. At the stage of defining a market, some association is found between the activity that qualifies a person as a prospect and demographic variables; within the market as defined, there is no basis to expect that demographic variables are useful. Moreover, their broad categories lack content appropriate to reflecting diverse conditions within individual product categories and domains of activity. Accordingly, in many cases, textbook authors misleadingly illustrate a key term of marketing analysis, i.e., market segment, by examples using groups described by demographic characteristics. The oxymoron, "target market," is a further example of an error of commission, in that it purports to say something meaningful about marketing analysis and offers, instead, a nonsense term, inappropriate to the nature of marketing analysis and strategy. Conclusion

In this paper, we have addressed what the reader might reasonably expect to be a basic aspect of marketing analysis and implementation, i.e., how management selects, plans to serve, and communicate with, the targets of its strategy. As is apparent, the topic is not without its complexities and consequent interest. Moreover, it is remarkable to consider how the marketing discipline can have existed for so many decades lacking adequate treatment of such topics in its basic textbooks. As we have seen, errors of omission and commission have played their part. Current principles of marketing

texts, for example, fail to discuss: - The practice of specifying the outside limit of a

universe for a venture by screening individuals as predisposed to buy/use some version of the focal product category;

- The targets of marketing strategy as those prospects for whose wants, within the product category, management's brand has been fashioned;

- The means available to management to communicate with targets, i.e., selecting media vehicles so as to maximize exposure to prospects, and using elements within its brand's message to communicate with targets.

It is hoped that this paper has taken a step

toward repairing the deficit, permitting others to integrate such issues into their view of marketing as science and managerial function.

REFERENCES American Marketing Association (1985), "Market

Segmentation," in Peter D. Bennett (Ed.), Dictionary of Marketing Terms, 2nd edition, Chicago: American Marketing Association (pp. 165-6).

Assael, Henry (1990), Marketing: Principles and Strategy, Chicago: The Dryden Press.Bearden, William O., Thomas N. Ingram and Raymond LaForge (1995), Marketing: Principles and Perspectives, Chicago: Irwin.

Berkowitz, Eric N., Roger A. Kerin, Steven W. Hartley and William Rudelius (1997), Marketing, 5th Edition, Chicago: Irwin.

Boone, Louis E. and David L. Kurz (1995), Contemporary Marketing, 8th Edition, Fort Worth, The Dryden Press.

Bovée, Courtland L., Michael J. Houston, and John V. Thill (1995), Marketing, 2nd Edition, New York: McGraw-Hill, Inc.

Churchill, Gilbert A. Jr. and J. Paul Peter (1998), Marketing: Creating Value for Customers, 2nd Edition, Burr Ridge, Illinois; Irwin.

Etzel, Michael J., Bruce J. Walker and William J. Stanton (1997), “Marketing, 11th Edition, New York: McGraw-Hill

Evans, Joel R. and Barry Berman (1997), Marketing, 7th Edition, Upper Saddle River, NJ, Prentice Hall.

Fennell, G. (1978), “Consumers' Perceptions of the Product-Use Situation,” Journal of Marketing, 42, (April), 38-47.

199

Fennell, G. (1979), “Attention Engagement,” In J.H. Leigh & C.R. Martin, Jr.,(Eds.), Current Issues & Research in Advertising. Ann Arbor: University of Michigan.

Fennell, G. (1997). Value and Values: Relevance to Advertising. In L. Kahle and L. Chiagouris (Eds.), Values, Lifestyles, and Psychographics. Hillsdale, NJ: Erlbaum.

Fennell, G. and Joel Saegert (1998), “Implications of the Marketing Concept: What the Textbooks Fail to State,” in Proceedings of the Society for Consumer Psychology, eds. Karen Machleit and Meg Campbell. Washington, D.C.: Society for Consumer Psychology (Division 23), American Psychological Association, 65-74.

Fennell, G. and Joel Saegert (2000), “Market Segmentation: Textbook Definitions,” In John P. Workman and William D. Perreault, (Eds.), Marketing Theory and Applications, Vol. 1I, Chicago: American Marketing Association, 136-145

Fennell, G., Joel Saegert and Robert J. Hoover (1999), “An Oxymoron in Principles of Marketing and International Marketing Textbooks,” , in K. Fatemi and S. Nichols (Eds.), Proceedings of the International Trade and Finance Association, 471-492.

Fennell, G., Saegert, Joel and Hoover, Robert (2002), “Marketing’s Universe: Implicit Presence in Marketing Research Texts,” in K.R. Evans and L.K. Scheer (Eds), Proceedings of the American Marketing Association Winter Educator’s Conference. Chicago: American Marketing Association, 188-195.

Harrell, Gilbert D. and Gary L. Frazier (1999), Marketing: Connecting with Customers, Upper Saddle River, NJ: Prentice-Hall Inc.

Husted, Stewart W., Dale L Varble and James R. Lowry (1989), Principles of Modern Marketing. Boston: Allyn and Bacon.

Keegan, Warren J., Sandra E. Moriarty and Thomas, R. Duncan (1995), Marketing, 2nd Edition, Englewood Cliffs, NJ: Prentice Hall.

Kinnear, Thomas C., Kenneth L. Bernhardt and Kathleen A. Krentler (1995), Principles of Marketing, 4th Edition, New York: HarperCollins

Kotler, Philip and Gary Armstrong (1997), Marketing: an Introduction, 4th Edition, Upper Saddle River, NJ: Prentice Hall.

Lamb, Charles W. Jr., Joseph F. Hair and Carl McDaniel (1996), Marketing, 3rd Edition, Cincinnati, OH: South-Western College Publishing.

Nichels, William G. and Marian Burk Wood (1997), Marketing: Relationships, Quality, and Value, New York: Worth Publishers.

Perreault, William D., Jr. and E. Jerome McCarthy (1996), Basic Marketing: a Global-Managerial Approach, 12th Edition. Chicago: Irwin.

Pride, William M., and O.C. Ferrell (1997), Marketing: Concepts and Strategies, 10th Edition, Boston: Houghton Mifflin Company.

Semenik, Richard J., and Gary J. Bamossy (1995), Principles of Marketing: A Global Perspective, 2nd Edition. Cincinnati, Ohio: South-Western College Publishing.

Smith, W.R. (1956), “Product Differentiation and Market Segmentation as Alternative Marketing Strategies,” Journal of Marketing, 21(July), 3-8.

Solomon, Michael R., and Elnora Stuart (1997), Marketing: Real People, Real Choices, Upper Saddle River, NJ: Prentice Hall.

Zikmund, William G. and Michael D’Amico (1996), Marketing, 5th Edition, Minneapolis/St. Paul: West Publishing Company.

For further information contact: Dr. Joel Saegert, Department Of Marketing, The University Of Texas At San Antonio San Antonio, TX 78249. Phone: 210-458-5375; Email: jsaegert@utsa.edu

200

Figure 1. EXAMPLES OF TILT: TARGET V. MARKET

TILT TO TARGET [ 7, p. 158] TILT TO MARKET [15, p. 43] Whether it is large or small, the specific group of customers (people or organizations) for whom the seller designs a particular marketing mix is a target market. Before target markets can be selected, however, they must be identified and described. This process is called market segmentation...a process of dividing the total market for a good or service into several smaller, internally homogeneous groups.

A target market is the group of people or organizations whose needs the marketer's products are specifically designed to satisfy. Target marketing narrows the organization's focus to an appropriate set of potential customers whose needs can be profitably met...Market segmentation identifies separate groups (segments) of people or organizations within a broad target market, each with similar needs, characteristics, or behaviors.

Figure 2. PEOPLE DIMENSION IN AUTHORS' DEFINITIONS OF TARGET MARKET QUALIFYING WORD COLLECTIVE

Particular (1) Specific (1)

[6] [20]

Market (1) Market Segment(s) (4)

[5] [6,18,19,20]

Specific (3) Defined (1) Particular (1) Fairly homogeneous (1) None (9)

[3,12,17] [2]

[8] [16]

[remainder]

Group(s)/ Set of: Buyers (2) Consumers (2) Customers (3)

Individuals/people (5) Potential customers (2) Group (unspecified) (1)

[13,17]

[1,2] [7,8,1]

[4,10,11,12,15] [3, 9]

[14] Note: Number of instances shown in parentheses; source shown in square brackets

201

Figure 3 DEFINING PROSPECTS and SELECTING TARGETS

A. TOTAL EARTH AREA

B. SELECTED GEOGRAPHIC DOMAIN C. GENERAL POPULATION w/i SELECTED GEOGRAPHIC DOMAIN

iiiii iiiii iiiii iiiii

D. EXCHANGE-COMPETENT INDIVIDUALS w/i GENERAL POPULATION

iiii i iii iii i i ii

E. PROSPECTS for the VENTURE w/i EXCHANGE-COMPETENT

i ii i i i ii i

F. OCCASIONS for FOCAL ACTIVITY (PROSPECTS x OCCASIONS )

G. WANTS PROMPTING INDIVIDUAL OCCASIONS for FOCAL ACTIVITY

H. WANTS SELECTED as FOCUS of MARKETING STRATEGY

______________________________________________________________________________

202

Appendix AUTHORS' DEFINITIONS OF TARGET MARKET

AUTHOR PEOPLE DIMENSION STRATEGIC RELEVANCE 1. Assael A group of consumers with similar needs that can be identified and appealed to by a

specific product or product line 2. Bearden

{D} A defined group of consumers or organizations

with whom a firm wants to create marketing exchanges

3. Berkowitz {D}

One or more specific groups of potential customers

toward which an organization directs its marketing program

4. Boone {D}

Group of people

toward whom a firm markets its goods, or ideas with a strategy designed to satisfy their specific needs and preferences

5. Bovée The market you’ve selected as the focus of your marketing program

it covers the potential customers you think are most likely to need or want your product

6. Churchill TM=MS {D}

The particular market segment that an organization selects to serve is called target market

7. Etzel A group of customers (people or organizations)

or whom a seller designs a particular marketing mix

8. Evans

The particular group(s) of customers

a firm proposes to serve, or whose needs it proposes to satisfy, with a particular marketing program

9. Harrell {D}

A group of potential customers with similar characteristics that a marketer is trying to satisfy better than the competition

10. Husted The group of individuals that the marketer wants to satisfy 11.Keegan The group of people most likely to buy or influence the purchase of a given product

12Kinnear {D}

A specific group of people for whom a marketer develops a marketing mix

13. Kotler A set of buyers sharing common needs or characteristics that the company decides to serve

14. Lamb Group

for which an organization designs, implements, and maintains a marketing mix intended to meet the needs of that group, resulting in mutually satisfying exchanges

15. Nickels The group of people or organizations

whose needs a product is specifically designed to satisfy

16. Perreault A fairly homogeneous (similar) group of customers

to whom a company wishes to appeal

17. Pride A specific group of buyers on whose needs and wants a company focuses its marketing efforts

18. Semenik TM=MS {D}

Target markets are market segments

to which a firm has decided to direct its marketing efforts—segment of people to which a firm hopes to sell satisfying products

19. Solomon TM=MS

The market segment(s) on which an organization focuses its marketing plan and toward which it directs its marketing effort

20. Zikmund TM=MS

A specific market segment toward which an organization aims its marketing plan

Source: Textbook glossaries, except two cases where there is no glossary or entry [13, p. 219; 18, pp. 50-51]. Name of first author only shown here; complete citations in References. D in curly brackets: authors illustrate the collective by demographic categories. TM=MS: authors say that target market is a market segment.

203

List of Contributors (Indexed by Session)

Contributor Session Jennifer Aaker, Stanford University 7.1 Nidhi Agarwal, New York University 7.1 Pankaj Aggarwal, University of Toronto 2.3 On Amir, Yale University 1.3 Jennifer J. Argo, University of Alberta 5.2 Dan Ariely, Massachusetts Institute of Technology 1.3, 5.3, 8.3 Rita A. Balaban, College of Charleston 5.2 Brad Barber, University of California, Davis 1.1 Lawrence W. Barsalou, Emory University 5.1 Debra Z. Basil, University of Lethbridge 6.2, 7.1 Michael D. Basil, University of Lethbridge 6.2, 7.1 Mike Bastin, University College Northampton 1.2 Colleen Bee, University of Oregon 4.2, 6.4 Julia E. Blose, College of Charleston 5.2 Carolyn Bonifield, University of Vermont 4.2 Maja Makovec Brenčič, University of Ljubljana 6.2 Susan M. Broniarczyk, University of Virginia (visiting) 5.3, 8.1 Liesbeth Bulkmans, University of Antwerp 6.4 Kurt A. Carlson, Duke University 2.2 Ziv Carmon, INSEAD 1.3 Tanya Chartrand, Duke University 2.2 Haipeng Chen, University of Miami 4.3 Alexander Chernev, Northwestern University 8.1 Ravi Chitturi, University of Texas at Austin 7.3 Catherine A. Cole, University of Iowa 4.2, 2.4 Kim P. Corfman, New York University 7.3 Robin A. Coulter, University of Connecticut 7.2 Maria Cronley, Miami University of Ohio 2.1, 3.4, 5.3 Sandor Czellar, University of Lausanne 8.2 Darren W. Dahl, University of British Columbia P.06 Yang Dai, University of California, Los Angeles P.01 Vassilis Dalakas, Berry College 6.2 Peter R. Darke, University of British Columbia P.06 Prabu David, The Ohio State University P.01 Edith F. Davidson, University of Tennessee P.02 Jan De Houwer, Ghent University 6.2 Gustavo de Mello, University of Southern California 3.4 Patrick De Pelsmacker, University of Antwerp 6.2, 6.4 Stephanie Dellande, Chapman University 7.1 Kalpesh K. Desai, State University of New York, Buffalo 8.1 Ravi Dhar, Yale University 1.1, 6.1 Aimee Drolet, University of California, Los Angeles 2.4 Laurette Dubé, McGill University 3.1 Tine Faseur, Ghent University 6.4 Alexander (Sasha) Fedorikhin, University of Southern California 6.3 Raymond Fisman, Columbia University 8.3 Gavan J. Fitzsimons, Duke University 6.3 Shane Frederick, Massachusetts Institute of Technology 6.1 Jeana Frost, Massachusetts Institute of Technology 8.3 Ellen Garbarino, Case Western Reserve University 5.1 Nitika Garg, University of Pittsburgh 6.3

204

Contributor Session Andrew D. Gershoff, Columbia University 5.3 Maggie Geuens, Ghent University 6.2 Bryan Gibson, Central Michigan University 2.3 Jennifer Gregan-Paxton, University of Delaware 2.1 Reetika Gupta, Baruch College P.03 Zeynep Gurhan-Canli, University of Michigan P.10 Nicole Hafey, Central Queensland University 7.2 Steve E. Hoeffler, UNC – Chapel Hill 2.1, 2.2 Susan Hogan, Emory University 5.3 Jianwei Hou, University of Mississippi P.04 Christopher K. Hsee, University of Chicago 6.1 Joel Huber, Duke University 2.2 Jeff Inman, University of Pittsburgh 6.3 Paul A. Jackson, University of Southern California P.09 Wim Janssens, University of Antwerp 6.4 Gita V. Johar, Columbia University 3.4, 4.1 Barbara E. Kahn, University of Pennsylvania 8.1 Frank Kardes, University of Cincinnati 2.1 Payal Karumbiah, Ohio State University 3.2 James Kellaris, University of Cincinnati 2.1 Tina Kiesler, California State University, Northridge 2.3 H. Shanker Krishnan, Indiana University 4.2 Monica LaBarge, University of Oregon 4.2 Raphaëlle Lambert-Pandraud, Negocia 2.4 Ying Lan, University of Mississippi P.04 Michel Laroche, Concordia University 8.2 Gilles Laurent, HEC 2.4 Sharmistha Law, University of Toronto 2.3 Jordan L. LeBel, Concordia University 3.1 Angela Y. Lee, Northwestern University 4.3 Leonard Lee, Massachusetts Institute of Technology 5.3 Sooyeon Nikki Lee, New York University 7.3 Irwin Levin, University of Iowa 2.2 Annalisa Libermann, Miami University P.05 Patricia W. Linville, Duke University 2.2 Mary Frances Luce, University of Pennsylvania 8.1 David Luna, Baruch College 6.4 Deborah MacInnis 3.4 Rhonda W. Mack, College of Charleston 5.2 Robert Madrigal, University of Oregon 4.2, 6.4 Vijay Mahajan, University of Texas at Austin 7.3 Kelley J. Main, University of British Columbia P.06, 5.2 Selin A. Malkoc, University of North Carolina 3.2, 3.2 Alan J. Malter, University of Arizona 5.1 Susan Mantel, IUPUI 1.3 Ereni Markos, Emmerson College P.08 Margaret G. Meloy, Penn State University 2.2 Norma A. Mendoza, University of Arkansas 1.3 Geeta Menon, New York University 7.1 Tracy Meyer, University of Cincinnati 2.1 Tom Meyvis, New York University 3.3 Camelia Micu, University of Connecticut 7.2 Ashwani Monga, University of Minnesota 4.3 Kent Monroe, University of Illinois 3.3

205

Contributor Session Andrea C. Morales, University of Southern California 6.3 Vicki Morwitz, New York University 3.3 Mehdi Mourali, Concordia University 8.2 Ruth Mugge, Delft University of Technology, Netherlands 8.2 Anirban Mukhopadhyay, Columbia University 3.4 Sucheta Nadkarni, University of Nebraska - Lincoln P.03 Xiaoli Nan, University of Minnesota 3.2 Sabrina Neeley, Miami University of Ohio P.05, 3.4, 4.2 Sharon Ng, University of Minnesota 4.3 Jesper Nielsen, University of Arizona 2.2 Nathan Novemsky, Yale University 6.1 Stephen M. Nowlis, Arizona State University 3.1 Prashanth Nyer, Chapman University 7.1 Terrance Odean, University of California, Berkeley 1.1 Christine Page, Skidmore College 4.2 Catherine Paquet, McGill University 3.1 Kiwan Park, University of Michigan 2.3 Vanessa Patrick, University of Southern California P.09, 6.3 Richard E. Petty, Ohio State University 3.4, 4.1 Frank Pons, University of San Diego 8.2 Steven Posavac, University of Rochester 2.1 Linda L. Price, University of Nebraska 7.2 Joseph R. Priester, University of Michigan 2.3 Priya Raghubir, New York University 7.3 Priya Raghubir, University of California, Berkeley 5.2 Raj Raghunathan, University of Texas at Austin 7.3 Sekar Raju, State University of New York at Buffalo 1.2 Seshan Ramaswami, Singapore Management University 8.1 Akshay R. Rao, University of Minnesota 4.3 Aparna Rao, University of Michigan P.10 S. (Ratti) Ratneshwar, University of Missouri - Columbia 8.1 Anne Roggeveen, Babson College 4.1 José Antonio Rosa, Case Western Reserve University 5.1 Gregory M. Rose, University of Washington - Tacoma 7.1 Derek Rucker, Ohio State University 3.4 J. Edward Russo, Cornell University 2.2 Julie A. Ruth, Rutgers University 1.2 Mototaka Sakashita, Kobe University, Japan P.11 Sela Sar, University of Minnesota 3.2 Ann E. Schlosser, University of Washington 1.3 Jan P.L. Schoormans, Delft University of Technology, Netherlands 8.2 David W. Schumann, University of Tennessee P.02 Norbert Schwarz, University of Michigan 2.4, 6.1 Hendrik N.J. Schifferstein,Delft University of Technology,Netherlands 8.2 Sheena Sethi-Iyengar, Columbia University 8.3 Eldar Shafir, Princeton University 6.1 Baba Shiv, University of Iowa 2.2, 3.1 Aviv Shoham, University of Haifa 7.1 Jennifer C. Siemens, University of Dayton 1.3 Itamar Simonson, Stanford University 6.1, 8.3 Lisa Sinclair, University of Winnipeg P.07 Ian Skurnik, University of Toronto 2.4 Douglas M. Stayman, Cornell University 4.1 Michal Strahilevitz, University of Arizona 1.1, 7.3

206

Contributor Session Tao Sun, Emerson College 7.2 Dobrina G. Teifel, Cornell University 4.1 Manoj Thomas, New York University 3.3 Tilottama G. Chowdhury, University of Connecticut 8.1 Zakary L. Tormala, Indiana University 4.1 Michel Tuan Pham, Columbia University 1.1 Canan Ulu, Duke University 3.2 H. Rao Unnava. Ohio State University 1.2, 3.2 Nancy Upton (Puccinelli), Suffolk University P.08 Wouter Vanhouche, University of Florida 2.1 Stijn M.J. van Osselaer, Erasmus University 2.1 Delphine Vantomme, Ghent University 6.2 Ana Velenzuela, San Francisco State University 5.2 Nicole L. Votolato, Ohio State University 1.2, 3.2 Jing (Alice) Wang, Northwestern University 4.3 Brian Wansink, University of Illinois at Urbana-Champaign 6.3 Tony Ward, Central Queensland University 7.2 Kittichai Watchravesringkan, University of Arizona P.12 Deanne Weber, Porter Novelli International 7.1 Michael Wiles, Indiana University P.13 Patti Williams, University of Pennsylvania 2.4 Guohua Mark Wu, San Jose State University 7.2 Ling-Ling Wu, National Taiwan University 5.1 Lan Xia, Bentley College 3.3 Jinhong Xie, University of Florida 3.3 Xiaojing Yang, Indiana University 4.2 Carolyn Yoon, University of Michigan 2.4 Anne York, University of North Carolina 1.2 Seounmi Youn, Emerson College 7.2 Jennifer Yurchisin, University of Arizona P.12 Gal Zauberman, University of North Carolina 3.2 Jiao Zhang, University of Chicago 6.1 Min Zhao, University of North Carolina 2.1 Rongrong Zhou, Hong Kong University of Science and Technology 1.1 Ning Zhu, University of California, Davis 1.1