xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam 2200421Ka 4500
controlfield tag 001 001681093
006 m d s
007 cr mnu uuuuu
008 051229s2005 flu sbm s000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0001021
Carrillat, Francois Anthony.
The effect of perceived entitativity on implicit image transfer in multiple sponsorships
h [electronic resource] /
by Francois Anthony Carrillat.
[Tampa, Fla.] :
b University of South Florida,
Thesis (Ph.D.)--University of South Florida, 2005.
Includes bibliographical references.
Text (Electronic thesis) in PDF format.
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
Title from PDF of title page.
Document formatted into pages; contains 199 pages.
ABSTRACT: This dissertation proposes that in the case of multiple sponsorships (i.e., brands sponsoring concomitantly the same event), the group constituted by the sponsoring brands and the sponsored event will be perceived as an entity; a phenomenon that Campbell (1958) called entitativity. The extent to which a group of brands and a sponsored event is seen as being entitative will result in stereotypic processing of the group members (Brewer and Harasty 1996). Information about an entitative group is abstracted and used to form judgments about every group member (McConnell, Sherman, and Hamilton 1997). Characteristics tied to one brand or to the event will become associated to the other brands due to category-based information processing (Fiske and Neuberg 1990).As a result, images associated with a brand or an event that belongs to an entitative group will be transferred to other brands of that group due to stereotyping.Image transfer effects were investigated through an experiment. Image transfer in sponsorship occurs primarily at an implicit level because sponsorship messages are subtle (Pham and Vanhuele 1997). As a consequence, the savings in relearning paradigm (Ebbinghaus 1885/1964) was the methodology used. It allows investigating implicit memory by comparing the recall of paired-associations between brands and image-traits across a multiple sponsorship and a no sponsorship condition. The findings confirmed that the event and the concomitant sponsoring brands were perceived as an entitative group, which resulted in an implicit transfer of image among the brands (Brand Image Transfer, BIT) as well as from the event to the brands (Event Image Transfer, EIT).
Adviser: Dr. Paul J. Solomon.
t USF Electronic Theses and Dissertations.
The Effect of Perceived Entitativity on Implic it Image Transfer in Multiple Sponsorships by Franois Anthony Carrillat A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Marketing College of Business Administration University of South Florida Major Professor: Paul J. Solomon, Ph.D. Michael T. Brannick, Ph.D. Eric G. Harris, Ph.D. Barbara A. Lafferty, Ph.D. Date of Approval: March 25, 2005 Keywords: Brand, Similarity, Leverage, Equity, Memory Copyright 2005, Franois A. Carrillat
Dedication This dissertation and degree are dedicated to my family and to Dana. My father Claude and my mother Eliane offered me their unconditional love and support during this doctoral endeavor and they did everything they could to make this experience successful. This dissertation is also dedi cated to the memory of my late grandmother Charlotte who had an important role in my life. Dana, I want to thank you for your love and your belief in me. Your sacrifices and your patience carried me through. Y our companionship and your support have made this journey complete.
Acknowledgments This work would have never been done without Dr. Paul Solomon, my dissertation chair. I am extrem ely thankful to him for several reasons. First, he believed in this dissertation and saw its potential contribution from the beginning. Second, he spent countless hours refini ng the conceptualization, rearranging the design, and reviewing the manuscript. Finally, he was a very good mentor not only when supervising this dissertation but also by guiding me on the path of academe. I also want to acknowledge my committee members, Dr. Michael Brannick, Dr. Er ic Harris, and Dr. Barbara Lafferty, who played a critical role in improving this dissertati on at every step of the research. In addition, I acknowledge th e role of Dr. William Locander who had an important influence on my scholarship a nd whose guidance and advices have been critical. I want to thank Dr. Greg Marshall who contributed extensively to my academic progress as well as Dr. Jean-Louis Chandon whos e dedication to research inspired me. I also want to acknowledge Dr Walter Borman, Dr. Michael Coovert, Dr. James Lee, Dr. Loyd Pettegrew and Dr. James Stock who contributed to my intellectual development. The support of my colleagues Cynthia Cano, Fernando Jaramillo, Jay Mulki and Bobby Riggle is also acknowledged; they were alwa ys there for exchanging ideas, helping me, or engaging in research projects. Finall y, a special thank you for Wendy Jennings who helped me with editing this manuscript and made sure I got through the administrative part of this journey.
i Table of Contents List of Tables iv List of Figures vi Abstract viii Chapter One: Introduction 1 Sponsorship and the Brand Leverage Model 3 Brand Knowledge 3 Image Transfer 3 Multiple Sponsorships 4 Statement of Research Problem 5 The Savings in Relearning Paradigm 6 Intended Contribution 7 An Entitativity-Based Alternative to the Brand Leverage Model 8 Implicit Knowledge Transfer 10 Research Method Employed 11 Chapter Two: Literature Review 13 Sponsorship Research 14 The Sponsorship Medium 14 The Effect of Sponsorship Communication 15 Theories Associated with Sponsorship Effects 15 Multiple Sponsorships as a Type of Brand Association 18 Brand Extensions 18 Celebrity Endorsement 20 The Ambiguous Role of Similarity 22 The Traditional View 22 A New Stance on Similarity 23 The Role of Entitativity in Multiple Sponsorships 24 The Antecedents of Entitativity 27 Individual Differences 27 Group Characteristics 29 The Entitativity of the Concomitant Sponsoring Brands and the Event 30 The Influence of Entitativity on Impression Formation 31 On Line vs. Memory-Based Judgment 32 Category-Based vs. Individuating Information Processing 34
ii Stereotypic vs. Exemplar Processing 35 The Cognitive Busyness of Entitativity 36 Group Essentialization 37 The Impact of Concept Similarity on Entitativity 39 Categorization and Conceptual Coherence 39 Brand-Concept Consistency 41 Savings in Relearning as a Measure of Implicit Memory 43 Abstraction as Implicit Knowledge Creation 44 Implicit Memory 45 Savings in Relearning 46 Implicit Image Transfer 48 Trait-Transference 49 Summary and Hypotheses Development 50 Outcome Hypotheses 50 Process Hypotheses 54 Chapter Three: Methodology 60 Overview 60 Respondents and Experimental Design 61 Material 62 Pretest 63 Procedure 65 Sponsorships Manipulation 67 Target Brands Exposure 67 Foil Brands Exposure 68 Paired-Associations Task 68 Filler Task 71 Cued-Recall Task 71 Tag-Line Recognition Task 71 Manipulation Checks and Covariate 72 Chapter Four: Results 74 Manipulation Checks 74 Outcome Hypotheses: Implicit Br and and Event Image Transfer 76 Additional Analysis 79 Brand Image Transfer Analysis 82 Event Image Transfer Analysis 82 Process Hypotheses: Brand Image Reinforcement (BIR) 83 Process Hypotheses: BIR vs. BIT and EIT Trials for Sponsors with Similar Brand-Concepts 86 Additional Analysis 89 Process Hypotheses: BIR vs. BIT and EIT Trials for Sponsors with Dissimilar Brand-Concepts 92 Additional Analysis 96 Process Hypotheses: Tag-Line Recognition Data 99
iii Additional Analysis 100 Mediation Analysis 102 Strength of BIT vs. Strength of EIT 105 Chapter Five: Discussion and Conclusion 108 Discussion of Experimental Results 109 Outcome Hypotheses 109 Evidence of Brand Image and Event Image Transfer 109 The Impact of Brand-Concept Similarity 110 The Role of Individual Differences 110 Other Unexpected Results 110 Process Hypotheses 114 Brand Image Reinforcement 114 The Impact of Multiple Sponsorships on BIR 114 Comparison of BIR Trials versus BIT and EIT Trials 116 The Impact of Individual Differences 117 Tag-Line Recognition 118 The Entitativity Model of Image Transfer in Multiple Sponsorships 119 Comparing the Importance of BIT with the Importance of EIT 120 Contribution of the Savings in Relearning Paradigm to the Marketing Literature 122 Managerial Implications 123 Limitations and Further Research 126 Conclusion 131 References 132 Appendices 147 Appendix 1: Summary of Empirical Studies on Sponsorship 148 Appendix 2: Experimental Stimuli Used in the Study 156 Appendix 3: Scale Items 165 Appendix 4: Study Experimental Design 171 Appendix 5: Results of Experimental Stimuli Pretest 143 Appendix 6: The Multiple Sponsorships Manipulation 175 Appendix 7: Manipulation Checks Results 176 Appendix 8: ANOVA and ANCOVA Tables 178 Appendix 9: A-priori C ontrast Tests Tables 183 About the Author End Page
iv List of Tables Table 1 Concept Similarity and Image of the Brands Used in the Experiment 6 Table 2 Mediation of the Impact of Multiple Sponsorships on Image Transfer by a Group Impression 105 Table 5-1 Image and Relevance Ratings of the Target Brands and the Event 173 Table 5-2 Image and Relevance Ratings of the Foil Brands 174 Table 7-1 Similarity Ratings of the Ta rget Brands with the Sport Concept 176 Table 7-2 Image Ratings of the Target Brands 177 Table 8-1 ANCOVA Table for the Outcome Hypotheses H1a to H2b 178 Table 8-2 ANOVA Table for Brand Imag e Reinforcement (H3a to H4) 179 Table 8-3 ANCOVA Table for the Process Hypotheses H5a to H5d 180 Table 8-4 ANCOVA Table for the Process Hypotheses H6a to H6d 181 Table 8-5 ANCOVA Table for the Process Hypotheses H7a and H7b 182 Table 9-1 Contrasts Tests for the Outcome Hypotheses H1a to H2b 183 Table 9-2 Contrast Test for the Need for Closure Analysis of the Outcome Hypotheses 183 Table 9-3 Contrast Tests For Brand Imag e Reinforcement Analysis (H3a to H4) 185 Table 9-4 Contrast Tests for the Processes Hypotheses H5a to H5d 185 Table 9-5 Contrast Tests for the Need for Closure Analysis of the Process Hypotheses H5a to H5d 186 Table 9-6 Contrast Tests for Process Hypotheses H6a to H6d 187 Table 9-7 Contrast Tests for the Need for Closure Analysis of the Process Hypotheses H6a to H6d 188
v Table 9-8 Contrast Tests for th e Process Hypotheses H7a and H7b 189 Table 9-9 Contrast Tests for the Need for Closure Analysis of the Process Hypotheses H7a and H7b 189
vi List of Figures Figure 1 The Brand Leverage Model 4 Figure 2 An Entitativity Model of Implicit Image Transfer in Multiple Sponsorships 10 Figure 3 Antecedents and Consequences of Group Entitativity 26 Figure 4 Methodological Steps of the Experimental Procedure 66 Figure 5 Examples of Exposure, Memorization and Cued-Recall Tasks in the Brand Image Related and the Event Image Related Conditions 72 Figure 5a Implicit Brand Image Transfer 76 Figure 5b Implicit Event Image Transfer 77 Figure 6a Implicit Brand Image Transfer: Individuals with a Low Need for Closure 80 Figure 6b Implicit Brand Image Transfer: Individuals with a High Need for Closure 80 Figure 7a Implicit Event Image Transfer: Individuals with a Low Need for Closure 81 Figure 7b Implicit Event Image Transfer: Individuals with a High Need for Closure 81 Figure 8 Brand Image Reinforcement 84 Figure 9a Consistent vs. Inconsistent Paired-Associations for Brands with Similar Concepts: BIR vs. BIT 87 Figure 9b Consistent vs. Inconsistent Paired-Associations for Brands with Similar Concepts: BIR vs. EIT 88
vii Figure 10a BIR vs. BIT for Sponsors with Similar Brand-Concepts: Low Need for Closure 90 Figure 10b BIR vs. BIT for Sponsors with Similar Brand-Concepts: High Need for Closure 91 Figure 11a BIR vs. EIT for Sponsors with Similar Brand-Concepts: Low Need for Closure 91 Figure 11b BIR vs. EIT for Sponsors with Similar Brand-Concepts: High Need for Closure 92 Figure 12a Consistent versus Inconsis tent Paired-Associations for Brands with Dissimilar Concepts: BIR vs. BIT 94 Figure 12b Consistent versus Inconsistent Paired-Associations for Brands with Dissimilar Concepts: BIR vs. EIT 94 Figure 13a BIR vs. BIT for Sponsors with Dissimilar Brand-Concepts: Low Need for Closure 97 Figure 13b BIR vs. BIT for Sponsors with Dissimilar Brand-Concepts: High Need for Closure 97 Figure 14a BIR vs. EIT for Sponsors with Dissimilar Brand-Concepts: Low Need for Closure 98 Figure 14b BIR vs. EIT for Sponsors with Dissimilar Brand-Concepts: High Need for Closure 98 Figure 15 Tag-Line Recognition 99 Figure 16a Tag-Line Recognition By Level of Need for Closure: Low Need for Closure 102 Figure 16b Tag-Line Recognition By Level of Need for Closure: High Need for Closure 102 Figure 17 Multiple Sponsorships Decision rules 125
viii The Impact of Perceived Entitativity on Implic it Image Transfer in Multiple Sponsorships Franois A. Carrillat ABSTRACT This dissertation proposes that in the cas e of multiple sponsorships (i.e., brands sponsoring concomitantly the same event) the group constitute d by the sponsoring brands and the sponsored event will be perceived as an entity; a phenomenon that Campbell (1958) called entitati vity. The extent to whic h a group of brands and a sponsored event is seen as being entitative will result in stereot ypic processing of the group members (Brewer and Harasty 1996). In formation about an entitative group is abstracted and used to form judgments about every group member (McConnell, Sherman, and Hamilton 1997). Characteristics tied to one brand or to the event will become associated to the other brands due to category-based information processing (Fiske and Neuberg 1990). As a result, images associated with a brand or an ev ent that belongs to an entitative group will be tran sferred to other bra nds of that group due to stereotyping. Image transfer effects were investigated through an experiment. Image transfer in sponsorship occurs primarily at an implicit le vel because sponsorship messages are subtle (Pham and Vanhuele 1997). As a consequen ce, the savings in relearning paradigm (Ebbinghaus 1885/1964) was the methodology used. It allows investigating implicit memory by comparing the recall of paired-asso ciations between brands and image-traits across a multiple sponsorship and a no sponsorship condition. The findings confirmed
ix that the event and the concomitant sponsoring brands were perceived as an entitative group, which resulted in an implicit transfer of image among the brands ( Brand Image Transfer BIT) as well as from the event to the brands ( Event Image Transfer EIT). These transfer effects were moderated by the brand-concept associat ed with the sponsors considered. BIT was only found for sponsors wi th a similar brand-concept (i.e., sport) whereas EIT was only found for sponsors with a dissimilar brand-concept (i.e., no sport). Further analyses confirmed that th ese phenomena of implicit transfer of image were due to a category-based as opposed to an individuating processing of information. Due to high entitativity, perceivers relied on a group impression to process information. As a result, the group of brands and th e event were seen as interchangeable.
1 CHAPTER 1 Introduction Sponsorships have become a major tool of marketing communication. In 2004, corporations spent $28 billion worldwide in sponsorship activities, a growth of 8.1 % compared to 2003. In North America alone, 2004 sponsorship spending reached $11.14 billion, up 8.7 % from 2003 (IEG 2003). The first demonstrations of commercial sponsorship activities can be traced back to the 18 th century when, in England, local sponsors supported horse races. Later, in 1891, Michelin sponsored the French cyclist Charles Terront on the race Paris-Brest (D ambron 1993). Today, sponsorship activities have evolved dramatically. They range from corporate sponsorship of sporting events such as the Olympic Games by Coca-Cola, to the sponsorship of ar tistic events like the Tribeca Film Festival in New York by American Express. As a consequence of the growing importance of sponsorship, academics began studying it in the early 1980s and the literature has been expanding at a steadfast rate ever since. From the numerous definitions of sponsorship that can be found, two components are common among most sponsorship rese archers (e.g., Cornwell and Maignan 1998; Meenaghan 1991; Roy and Cornwell 2003): 1) the possibility for the s ponsor to associate itself with the sponsee (i.e., the sponsored entity ) at the corporate, product, or brand level, in exchange for financial or in-kind support, and 2) a set of market ing activities centered on that association. The same definitional elements are provided by the International
2 Events Group (IEG Glossary and Lexicon 2003), a leading source of sponsorship information. The sponsee can be an individual (e.g., an athlete, an artist), a group of individuals (e.g., a sports team, an associati on) or an event (e.g., a sporting event, a concert, or an art exhibit). The literature focuses principally on s porting events. Sporting events are the most popular for sponsorship arrangements, largely because they convey positive values and consumers associate th em with notions of excitement and entertainment (Nichols, Roslow, and Dublis h 1999; Roy and Cornwell 2002). In 2004 they represented 69% of sponsorship sp ending in North America (IEG 2003). In addition, sporting events have an image capital similar to corporations or brands (Ferrand and Pages 1999). They are often the preferre d choices of sponsors due to the opportunity of being associated with them. Sponsorship is distinct from philanthropy and advertising (Hoek, Phillip, Jeffcoat, and Orsman 1997). Compared with sponsor ship, corporate philanthropy consists in making an anonymous donation or providing material support to charities (Berger, Cunningham, and Kozinets 1999; Ulibarri 2000 ) with no reciprocat ion expected by the firm (Gillies 1991). Contrary to sponsorship, an advertising message is totally controlled by the firm. In sponsorships, the content of the messages is contingent on the sponsee (Meenaghan 1991; Meenaghan and Shipley 1999). Advertising is seen as a less subtle activity than sponsorships that concentrates on short, rather than long term, objectives. As a result, consumers generally perceive advertising as being a more commercially oriented and obtrusive medium than sponsor ship. In fact, studies have shown that
3 consumers have a more favorable attitude to ward sponsorship than toward advertising (MacDonald 1991; Meenaghan 2001; Meenaghan and Shipley 1999). Sponsorship and the Brand Leverage Model Sponsorships most attractive characteristic is that it allows associating a brand to very popular and well-liked events. At a time of massive product offerings, media clutter, and savvy consumers, it has beco me increasingly difficult for marketing managers to build stronger brand equity than the competition. Brand equity is defined as the value that marketing effort and activi ties add to a product or service (e.g., Aaker 1991; Broniarczyk and Alba 1994; Farqhar 1989; Moore, Wilkie, and Lutz 2002). Brand equity is composed of both brand aw areness and brand meaning (e.g., Berry 2000). Brand Knowledge According to Keller (1993, 2003), the bra nd leverage process (see Figure 1) consists of associating brands with other entities (e.g., event, person, place, brand, and things) that carry desirable meanings for c onsumers in order to develop brand knowledge (e.g., attitude, thought, image, f eelings, awareness, and expe rience). Leveraging a brand through associations allows building brand equi ty more efficiently than traditional means of marketing communication (e .g., advertising). The leveraging process enables brands to borrow equity from other entities thr ough a process of knowledge transfer (Keller 2003). Associating a brand with an event (e .g., sporting, artistic, or charitable) through sponsorship activities has become a major brand leveraging tool in recent years.
Image Transfer Knowledge is said to be transferred between an entity and a brand when associations (e.g., attitude, feeling, image, etc.) linked to the entity become linked to the brand as well (e.g., through a sponsorship agreement). According to Keller (2003), brand image is a very important component of brand knowledge. The image of a brand encompasses all the meanings and symbols consumers associate to it (Durgee and Stuart 1987; Levy 1958). Based on this, image transfer occurring in sponsorship can be thought of as the meanings or symbols that become associated with the brand as a result of a sponsorship activity. For example, UPS might be perceived as being international due to its sponsorship of the Olympic Games which image has an international meaning. Figure 1. The Brand Leverage Model (Keller 2003) Multiple Sponsorships Although researchers have conceptually and empirically investigated the brand leverage phenomenon in situations of single sponsorship (i.e., one sponsoring brand tied to one event) (e.g., Keller 2003; Gwinner 1997; Gwinner and Eaton 1999), multiple 4
5 sponsorships (i.e., two or more firms spons oring the same event concomitantly) have been overlooked. As a consequence, key factors concerning the impact of multiple sponsorships on knowledge transfer are st ill unknown for marketing managers and researchers due to a lack of theoretical deve lopment and empirical research. This is a serious gap in the existing literature since th e association of a single sponsor with an event seldom characterizes event sponsorship. In fact, most of the largest sporting events are tied to multiple sponsors. For example, ten brands, including Coca Cola, McDonalds, and Visa, were the multiple majo r sponsors (i.e., concomitant sponsors) of the 2004 Athens Olympic Games. In view of th is, the need to investigate image transfer mechanisms in multiple sponsorships appears fundamental for achieving a better understanding of sponsorship effects. Statement of Research Problem This dissertation proposes a concep tual framework for multiple event sponsorships (i.e., when more than one brand sponsors the same event simultaneously). Based on the notion of entitativity, this framework posits that multiple sponsorships influence the sponsoring brands images. Enti tativity is the extent to which a group of elements is perceived as being an entity by itself (Campbell 1958). It is proposed that when several brands sponsor the same even t concomitantly (i.e., multiple sponsorships), these brands and the event are likely to be pe rceived as an entitativ e group. Psychologists have shown that social percei vers abstract a stereotype (i.e., a core identity) from entitative groups and perceive each group member in light of that stereotype (Crawford, Sherman, and Hamilton 2002). Similarly, it is argued that attributes of concomitant
6 sponsoring brands, such as images, are likely to be influenced by a group stereotype of the sponsors and the event. Entitative groups are perceived throu gh a group impression that is based on specific information associated to each indi vidual member of the group (Yzerbyt, Rocher, and Schandron 1997). This group impression is applied to each member of the group and, as a result, knowledge (i.e., image) asso ciated to one particular member becomes associated with other members. This re sults in a transfer of knowledge among group members (Crawford et al. 2002). The Savings in Relearning Paradigm Image transfer in sponsorship is likely to be very subtle (Pham and Vanhuele 1997) and could constitute implicit know ledge (i.e., knowledge not consciously accessible by the individual) (Reber 1989). Im plicit knowledge is much more stable over time than explicit knowledge (i.e., knowledge readily accessible by th e individual) (e.g., Matthews, Russ, Stanley, Blanchard-Fiel ds, Cho, and Druhan 1997). Implicit image transfer can be captured by the savings in relearning paradigm (Ebbinghaus 1885/1964). This paradigm has been used to measure implicit learning in experimental psychology (e.g., MacLeod 1988). It has also been used in social psychology to test for implicit traittransference among social group members (e.g., Crawford et al. 2002) or implicit traitinference about social actors (e.g., Skowrons ki, Carlston, and Crawford 1998). It has not been used to measure brand or event image tran sfer to date. This research introduces the savings technique to measure image transfer in multiple sponsorships. According to this paradigm, differences in savings acr oss treatment and control conditions are a
7 measure of implicit memory. In this rese arch, savings are operationalized as the performance of individuals on cued-recall of associations be tween an image-trait and a brand in the case of multiple sponsorships and no sponsorship. Intended Contribution Firms are often interested in consumers perceptions of events and make their sponsorship decisions on the hope that consum ers will associate their perceptions of an event with their perceptions of its sponsori ng brands. However, entitativity in multiple sponsorships is an important phenomenon for br and managers because it implies that not only the event has to be considered when making sponsorship decisions, but also the concomitant sponsoring brands of that event. It is proposed that, in the case of multiple sponsorships, the image of a sponsoring brand can be impacted in two ways. First, the images of the other brands can be transf erred to the sponsoring brand. Second, the events image can be transferred to the sponsoring brand. In addition, evidence of the implicit nature of the image transferred would be very important because of the long-term effects that implicit knowledge could have on brand image and equity. The above discussion sugge sts that brand managers would have to consider the core identity conveyed by the group of concomitant sponsors when designing brand image and brand equity building strategies through a multiple sponsorships agreement. This dissertation makes three major theo retical contributions to the existing literature on sponsorship and branding by:
8 1. Conceptualizing the role of percei ved entitativity in the formation and generalization of a group stereotype in multiple sponsorships, 2. Empirically investigating im plicit image transfer among the sponsoring brands as well as from the event to the sponsoring brands, 3. Adapting the savings in relearni ng paradigm to investigate these phenomena. An Entitativity-Based Alternative to the Brand Leverage Model The brand leverage model (Keller 2003) is not well adapted for situations of multiple sponsorships. According to the bra nd leverage model, knowledge is transferred directly from the entity to the brands. Such transfer effects have been identified from an event to its sponsoring brand (i.e., Gwinne r and Eaton 1999), as well as between two concomitant sponsoring brands of the same event (Carrillat, Harri s, and Lafferty 2004b). The brand leverage model, however, does not conceptually account for the multiple transfer effects that could occur when the concomitant sponsors are numerous. As illustrated in Figure 1, the brand leverage model posits that knowledge is transferred between pairs of entities. Therefore, accordi ng to this model, in the case of numerous and simultaneous brand associations such as multiple sponsorships, knowledge would be transferred from one brand to another and fr om the event to each brand in a pair-wise manner.
9 Alternatively, it is proposed that when mu ltiple brands sponsor the same event, perceivers abstract a gro up impression from the brands and the event due to high entitativity. That is, in the researchers view, image transf er is not due to the direct association between pairs of brands as well as between th e event and each brand but, rather, is due to the associat ion of the brands with a gr oup impression. In the case of multiple sponsorships, the association of an event with numerous sponsoring brands will be subjected to the phenomenon of entitativity. Entitativity is the extent to which a group of elements is perceived as having the nature of an enti ty, a real existence (Campbell 1958). More specifically, because concomita nt sponsoring brands and the event are likely to form an entitative group, they will be perceived as being a single entity rather than distinguishable elements. The social psychology literatur e has investigated the noti on of entitativity and its impacts on the perception of persons and social groups. The degree of entitativity affects how knowledge about one group member char acterizes the entire group and is then transferred to each group members (Crawford et al. 2002). Highly entitative groups are subjected to stereotypic pr ocessing (i.e., group impression fo rmation) (Crawford et al. 2002). In the case of high entitativity, indi vidual traits are abst racted from the group members simultaneously with the processing of information (i.e., in an on-line manner) (McConnell, Sherman, and Hamilton 1997), and ar e associated with the group and all its members due to category-based as opposed to individuating processi ng (in the case of low entitativity) (Fiske and Neuberg 1990). Figure 2 illustrates these principles in a multiple sponsorships context. Due to the high entitativity of the concomitant sponsoring brands and the sponsored event, a group stereotype will be formed and each member will
be perceived in light of that stereotype. Knowledge (i.e., image) will be transferred from the group stereotype to each member of the group. As a consequence, to the extent that the stereotype is composed of the image of the sponsoring brands, knowledge associated with one brand will be generalized to every other brand. In addition, since the event contributes to the group stereotype, its image will also be transferred to the sponsoring brands. Figure 2. An Entitativity Model of Implicit Image Transfer in Multiple Sponsorships F Image Event F Implicit Knowledge Transfer The marketing field recognized that implicit learning could take place when consumers are exposed to fragments of advertising or subtle communication messages such as sponsorship-like communications (Johar and Pham, 1999; Pham and Vanhuele 10
11 1997). Although the use of implicit measures of social cognition has been advocated in consumer research (Brunel, Tietje, and Greenwald 2004), to date no paradigm has been brought forward that can direc tly investigate imp licit consumer knowledge. As will be shown subsequently, the savings in rel earning paradigm (Ebbinghaus 1885/1964) allows for assessing implicit image transfer phenomena in multiple sponsorships. The main objective of this dissertation will be to investigate the effects of multiple sponsorships on consumer knowle dge about concomitant sponsoring brands images through the theoretical perspective of group impression formation and trait transference developed in the person and group perception litera ture (e.g., Crawford et al. 2002). Specifically, this di ssertation will address the following research questions: 1. Do multiple sponsorships result in implicit image transfer among concomitant sponsoring brands? 2. Is there evidence of implicit image transfer from the event to the sponsoring brands? 3. Are the image transfer phenomena due to the group entitativity of the sponsoring brands and the event? Research Method Employed An experiment that relied on the savi ngs in relearning paradigm was conducted for investigating these questi ons. Respondents were introduced to a set of fictitious
12 sponsoring brands and a fictiti ous event (created for use in this study). En titativity was manipulated through a multiple sponsorships (high entitativity) and a no sponsorship (low entitativity) conditions and the image transfer phenomena were measured through a memorization task (the savings measure). This first chapter presented a brief in troduction of sponsorship, the objective of this study, and an overview of the method th at will be used. The next chapter will expand the different streams of literature this research draws upon. Th is will lead to the formulation of empirical pred ictions in a hypothesis format.
13 CHAPTER 2 Literature Review The review of the literature covers the four areas of i nquiry that are relevant for this dissertation. First, academic literature on sponsorship is reviewed. A description of sponsorship as a medium of marketing communication, as well as the main theoretical perspectives and empirical findings from spons orship studies, are provided. Second, the branding literature is reviewed in order to put multiple sponsorships in the general perspective of brand associations. For that purpose, brand associati on techniques such as brand extensions or celebrity endorsement ar e presented. The partic ularities of multiple sponsorships as a method of brand association are also emphasized. Third, the social psychology literature is reviewed in order to provide an understanding of the concept of entitativity. Its antecedents are descri bed, as well as its impact on information processing. In addition, conceptual arguments are presented as to how entitativity could stem from multiple sponsorship situati ons. Fourth, the experimental psychology literature is reviewed. A presentation of implicit memory is first undertaken followed by a description of the savings in relear ning paradigm and how it captures implicit knowledge. In the last part of this chap ter, a summary of the conceptual background leading to the hypotheses that will be empirically tested is provided.
14 Sponsorship Research The Sponsorship Medium Compared to advertising, sponsorship has two distinctive charac teristics: 1) the medium of sponsorship is inseparable from the content of the information it conveys since the message is the sponsorship itself, and 2) consumers assign a notion of goodwill to it. Sponsorship is the medium by which a sponsor (e.g., a brand) is associated with a sponsee (e.g., a sporting event). Meenagha n and Shipley (1999) noted that this association is of a very peculiar type both medium and media elements are not separate, but are inextricably li nked (p. 6). As a result, the message of the sponsor of a sporting event is not indepe ndent from the environment of the sponsorship operation since the same means are used to produ ce and deliver the message (Meenaghan and Shipley 1999). The context of the communica tion is shaped by the event situation, which plays a key role in the resulting message perceived by consumers. This is the conceptual basis of Meena ghan and Shipley (1999) who identified the possibility of the transfer of inherent values of the event to the sponsor. They considered the sponsor and the event as parts of a sym biotic relationship. A sponsor inherits attributes from the event to which it is linked. However, the literature points out that attribute transfer is not well understood because it lacks an underlying theoretical explanation as well as empirical investiga tions (i.e., Javalgi, Traylor, Gross, and Lampman 1994; Lee, Sandler, and Shani 1997). Sponsorship has a connotation of goodwill (Meenaghan 2001). In effect, consumers often consider the association betw een a sponsor and a sponsee as a legitimate connection. They believe the sponsor takes risks by associating its name to an event
15 because they consider the sponsor has litt le control on its outcomes (MacDonald 1991). Consumers consider a sponsoring company as being less commercially oriented than a traditional advertiser because the sponsor directly assists in the development of the event, team, or individual. As a result consumers defense mechanisms against sponsorships are weaker than agai nst advertising (Meenaghan 1998, 2001). The Effects of Sponsorships Communication Early academic works on sponsorships focused on the role of sponsorship in marketing communication. Studies aimed at introducing sponsorship as a legitimate component of the marketing mix (e.g., Meenaghan 1983) that does not result from the altruistic inclination of th e firms CEO (e.g., Crimmins and Horn 1996). Research was mostly descriptive and inves tigated the objectives of sponsorships as well as how organizations managed event sponsorships (e .g., Abratt, Clayton, and Pitt 1987; Crowley 1991; Marshall and Cook 1992). Theories associated with sponsorship effects More recently, research on sponsorship has attempted to integrate vari ous theories to under stand the effects of sponsorship on consumers (see Appendix 1). Schema theory, which asserts that past experiences or activities are organized as an active network in memory (Bartlett 1932), has been used by researchers on the basis that the information about the sponsor and the event is accessed from memory. Information about the sponsorship agreement is then compared with the schema of the event a nd the sponsor (i.e., Gwinner and Eaton 1999; McDaniel 1999; Roy and Cornwell 2003; 2004; Speed and Thompson 2000). Key
16 findings based on schema theory include th e positive relationship between the eventsponsor congruency (in terms of brand equity ) and attitude toward the sponsoring brand (Roy and Cornwell 2003). Also, Gwinner and Eaton (1999) found that image could be transferred from a sporting event to a sponsoring brand. They provided evidence that image congruency between the sponsor and the sponsoring brand was greater under a yes sponsorship than under a no sponsorship condition. Human Associative Memory (HAM), which considers declarative knowledge as a network of concept nodes (cues and outcomes) associated by links of different strength (Anderson and Bower 1973), has also been used for investigating sponsorship effects. According to the HAM framework, the associ ations linked to the event modify the associations linked to a brand in consumers memory in the case of a sponsorship agreement. Based on this framework, studies have found that the impact of sponsorship on attitude and purchase inte ntion was positive (Chapman and Aylesworth 1999; Simonin and Ruth 1998), but stronger for unfamiliar than for familiar brands (Carrillat, Harris, and Lafferty 2004a). Other theoretical perspectives of spon sorship effects include balance theory, congruity theory, attribution th eory, heuristics, halo effect, and social identity theory. Based on balance theory, Dean (1999) f ound that sponsorship positively impacts perceived corporate citizenship of the sponsoring companies. Based on congruity and attribution theory, Dean (2002) also found that corporate sponsorship of a well-liked event resulted in an improved perceived corporate community relations for the sponsor. Madrigal (2000, 2001) found that pu rchase intention of the spons ors product, as well as the link between attitude and purchase intenti on, were stronger for consumers with a high
17 versus a low level of identification with the sponsee. Johar and Pham (1999) supported the assertion that consumers rely on heuristics for identifying sponsors of events when they cannot retrieve the sponsor s directly from memory. Brands that are prominent in the market or that are related to the event are more likely to be identified as sponsors of that event. Other empirical studies indicated a positive impact of sporting event sponsorships on brand image (e.g., Crimmins and Horn 1996), corporate image (e.g., Javalgi et al 1994; Stipp 1998; Stipp and Schiavone 1996) or brand recall (e.g., La rdinoit and Derbaix 2001). The above empirical findings on sponsorshi p research show that studies of the knowledge transfer phenomenon ar e sparse (see Appendix 1). Sponsorship is described as a prominent brand leveraging tool (Ke ller 2003, Roy and Cornwell 2003; Ruth and Simonin 2003), however, only a fe w studies investigate the role of sponsorships as such a tool (i.e., transferring knowledge from other entities to a sponsoring brand). The large majority of studies assessed the impact of sponsorships on recall, image, or attitudinal variables. Nonetheless, Gwinner and Eatons (1999) work supports the notion that the image of the sponsored event can be transfe rred to the sponsoring brands in the case of single sponsorship. Two studies empirically investigated multiple sponsorships. In one study, Ruth and Simonin (2003) found that prio r attitude toward the sponsors had a positive impact on attitude toward the sponsor ed event. This study did not, however, investigate knowledge transfer, but rather assumed the existence of this phenomenon for predicting directional changes of attitude. Besides, it examined the impact of the sponsors on the event, but it did not investig ate the possibility of knowledge transfer between the sponsoring brands.
18 Another study showed that knowledge (i.e ., image) could be transferred not only from the event to the sponsoring brands, but also between the concomitant sponsoring brands of the same event (Carri llat et al. 2004b). However, it is important to notice that these two studies included a multiple sponsor ships situation with two sponsoring brands only and neither considered the possibility of the formation of a group impression that could impact consumers implicit knowledge about each sponsor. Multiple Sponsorships as a Type of Brand Association As mentioned before, a sponsorship can be categorized as a brand leveraging method that leads to a transfer of knowle dge between entities. Other marketing techniques such as brand extensions and cel ebrity endorsements can be used as brand leveraging tools as well. Both these techni ques allow transferri ng knowledge to a brand through association with another entity (K eller 2003). Therefore, examining the literature in these areas would provide a better understanding of how researchers approached knowledge transfer in re lated but quite different domains. Brand extensions Marketers extend br ands beyond their orig inal categories to reduce risks and cost inherent to enteri ng a new product category (Aaker 1991). Different strategies are then available that all imply some type of knowledge transfer: brand alliance, brand extens ion, composite brand extensi on, or ingredient branding. Simonin and Ruth (1998) investig ated the spillover effect that occurs in the case of brand alliance. They found that consumers prior attitude toward th e alliance itself can influence attitude toward each brand com posing the alliance. Based on the signaling theory of information economics (i.e., people infer that claims about unobservable quality
19 are true otherwise it would be too moneta rily detrimental to the claimer), Rao and Ruekert (1999) investigated th e role of an ally in a bra nd alliance. They found that consumers evaluation of the unobservable quality of brands is enhanced when it is allied with a brand that could be harmed by consum er sanctions in the case of a false claim. Boush and Loken (1991) investigated br and extension from a categorization perspective. By assuming that a brand a nd its products compose a category on their own, they found a positive relationship between the typicality of the extension (i.e., its degree of representativeness of the ca tegory of the brand and its pr oducts) and the evaluation of this extension. Using a si milar categorization approac h, Loken and John (1993) found that when brand extension attributes are inc onsistent with the fam ily brand, beliefs held toward the attributes associated with the brand name are diluted. Park, Jung, and Schocker (1996) examined composite brand extensions. Their results showed that combining two brands with complementary attributes into a new product (a Slim-Fast cake mix flavored with G odiva chocolate) could generate superior consumer attitudes. Desai and Keller (2002) showed that a self-branded ingredient strategy leads to more favorable evaluation of slot-filler expansion (i.e., change in the level of one product attribute), whereas a co-b randed ingredient strategy leads to more favorable evaluation of new attribute expansions (i.e., an entirely new attribute or characteristic is added to the product). These studies give insights for unders tanding the different forms of brand alliances; however, they do not solve the main problems raised by multiple sponsorships. First, the transfer of knowledge is not investig ated per se but is posited to take place in order to explain phenomena such as improved br and attitudes or better brand evaluations.
20 Second, ingredient branding, brand extensions, brand alliance, or composite brand extensions involve only one source of knowle dge transfer (the other brand) whereas concomitant sponsorship involves simultaneous sources of knowledge transfer (the event and all the other sponsoring brands). As pointed out by Keller (2003), multiple simultaneous sources of knowledge transfer re main an unexplored aspect of the brand leverage process. Celebrity endorsement Celebrity endorsement is also a way to leverage a brand. However, in this case, knowledge transfer does not take place between an event and a sponsoring brand, but between a celebrity and an endorsed br and. Celebrity endorsement is different from sponsorship due to th e fact that the association between the product/brand and the entity (i.e., celebrity) is made explicit through an advertisement (McCracken 1989). Similar to brand extension studies, knowledge transfer has been posited rather than being the focus of the studies on celebrity endorsement and most findings concern traditional outcome variable s. Celebrity endorsement has been shown to increase brand recognition (Petty, C acioppo, and Schumann 1983), message recall (Friedman and Friedman 1979), attitude toward the brand (Kamins, Brand, and Hoeke 1989), and brand choice (Heath, McCarthy, and Motherbaugh 1994; Kamins et al. 1989; Khale and Homer 1985; Ohnian 1991). Several characteristics of the source of th e message (i.e., celebrity) can impact the endorsed product. Source credibility (exper tise and trustworthine ss of the celebrity) (Tripp, Jensen, and Carlson 1994) is positively related to attitude toward the ad and products image evaluation (Atkins and Block 1983). In addition, the match-up hypothesis suggests that celebrity endorsement is more eff ective when there is a fit
21 between the endorsed product and the endor ser (Kamins 1990; Khale and Homer 1985; Lynch and Schuler 1994). Most of the evid ence focuses on physical attractiveness and shows that endorsements are more effectiv e when an attractive celebrity endorses attractiveness enhancing products (Kam ins 1990; Khale and Homer 1985). Other findings suggest, however, that the match-up hypothesis is more valid when it is based on expertise rather than physical attractiven ess. Till and Busler (1998, 2000) found that endorsements were more effective in impr oving brand attitude when endorsers were perceived as being experts in the produc t domain than when attractive celebrities endorsed products that enhance ones attractiveness. Source meaningfulness also plays an impor tant role in the effectiveness of celebrity endorsers. McCracken (1989) referred to celebri ty endorsement as a way to transfer social meaning from the celebri ties to the products endorsed, which was supported by findings from Langmeyer and Walker (1991). The effectiveness of multiple endorseme nts (i.e., one celebrity endorsing more than one product or brand) has been investigated. Ther e is not, however, a strong conceptual foundation in this research and the evidence is sparse and contains mixed findings. A study by Mowen and Brown (1981) revealed that consumers attitudes toward the endorsed brands were negativel y affected when the celebrity endorsed multiple brands. Alternatively, Tripp et al. (1994) did not find a negative impact of multiple endorsements on attitude toward the endorsed brands. Therefore, guidance from previous works on multiple celebrity endorsement is limited when investigating multiple sponsorships.
22 The Ambiguous Role of Similarity The traditional view. Categorization research suggested that similarity is the most important antecedent of categorization because people tend to classify similar objects in the same category (e.g., Fiske and Taylor 1991 ; Rosch and Mervis 1975; Srull and Wyer 1989; Sujan 1985; Tversky 1977). According to this view, items associated with similar objects are judged as falling under the same cognitive category and are, therefore, associated to both objects (Boush, Shipp, L oken, Gencturk, Crockett, Kennedy, Minshall, Misurell, Rochford, and Strobel 1987; Smith a nd Medin 1981). As a consequence, to the extent that two entities are perceived as bei ng similar, they will be categorized together and knowledge transfer will ta ke place between them. Based on this view of similarity, the branding literature has widely investigated the role of categoriza tion in the evaluation of brand extensions (e.g., Aaker and Keller 1990; Boush and Loken 1991; Keller and Aa ker 1992). Researchers posited that consumers use similarity judgment when ev aluating the fit between the brand and the extension. The greater the fit, the more li kely the extension is to be perceived as belonging to the category of th e parent brand and the greater is the knowledge transferred from the brand to the extension (Levy and Tybout 1989; Martin and Stewart 2001; Sujan 1985). Sponsorship research has also assumed that fit or similarity between the event sponsored and the sponsoring brand is benefi cial for the sponsorship (e.g., MacDonald 1991; Meenaghan and Shipley 1999). Johar and Pham (1999) found that similarity between the event and a brand influenced spons orship recognition in favor of that brand even when it was not a sponsor of that event. Their results indicated that when a brand is
23 perceived to be related to a sponsored event (i.e., semantic overlap between features of the event and those of the sponsors), consumers are more likely to attribute the sponsorship of the event to that brand rath er than any other brand when they fail to directly recall the co rrect sponsoring brand. Gwinner and Eaton (1999) found that image transfer between the event and the sponsor of the event is positively related to their functionaland image-based similarity. A new stance on similarity More recently, research ers pointed out that the traditional notion of similarity stands on a fragile conceptual ground and that the relevance of its role in ca tegorization and knowledge transf er should be reconsidered. Indeed, some findings in the brand extension area indicate that the notion of similarity is secondary to the notion of specific brand asso ciations in knowledge transfer (Broniarczyk and Alba 1994). Other findings indicate that the degree of fi t between the brand and the extension depends more on the extent to which the extension product accommodates the brand-concept (brand-concept consistency) (Park, Milberg, and Lawson 1991). More recently, Meyvis and Janiszewski (2004) showed that not only similarity but also brand benefit accessibility play an important role in the evaluation of brand extension. They found that the extension of a broad brand (i.e., one brand subsuming products from different categories) is preferred to the extension of a narrow brand (i.e., one brand subsuming products from the same category) when the brands are both extended to a product category similar to them. This is due to the greater accessibility of brand benefit associations of the broad brand (i.e., broad brands do not have strong category associations that interfere w ith their benefit associations). The above mentioned studies indicate that similarity may not be the key variable for investigating knowledge transfer.
24 It is possible that similarity may be as mu ch a consequence of knowledge transfer as a cause (Murphy and Medin 1985). Given th is ambiguous role of similarity in categorization, there exists a need to inve stigate other possible drivers of knowledge transfer. It is suggested that the extent to which knowledge can be transferred among objects depends on their percei ved entitativity as a group of objects (Campbell 1958). The next section is devoted to the notion of entitativity in multiple sponsorships and how it contributes to the understanding of imag e transfer mechanisms among concomitant sponsoring brands as well as between the event and the sponsoring brands. The Role of Entitativity in Multiple Sponsorships Goodman (1972) stated that similarity is invidious, insidious, a pretender, an imposter, a quack (p. 437). Although this pos ition might appear extreme, it adequately conveys the notion that similarity is not all that useful for explai ning the categorization process. First, similarity is too flexible (Medin 1989; Medin, Goldstone, and Gentner 1993). In effect, according to Tverskys ( 1977) contrast model, similarity between objects is a function of their common and distin ctive features weighted by their salience, with common features increasi ng similarity and distinctive on es decreasing similarity. However, these weights are relative to the c ontext of judgment. In the context of a similarity judgment where a subject is compar ed to a referent, the focus is on the subject and its features are then weighted more h eavily than the referents. For example, subjects ratings of the similarity between No rth Korea (i.e., subject) and Red China (i.e., referent) are greater than thei r ratings of the similarity be tween Red China (i.e., subject)
25 and North Korea (i.e., referent). This is e xplained by the fact that Red China has more distinctive features than North Korea and, according to the contrast model, detracts more from similarity than when North Korea is the subject (Tversky 1977). Second, Murphy and Medin (1985) have ar gued that any two things may be arbitrarily similar or dissimilar. In effect, there could a countless number of features that two objects have in common. As an extreme example, a plum and a lawn mower share an infinite number of commonalities: bot h weight less than 1,000 kg, are found on earth, both are found in our solar system, both cannot hear well, both have an odor, both are not worn by elephants (see Medin 1989, p. 1473). Although similarity raises some problems, it is still an important element of categorization. Medin (1989) developed a model of categorization that takes into account both similarity and conceptual coherence. A ccording to him, superf icial similar features lead perceivers to infer that these features are caused by deeper underlying factors, a phenomenon called essentialization (Medin 1989). Objects will be categorized together if they share the same underlying core esse nce inferred from surface similarities. The concept of entitativity refers to the extent to which a gr oup of objects is perceived to share a common core, i.e., an essence (Cam pbell 1958). As a consequence, members of an entitative group will be categorized toge ther and a transfer of knowledge will take place between the group and its members (i.e ., information linked to the group becomes linked to the members of the group as well). Therefore, entitativity is a more relevant variable than similarity when trying to inve stigate knowledge transfer In fact, empirical evidence supports the noti on that entitativity is a key driver of gr oup stereotype formation (Crawford et al. 2002). Similarity has an ambiguous role and is, in fact, one of the
antecedents of entitativity (Campbell 1958). Welbourne (1999), however, showed that it is not as an important variable as unity or perceived cohesiveness in explaining the categorization of group members. The concept of entitativity, as well as its causes and consequences, are explained in the following section. In addition, Figure 3 graphically displays the antecedents of entitativity and its impact on information processing. Figure 3. Antecedents and Consequences of Group Entitativity On-line vs. Memory-Based Judgment Integrative Processing Group Impression Abstraction Individual Differences: Need for Closure Perceived Group Entitativity Group Characteristics: Similarity Proximity* Common Fate Pregnance Interdependence of Group Members* Group Interaction Common Goals Common Outcomes Group Essential Core Group Permeability Group Duration Group Importance Category Based vs. Individuating Processing Group Members Subsumed Under Grou p I m p res si on Stereotypic vs. Exemplar Processing of Group Members *: Factors leading to a high entitativity of the group of brands and the event in multiple sponsorship situations. 26
27 The Antecedents of Entitativity Drawing on the Gestalt principles of perceptual organization, Campbell (1958) stated that a group of individual elements is pe rceived as an organized whole to the extent that this group is entitative According to Campbell (1958), entitativity refers to the degree with which a social aggr egate is perceived as having t he nature of an entity, of having real existence (1958, p. 17). Campbell ( 1958) also claimed that different groups have different degrees of entitativity. Re sults from Lickel, Hamilton, Lewis, Sherman, Wieczorkowska, and Uhles (2000) confirmed that view. They showed that the perception of group entitativity differs greatly from one group to another. For instance, they found that members of a sport team or of a family are perceived much more like a single entity than people in the audience at a movi e or people waiting at a bus stop. Understanding what factors can lead to a higher perceived entitativity is important for multiple sponsorships because the degree of perceived entitativity will impact how information about each sponsoring brand is processed and what knowledge can be transferred from the event or the brands to each concomitant sponsoring brand. There are two types of antecedents to entitativity: 1) the cognitive characteristics of the perceiver can influence his/her tendency to perceive the group as being high in entitativity, and 2) the characteristics of the group can impact the extent to which it is perceived as being entitative. Individual differences Individual difference variables such as Need for Closure (Webster and Kruglanski 1994) may intensify peoples tendency to seek for coherence and meaningfulness in the perceptio n of groups (Lickel et al. 2000).
28 Need for Closure refers to the desire for a clear-cut opinion on a judgment topic (Webster and Kruglanski 1994). Two types of factors impact the need for closure: 1) contextual, and 2) individual differences. Firs t, when the costs outweigh the benefits of further information processing in a particul ar situation, the need for closure of an individual increases. This can occur when the predictability is important, when a decision is made under time pressure, or when the processing of further information is tedious (Kruglanski, Webster, and Klem 1993; Webster and Kruglanski 1994). Second, Need for Closure is also described as a stable individual dime nsion. Webster and Kruglanski (1994) described Need for Cl osure as a unidimensional individual difference variable composed of 5 facets. When the need for closure of an individual is high as opposed to low, this individual will have an increase in: 1) the preference for order and structure of the environment, 2) the affective discomfort occasioned by ambiguity, 3) the decisiveness of judgment and choices, 4) the tendency to afford predictability to future contexts, and 5) the unwillingness to have ones opinion contradicted by others or by inconsistent ev idence. Need for Closure has been shown to lead to stereotypically dr iven judgment in the group per ception literature (Kruglanski and Freund 1983). Individuals with a high need for closure tend to rely on stereotypes in order to limit their cognitive effort and to ma ke quicker decisions. Therefore, the extent to which an entitative group of brands and the event involved in multiple sponsorships are processed stereotypically will be greater fo r individuals with a high need for closure. This implies that individuals with a high n eed for closure are more likely to exhibit a greater transfer of implicit knowledge (i.e., know ledge that is not c onsciously accessible) than individuals with a low need for closure.
29 Group characteristics According to Campbell (1958) four factors could lead a group to be perceived as more entitative: 1) similarity, 2) proximit y, 3) common fate, and 4) pregnance. 1) Similarity has been esta blished as an antecede nt of perceived group entitativity. The aspects according to whic h groups can be perceived as similar are numerous (Lickel et al. 2000). For instance, Crawford et al. (2002) and McConnell et al. (1997) manipulated group members simila rity by informing respondents of the resemblance of these persons personality, opinions, beliefs, and behaviors across situations. 2) Proximity refers to the physic al distance between the different elements of the group (Campbell 1958). 3) Common fate is th e extent to which elements of a group move in the same direction successively in time (Campbell 1958). 4) Pregnance refers to the extent to which a group of elements ar e perceived as forming a part of a spatial organization (Campbell 1958). Common fate has been argued to be the most important driver of group entitativity before proxim ity, similarity, or pregnance (Campbell 1958). Apart from these factors, researchers have focused on other antecedents of entitativity. Lickel et al. (2000) found that the degree of interdependence between group members, the extent to which the group is perc eived as having an essential core, and the importance of the group for each individual member are key factors of perceived entitativity. The degree of interdependence among the members of a group impacts perceived entitativity because it influences the degree to which the group is perceived as a coherent unit (Gartner and Schopler 1998). Three properties impact the interdependence of group members: 1) the de gree of perceived gr oup interaction, 2) the extent to which members are perceived as ha ving common goals, and 3) the extent to which members are perceived as having comm on outcomes. Lickel et al.s (2000)
30 findings show that these variables are positively correlated with perceived group entitativity (r ranging from .37 to .58). The extent to which the group has an essent ial core leads people to perceive the group as being inalterable (Rothbart and Tayl or 1992). Two properties of the group can lead one to perceive the group as being as such (Lickel et al. 2000): 1) the permeability of the group boundaries (i.e., the di fficulty to exit or enter the group), and 2) the duration of the group. Lickel et al. ( 2000) found that permeability was negatively related to entitativity (r = -.24) whereas duration had a small positive correlati on with entitativity (r = .11). Importance of the group to the members is a crucial variable in the social identity and group dynamics domains (Cartwright a nd Zander 1960; Tajfel and Turner 1985). Importance has a fairly strong relationship with en titativity; according to Lickel et al. (2000) the correlation between these two variables is .50. The Entitativity of the Concomitant Sponsoring Brands and the Event It is likely that multiple brands sponsoring an event will be perceived as being more entitative than if the brands do not concomitantly sponsor that event. Two important antecedents of entitativity charac terize the brands and the event involved in a multiple sponsorships agreement: 1) proximity, and 2) interdependence. In effect, the brands and the event are often concomitantly presented to the diverse audiences (e.g., television viewers, on-site spectators) thr ough side-by-side signage (Ruth and Simonin 2003) that strengthens the im pression of proximity among the brands and the event. In addition, the sponsoring brands and th e event are likely to be seen as being interdependent because consumers will percei ve they have the same goals and outcomes
31 (Lickel et al. 2000). As pointed out by Meenaghan and Shipley (1999), consumers perception of the objectives of event sponsorsh ip depends on the type of event sponsored (e.g., sporting events sponsorship is percei ved as being commercial whereas charitable event sponsorship is perceived as being phila nthropic). Therefore, if the concomitant sponsoring brands are tied to the same event, th eir objectives are likely to be perceived as fairly similar. Empirical evidence confirms that assertion. Welbourne (1999) found that the perceived unity of th e targets intentions and goals se rve as a basis to the entitativity assumption made by respondents. Furthermore, since concomitant s ponsoring brands are tied to the same event, it is likely that cons umers will consider that the outcomes of the sponsorship are the same for all the partie s involved. As a consequence, physical proximity and interdependence suggest that concomitant spon soring brands and the event will be perceived as being high in entitativity. The Influence of Entitativity on Information Processing Perceived entitativity of a group has a strong influence on the way people form impressions about target elements of that group (Welbourne 1999). If a group is perceived as being high in entitativity, info rmation about its members will be processed in much the same way as about an individual target. When trying to form an impression about an individual target as opposed to a group, a perceiver expects consistency in the observed traits and behaviors of that targ et. This is because one assumes that dispositional characteristics, such as persona lity for instance, underl ie these observations (Asch 1946; Hamilton and Sherman 1996; McConnell et al. 1997).
32 Groups that are perceived as high in enti tativity are assumed to have a unity (Campbell 1958) similar to an individual target As a consequence, highly entitative groups are also expected to be consistent (e.g., given one observed behavior, the target would be expected to perform similar beha viors subsequently (Asch 1946; Kelley 1973)), which implies that information about such groups is processed in the same way as information about individual targets (Ham ilton and Sherman 1996; Welbourne 1999). As a consequence, members of entitative groups are perceived to be interchangeable and they lose some of their individuality; in fact, they become confounded with the group impression (Crawford et al. 2002). Information processing of highly entitative groups as compared to low entitative groups has three main characteristics: 1) on-line versus memory-based judgment, 2) category-based versus individuating, and 3) stereotypic versus exemplar. On-line vs. memory-based judgment On-line judgment implies that information about the target (individual or group) is processed into an integrating fashion (McConnell et al. 1997). Integrative information processi ng is aimed at forming an impression of the target (McConnell et al. 1994) that leads to th e reconciliation of inconsistent information and to more associative linkages in memo ry (Hastie and Park 1986; McConnell et al. 1994). This abstraction of a ta rget impression will then be used in order to form an online judgment at the same time the informa tion is processed (McConnell et al. 1997). Therefore, when respondents are asked to make a judgment after the information was processed, they do not have to retrieve the information from memory and then form a judgment because it is already av ailable (Hastie and Park 1986).
33 In the case of memory-based judgment, th e perceiver does not attempt to form an impression of the target group or reconcile in consistent information because information is not processed into an inte grative fashion (McConnell et al 1994). Therefore, there is no abstraction formed at the time the inform ation is processed (McConnell et al. 1997). As a consequence, in the case of memory-bas ed judgment, when respondents are asked to form a global impression, they have to form a judgment based on information stored in memory (Hastie and Park 1986). Empirical ev idence shows that high entitativity groups are characterized by on-line judgment, wher eas low entitativity groups are characterized by memory-based judgment (McConnell et al. 1997, 1994). Therefore, information about high entitativity groups vs. low entitativity groups will be processed in an integrative manner. Two factors can explain why high entit ativity leads to on-line as opposed to memory-based judgment: 1) the motivation to reconcile inconsistencies, and 2) the limited capacity to process diverse and comp lex information (McConnell et al. 1994). In the case where a target group is perceived as being high in entitati vity, the behaviors or personalities of its members are expected to be coherent (McConnell et al. 1994). Therefore, because small variance in the group characteristics is an ticipated, perceivers will be more motivated to form an integrat ed overall impression of the group. However, when greater variance in the group characterist ics is anticipated (i.e., low entitativity), perceivers will be less motivated to form an integrated overall impression of the group. In addition, integrating information about a lo w entitativity group is more complex than about a high entitativity gr oup due to the variance of the information. Therefore, a perceiver is more likely to have the sufficient cognitive resources available for on-line
34 rather than memory-based judgment when th e group is high in entitativity (McConnell et al. 1994). Category-based vs. individuating information processing Both the dual-process model proposed by Brewer (1988) and the continuum model proposed by Fiske and Neuberg (1990) distinguish between categ ory-based and individuating information processing. The category-based process implies that the information about individual members is subsumed under a group impressi on. As a consequence, the information is stored in association with all the members of the group. The individuating process implies that information about a group me mber is uniquely associated with that individual. Hamilton and Sherman (1996) suggested that perceivers organize information around the units they consider coherent. In line with this assertion, Crawford et al. (2002) found that in the case of a high enti tativity group, the impression of each group member is formed on a category-based manner whereas in the case of a low entitativity group, the impression of each group member is formed on an individuating manner. As a consequence, information about individual me mbers of a highly entitative group will be associated with all the other members of th e group in the perceivers memory. For a group low in entitativity, the informati on about group members will be associated uniquely to each individual. To summarize, the previous discussion has several implications for the processing of information related to concomitant sponsor ing brands of an event. As previously explained, concomitant sponsoring brands are likely to be perc eived as high in entitativity because of their proximity and interdependence (i.e., common goals and outcomes) (Lickel et al. 2000). Therefore, perceivers will rely on an on-line rather than memory-
35 based judgment when processing information re lated to concomitant sponsoring brands. They will try to integrate the diverse information by abstracting a group impression that will be applied to each brand. In the case of multiple sponsorships, information related to each brand is likely to be stored on a categ ory-based rather than on an individuating manner. Therefore, information about a spons oring brand will become associated to the other concomitant sponsoring brands in consumers memory. Both on-line judgment and category-based pr ocessing of information suggest that a transfer of image is likely to take place among the brands of a multiple sponsorships agreement as well as from the event to th e sponsoring brands due to an increased entitativity. These two processes lead to the creation of a group impression. Information associated with each concomitant sponsor ing brand is abstracted to form a group stereotype that will, in turn, influence each of the concomitant sponsoring brands of the multiple sponsorships agreement. This group stereotype is composed of the images of the concomitant sponsoring brands, as well as the event, and is generalized to the perception of each brand. As a consequence, a given brand will become associated with the images of all the other sponsoring brands For instance, if Coca-Cola, MacDonalds, and New Balance are all sponsors of the Olym pic Games, consumers will try to integrate the different images of these brands and th e event by abstracting a group stereotype that will influence the perception of each brand. As a result of this stereotypic processing, Coca-Colas image will be influenced by the images of MacDonalds, New Balance, and the Olympic Games. Stereotypic vs. exemplar processing Entitativity impacts the cognitive processes engaged when one is developing a represen tation, or impression of the group (Hamilton
36 and Sherman 1996; Sherman, Castelli, a nd Hamilton 2002). Group entitativity is positively related to the extent to which groups are mentally represented as prototypes and negatively related to the extent to which they are mentally represented as exemplars (Brewer and Harasty 1996; Brewer, Weber, and Carini 1995). This prototypic representation of members from groups high in entitativity is due to the fact that perceivers are motivated to form a simple representation of the group. In fact, a group impression is generalized to group members as a function of the strength of perceivers expectancies about the group stereotype (Stangor and Mc Millan 1992). Stereotypic processing is more likely for a high entitative group for two reasons. First, entitativity creates a greater cognitive load on the percei ver that stereotypes can alleviate (Macrae, Milne, and Bodenhausen 1994). Second, entitativity leads perceivers to form expectancies about the group based on the pe rception of its real essence (Yzerbyt, Rocher, and Schadron 1997). The cognitive busyness of entitativity Social psychology considers stereotypes as energy saving devices that social perceivers use to simplify information processing and their response to that information (Allpor t 1954; Fiske and Neuberg 1990; Macrae et al. 1994). This leads people to rely on st ereotypic processing when individuation necessitates too much cognitive resources (Brewer 1988; Fiske and Neuberg 1990). A large body of research empirically supports th is assertion. Studies have shown that stereotypic processing is a si mplification tool (Macrae et al. 1994) by demonstrating that an increase in stereotypic processing is obtained by inducing cognitively taxing conditions (e.g., Kruglanski and Freund 1983; M acrae, Hewstone, and Griffiths 1993). In fact, Macrae et al. (1994) suggest ed that stereotypic processing is often used as a heuristic
37 by perceivers in order to free up cognitive resources that can be applied to more rewarding mental activities. Therefore, b ecause concomitant sponsoring brands represent an extensive source of stimuli, consumers mi ght be tempted to rely on a stereotype in order to simplify their cognitive task. Ap art from a cognitive simplification tool, stereotypes are also considered as devices allo wing perceivers to find coherent patterns in the world that provide expl anations of the environm ent through a process of essentialization (Yzer byt et al. 1997). Group essentialization Medin (1989) defined psyc hological essentialism as the tendency of people to act as if things (e.g., objects, other peop le, entities) have an essence or an underlying nature that make them the thi ngs they are. This e ssence is composed of the underlying properties that cause observabl e superficial propertie s. According to Medin (1989), the esse ntialist heuristic consists in assigning objects to categories and makes inferences about these objects by genera lizing characteristics fr om their category. In the social perception literature, Rogier and Yz erbyt (1999) relied on the essentialization process to describe th e phenomenon by which perceivers make link between surface characteristics of people and un derlying features they consider as causal factors of these characteristics. Such internal causal structures generate some attribution and motivate perceivers to integrate informa tion into a coherent story. Evidence shows that this leads to stronger dispositional inferences about group members behaviors (Yzerbyt, et al. 1998) and greater correspondence bias (Rogier and Yzerbyt 1999). These underlying features constitute the essence of the group, which is the common core of all group members and is crucial to the groups identity (Rogier and Yzerbyt 1999).
38 According to the subjective essentialism vi ew of Yzerbyt et al. (1997), the more a social group is perceived as having an esse nce, the more rationale ground a perceiver has for relying on a stereotype to make judgme nt about members of the group. People find a justification to their stereotypic beliefs a bout group members of a group in the perceived essence of the group. Furthermore, the essentia list view of stereot ype implies that groups with a strong essence have a high inductive po tential and a highly in terconnected set of characteristics (Yzerbyt et al. 1997). Therefore, perceivers will easily infer characteristics of individuals based on inform ation pertaining to their group membership. In addition, because these charac teristics are highly interconn ected, all the characteristics of the group are likely to be inferred and tr ansferred from the group stereotype to each group member. People are more likely to re ly on an essentialist heuristic if the considered objects belong to a highly entita tive group. This suggest s that essentialization is a function of perceived enti tativity (Yzerbyt et al. 1998). Therefore, perceivers are likely to rely on a stereotype when judging members of highly entitative groups. This leads group members to be perceived as homoge neous and interchangeable (Brewer et al. 1995; Crawford et al. 2002). It is proposed that, similarly to what is argued in the person and group perception literature (i.e., Rogier and Yzerbyt 1999), the processing of concomitant sponsoring brands will be subjected to an essentialist heuristic. This view implies that the concomitant sponsorship of the same event wi ll be seen as the manifestation of the underlying features common to the brands and the event. For instance, people could perceive that the Olympic Games and its spons oring brands all share the Olympic values. This is in line with the conceptualization of Praceju s (1998) who suggested that
39 consumers make inference about the size of the sponsor (sponsoring brands belong to large companies), the legitimacy of the spons or (sponsoring brands care about the event they sponsor), and the facilitation of the event (sponsoring brands make the event possible). Therefore, the induction of char acteristics and interc onnectedness of these characteristics will increase with entitati vity and perceived group essence, which will improve the potential for image transfer. The Impact of Concept Si milarity on Entitativity It was previously argued that multiple sponsorships would increase the perceived entitativity of the sponsoring brands and the event sponsored, which then would be seen as a group. The grouping of objects by perc eivers, however, depends strongly on the underlying concept of the objec ts considered. Objects understood to share the same concept tend to be grouped together and to constitute a category (Murphy and Medin 1985). Therefore, perceived entitativity co uld be influenced not only by sponsorship activities, but also by the concepts that consumers asso ciate with the sponsoring brands and the event. For instance, if the sponsors and the event are related to the concept of sport, perceived entitativity for the group will be stronger th an if the sponsors are not related to the concept of s port. It is proposed that brand-concept similarity will moderate the role of multiple sponsorships in the perceived entitativity of the brands and the event. Multiple sponsorships will gene rate the creation of an entitative group, but only for sponsors with the same brand-concept as the event. Categorization and conceptual coherence Mervis and Rosch (1981) defined categorization as one of the most fundamenta l processes performed by living creatures.
40 A category exists when two or more distingu ishable elements are treated equivalently; these elements can be anything including objec ts, persons, events, or ideas (Mervis and Rosch 1981). Psychologists have argued th at the objective of categorization for the perceiver is to achieve the highest cognitive efficiency, which is obtained when the categories maximize within-category similarity relative to between-category similarity (Medin and Schaffer 1978; Rosch and Mervis 1975). In line with that stream of research, accentuation theory (Krueger and Clement 1994; Tajfel 1959) posits that categorization leads to an assimilation effect within the cate gories and to a contrast effect between these categories. In other words, perceivers atte mpt to minimize differences among stimuli that fall in the same category (assimilation) and to maximize the differences among stimuli that fall into different cate gories (contrast). In addition, categoriza tion of stimuli has been shown to impede the individualization of these stimuli. Stangor and McMillan (1992) found that stimuli in the same category ar e likely to be confused with each other. This finding is consistent with Crawford et al.s (2002) results th at group entitativity leads group members to be perceived as interchangeable. An important question relative to cate gorization concerns the basis on which it takes place; how are elements assigned to different categories by perceivers? Murphy and Medin (1985) provided an explanation of how categories are created based on conceptual consistency. This is an alternative to the classical view (i.e., by a set of defining attributes), the probabilistic view (i.e., by a set of typical attributes) or the exemplar view (i.e., by the exemplars). Murphy and Medin (1985) developed the argument that elements are grouped together on the basis of how coherent the re lationships between th eir underlying concept
41 are. These inter-concept relationships depend on the perceivers theo ry about the world. For example, Medin (1989) reported that people consider the terms white hairs and grey hairs to be more similar than the terms grey hairs and black hairs, whereas the terms white clouds and grey clouds were judged to be less similar than the terms grey clouds and black clouds This shows how elements can be grouped together differently depending on their underlying concept (i.e., ag ing or bad weather) as well as on the theory perceivers use to arti culate these concepts (i.e., darker colo rs indicating younger age for hairs and worse weather for clouds). Barsalous (1983) research on goal-derived categories illustrate s the importance of theories in establishing conceptual struct ure. According to him, people rely on their own theories about the world to categorize together elements they think fulfill similar goals. He found that a disparate set of elem ents composed of objects such as children, photo albums, paintings, manuscripts, and jewe lry could be categorized together under the label taking things out of ones home during a fire. Brand-concept consistency Brand concepts are abstra ct meanings that consumers uniquely associate to brands that result fr om product features (e.g., a higher price can induce a high status meaning) and the fi rm marketing activities (Park, Milberg and Lawson 1991). The marketing literature recogni zed the role of consumers a-priori category structure in mediating information processing. Cohen and Basu (1987) as well as Sujan (1985) argued that categories to which consumers assign products and brands impact how information is processed. Suja n (1985) found that when information about a brand was consistent with the knowledge about the category of that brand, perceivers relied on category-based information proces sing; whereas, when information about a
42 brand was not consistent with the knowledge about the category of that brand, perceivers relied on individuating information processing (Fiske and Neuberg 1990). Murphy and Medins (1985) work suggests th at the concepts associated to the brands are key for understanding how consum ers categorize the brands. In a brand extension study, Park, Milberg and Lawson ( 1991) found that the better the extension product accommodates the brand-concept, th e more likely the brand name and the extension product will be categor ized together (i.e., seen as belonging to the same group) and the more favorably the brand extension is perceived. 1 Martin and Stewart (2001) studied how a categorization process mediates th e impact of different types of similarity on attitudes and purchase intenti ons within the context of a br and extension. Their results showed that, under a situation of goal congr uency, moderate goal incongruency, or extreme goal incongruency between the brand and the extension product, the impact of similarity on attitude and purchase inten tion was mediated by the abstract meaning associated to the brands and the extension produ cts. Both these studies offer considerable support for taking into account brand-concep t when investigating the basis on which consumers categorize brands. Based on the above discussion, it is asse rted that perceived entitativity is a function of both the presence of a multiple sponsorships agreement and the extent to which the event and the sponsors sh are the same brand-concept. 1 For brands with either a functional or a prestige brand-concept.
43 Savings in Relearning as a Measure of Implicit memory Johar and Pham (1999) showed that sponsor identification is bi ased by heuristics that consumers rely on when they cannot retr ieve information directly from their explicit memory. These heuristics pertain to the diffe rent contingent processes that consumers use when trying to identify message source (Pham and Johar 1997). Specifically, when asked to identify the correct sponsor of an event among a lis t of alternatives, consumers tended to attribute the sponsorship of an ev ent to a brand prominent in a product category or to a brand semantically related to the event sponsored. These findings show that consumers have difficulties remembering what brand sponsored a given event. However, as Johar and Pham (1999) suggested, their results only concern explicit memory and consumer learning about spons oring brands could take a more implicit form during sponsorships arrangements (Pham and Vanhuele 1997; Pracejus 1998). Pham and Vanhuele (1997) showed that a dvertising fragments (i.e., advertising messages that only include the brand name or a few words expressing the brands positioning) impact the implicit knowledge of c onsumers about brands rather than their explicit knowledge. The Implicit Association Test measures implicit attitude based on response latency (Greenwald, McGhee, and Sc hwartz 1998) and has been shown to be useful in assessing consumer implicit cogni tion (Brunel, Tietje, and Greenwald 2004). However, it is not well suited for the objectives of this dissertation for two reasons: 1) it does not allow investigating th e processing mechanisms that entitativity triggers (i.e., category-based versus individua ting), and 2) it does not allow investigating implicit learning because it focuses on the strength of the implicit associations between concepts in the consumer social knowledge structur e (Greenwald et al. 1998). As it will be
44 subsequently shown, the savings in relearni ng paradigm allows investigating implicit memory (Carlston and Skowronski 1994; Ebbinghaus 1885/1964). Abstraction as Implicit Knowledge Creation Research has shown that the abstracti on of the underlying characteristics of a group due to stereotyping results in the loss of individual-level information, which leads the members of the group to be seen as interchangeable. Once a group impression has been abstracted, the specific information that was used to create that group impression is forgotten (Crawford et al. 2002). Findings in the attitude domain conf irm this assertion. Once a global attitude toward an object ha s been formed, the or iginal information on which the attitude is based cannot be recalled (Lingle and Ostrom 1979). In fact, the passive abstraction of the deep underlying structure of the group, its essence as previously seen (Yzerbyt et al 1997), is the mechanism that characterizes the creation of implicit knowledge (Matthews et al. 1997; Reber 1989; Whittlesea and Wright 1997). Implicit knowledge is intuitive and results from an unconscious inductive abstraction of the underlying structure of complex environmental stimuli (Reber 1989). Implicit learning is more robust over time th an explicit learning (Matthews et al. 1987; Seger 1994) and is directly related to implicit memory (Seger 1994). In fact, there is evidence that implicit knowledge about a phenomenon can remain long after explicit knowledge about that phenomenon no longer ex ists. This implies that even though an individuals remembrance of particular inform ation related to a bra nd after exposure to a sponsorship might be limited, his/her global impression of that brand could still be implicitly remembered. Therefore, the abst raction of a group impression resulting from
45 multiple sponsorships is likely to take place at an implicit level in consumer memory. Stimuli conveyed by sponsorship operations are not salient (Pham and Johar 1997; Pham and Vanhuele 1997) and are like ly to operate below the c onscious level of consumers (Greenwald 1992). The research paradigm used in this dissertation will allow studying if higher entitativity results in th e formation of implicit knowledge that is stored in implicit memory and subsequently impacts performance on cued-recall trials. Implicit memory. The field of experimental ps ychology distinguishes between explicit memory, which is the conscious reco llection of information from a previous learning episode, and implicit memory, which is the unconscious impact of previously learned information on the facilitation of test performance (Schacter 1987). Explicit and implicit memory are dissociated systems that work independently from each other. Amnesic patients have been shown to suffe r from decreased performance on explicit memory tests such as free recall or re cognition while their performance on implicit memory tests such as repetition priming (i.e ., partial words completion after having been exposed to the entire word) was as good as c ontrol patients (Roediger 1990). In addition, explicit memory performance has been show n to be greater than implicit memory performance in the case of conceptual prim ing (i.e., presentation of cues conceptually related to the target stimulus: respondents ar e given the word tool to prompt the word hammer they saw earlier), whereas implicit memory performance has been shown to be greater than explicit memory performance in the case of perceptual priming (i.e., presentation of a perceptually degraded ve rsion of the target stimulus as a cue: respondents are given the word treasure to pr ompt the word treason they saw earlier)
46 (Smith and Branscombe 1988). The previously mentioned evidence indicates that explicit and implicit memory tap different forms of retention (Roediger 1990). Neuro-psychologists have suggested that these two types of memory correspond to two different systems in the brain. Exp licit memory corresponds to the declarative memory system, whereas implicit memory corresponds to the procedural system (Cohen and Squire 1980). On the other hand, cogniti ve psychologists have argued that these two types of memory correspond to two different types of processing. Explicit memory relates to conceptually-driven processing whereas implicit memory relates to perceptually-driven processi ng (Roediger, Weldon, and Challis 1989). A theoretical perspective has been proposed that integrat es the neurological a nd cognitive approaches of memory. Tulving and Schacter (1990) ar gued that the different memory systems of the brain have to operate through different c ognitive processes, which is in line with the position of Hayman and Tulving (1989). As su ch, they proposed that explicit memory is based on a declarative system that operates through conceptual pr ocessing and implicit memory is based on a procedural system that operates through per ceptual processing. Savings in Relearning During the early 18 th century, Goottfried Wilhelm Leibniz pointed out that although some things could not be remember ed directly, they could be more easily conceived if they were formerly know n (Leibniz 1916). Similarly, Ebbinghaus (1885/1964) noticed that some retentions are concealed from consciousness but have effects that are significant, which demonstrates their previo us experience. Ebbinghaus observation corresponds to the modern defin ition of implicit memory (Roediger 1990).
47 Because in the late 19 th century all the measures of memo ry were measures of explicit memory (i.e., free recall, recognition), Ebbi nghaus (1885/1964) created a way to measure implicit memory: the savings in relearning met hod. In its original version, the savings in relearning method consisted of subjects learning nonsense sy llables and then recording the number of trials or the amount of time to achieve a perfect recita tion of the syllables. At a later time, the same process was repe ated requiring the subjects learn the same material again (after a delay or a distrac ting task). The saving s in relearning were measured as the difference in the number of trials or in the amount of time needed between the second and the first learning phase. Subjects attempt to learn the material the second time should take less trials or less time. When subjects cannot consciously remember the previously learned material savings during the re learning phase are a quantitative estimation of their implicit knowle dge about the material learned during the first phase. Savings in relearning effects have been s hown even after several years have past since the first learning phase (Titchener 1923). The savings in relearning method is still regularly used in the field of experimental and social psychology. Recent studies include spontaneous trait inferences of social act ors (Carlston and Skow ronski 1994, 1999) or trait transference among group members in a so cial setting (Crawford et al. 2002). Over the years, several characteristics of the sa vings in relearning method have been pointed out that are of particular inte rest for the purpose of this disse rtation. First, savings effects have been reported with material presented once rather than presented until memorized perfectly (i.e., learned to cr iterion) (Nelson 1985). Second, such effects hold not only for word pairs, but also for differe nt combinations of verbal an d pictorial materials. Third,
48 the number of trials or the amount of time as a measure of savings in relearning has been replaced with a comparison of cued recall for old (relearned) versus new (control) material (e.g., Carlston and Skowronski 1994,1999; Crawford et al. 2002; MacLeod 1988; Nelson 1985). Implicit Image Transfer Sponsorships effects on the images of sponsoring brands are likely to be subtle (e .g., Johar and Pham 1999; Pham and Vanhuele 1997; Pham and Johar 1997). In addition, our conceptualizati on posits that the images characterizing the brands and the event are abstracted into a group stereotype as a function of their brandconcept similarity and transferred from th e group to each brand due to category-based processing (Fiske and Neuberg 1990). Therefor e, image transfer effects might be below the consciousness level of individuals. As such, the savings in relearning paradigm will be used because savings in this paradigm constitute a measure of implicit memory. Previous studies in social psychology ha ve relied on a modifi ed version of the savings in relearning paradigm in order to capture trait tran sference about social actors at an implicit level. Skowronski et al. (1998) showed that a trait implied by a behavior influenced the judgment of the person comm unicating that behavior when referring to another person. If somebody describes a person as being brave, the person communicating this statement woul d be judged as being brave as well. Skowronski et al. (1998) referred to this as Spontaneous Trait Transference because the trait brave was transferred from the braveness behavior to the communicator of that behavior. Also relying on the savings in relear ning paradigm, Crawfo rd et al. (2002) showed that such trait tran sference could occur for other group members and for the group as a whole. In their study, th ey conceptualized the notion of Trait-Inference and
49 Trait-Transference. The subjects were first presente d with a behavioral description of individual group members that clearly implied a particular trait (e.g., intelligent). Then, they had to learn the association between each individual member of the group they had seen and either a word that matched the trai t implied by the behavior of that individual (e.g., intelligent) or a word that did not matc h the trait implied by the behavior of that individual, but that matched the trait implied by the behavior of anot her individual of the group (e.g., lazy). When the learning task i nvolved a trait and a word that matched, it was labeled as a Trait-Inference trial. When the learning ta sk involved a trait and a word that did not match it was labeled as a Trait-Transference trial. They found that subjects performance on Trait-Transference trials was higher when th e group of individuals was high on entitativity, whereas it was greater on Trait-Inference trials when the group was low on entitativity. In this research, the transference of so cial actors traits presented above is extended to the transfer of image among the bran ds as well as from an event to the brands during multiple sponsorships situations. Ba sed on Crawford et als (2002) work, the notion of Brand Image Reinforcement (BIR), as well as the notion of Brand Image Transfer (BIT) and Event Image Transfer (EIT), are conceptualized. Brand Image Reinforcement (BIR) refers to a situation in whic h subjects are asked to memorize the association between a brand and a word (i.e., image-trait) that is consistent with the image of the brand. Brand Image Transfer (BIT) refers to a task for which subjects have to learn the association between a brand a nd a word (i.e., imagetrait) that is not consistent with the image of that brand, but is consistent with the image of another brand sponsoring the event. Event Image Transfer (EIT) refers to a task for which subjects
50 have to learn the association between a brand and a word (i.e ., image trait) that is not consistent with the image of that brand, but is consistent with the image of the sponsored event. The extent to which subjects can better remember Brand Image Transfer (BIT) trials, as well as Event Image Transfer (EIT) trials, in a multi ple sponsorship scenario rather than in a no sponsorship condition constitutes eviden ce of image transfer among multiple sponsoring brands and from th e event to the brands, respectively. Summary and Hypotheses Development Hypotheses are formulated concerning both the outcomes of multiple sponsorships in terms of image transfer a nd the psychological process responsible for these outcomes. It is expected that ther e will be a transfer of image among the sponsoring brands, as well as from the event to the sponsoring brands, due to multiple sponsorships, but only for the sponsors with the same brand-concept as the event (i.e., outcome predictions). In addition, these tran sfer phenomena are expe cted to be due to category-based processing as opposed to individuating processing (i.e., process predictions). Outcome Hypotheses There are two factors that can impact the perceived entitativity of an event and some brands: 1) whether or not the brands ar e sponsoring the event concomitantly and, 2) whether or not the sponsors share the same brand-concepts as the even t. In the case of multiple sponsorships, the perceived entitativi ty of the sponsoring brands and the event will increase due to their proximity and interd ependence (Lickel et al. 2000). In the case
51 of no sponsorship, the brands and the even t will not be perceived as a group and, therefore, entitativity w ill be not be impacted. On the other hand, if the sponsors and th e event share the same brand-concept, they will form a group and consumers will perceive them as belonging to the same category (Murphy and Medin 1985). Accentuation theory (Krueger and Clement 1994; Tajfel 1959) indicates th at stimuli categorized together become more sim ilar through an assimilation effect whereas stimuli assi gned to different categories become more dissimilar through a contrast e ffect. Therefore, if the s ponsors and the event have the same brand concept, they will be perceived as similar to each other. If the sponsors and the event do not share the same brand-concep ts, they will be perceived as dissimilar to each other. According to Campbell (1958), the entitativity of a group is a positive function of how similar the elements of that gr oup are perceived to be. Therefore, if the event and the sponsors share the same brand-co ncept, it will increase their entitativity; whereas, should they have divergent brand-conc epts, it will decrease their entitativity. The above discussion indicates that multip le sponsorships and brand-concept are related to entitativity differently. They both increase entitativity wh en the brands engage in multiple sponsorships or have the same brand-concepts as the event. As suggested previously, the absence of multiple sponsorship s does not impact entitativity. When the brands do not engage in multiple sponsorships, however, the influence on entitativity is different from when they do not have the sa me brand-concepts. If the sponsors do not have the same brand-concepts, they will b ecome more dissimilar (i.e., brand-concept dissimilarity) (Krueger and Clement 1994; Tajfel 1959). Ba sed on Campbell (1958), this suggests that brand-concept diss imilarity will actually decrease the perceived entitativity
52 of the group. As a consequence, given a situ ation of multiple sponsorships, if the brandconcepts of the sponsors are similar, the enti tativity of the group will be greater than if the brand-concepts are dissimilar. The higher entitativity of the group of sponsors and the event will result in the abstraction of a group impression in light of which each group member will be perceived. Since the group impression is composed of the images of the individual members, it is abstracted from each sponsoring brands image and will be associated with the images of the other sponsors, as well as of the event. In the context of th e savings in relearning paradigm, higher entitativity should facil itate the memorization of a brand and an inconsistent image-trait ( Brand Image Transfer trials, BIT and Event Image Transfer trials, EIT). In such a learning task, because each brand is already associated with the other brands and the events images, re spondents will only have to relearn these associations and not learn them for the firs t time, which should gene rate more savings compared to low entitativity. As a result, respondents should exhibit more savings in the multiple sponsorships than in the no sponsorship condition for Brand Image Transfer trials (BIT) or Event Image Transfer trials (EIT) concerning sponsor s that share the same brandconcepts as the event. In that case, both the sponsorship and th e brand-concept factors increase the entitativity of the group. BIT or EIT trials for sponsors with dissimilar brand-concepts, however, should not gene rate greater savings in the multiple sponsorships than in the no sponsorship condition because the two factors (i.e., sponsorship and brand-concept) will contribute in opposite directions to the entitativity of the group. Since both multiple and no sponsorsh ip conditions will have a low entitativity,
53 respondents will not abstract a group impression in either condition. Inconsistent pairedassociations will have to be learned for the first time in both cases and respondents should not have more savings in multiple sponsorships as compared to no sponsorship. Given the above discussion, the following is hypothesized: H1a : The recall of Brand Image Transfer (BIT) trials for sponsors with similar brand-concepts (e.g., sport bra nd) will be greater in the multiple sponsorships versus the no sponsorship condition. H1b: The recall of Brand Image Transfer (BIT) trials for sponsors with dissimilar brand-concepts (e.g., no sport brands) will not be significantly different in the multiple sponsorships versus the no sponsorship condition. H2a: The recall of Event Image Transfer (EIT) trials for sponsors with similar brand-concepts (e.g., sport brands) will be greater in the multiple sponsorships versus the no sponsorship condition. H2b: The recall of Event Image Transfer (EIT) trials for sponsors with dissimilar brand-concepts (e.g., no sport brands) will not be significantly different in the multiple sponsorships versus the no sponsorship condition.
54 Processes Hypotheses It is posited that entitativity will trig ger category-based processing as opposed to individuating processing (Fis ke and Neuberg 1990). Category-based processing implies that information about a group member is asso ciated with all the other members of the group while individuating proces sing implies that informati on is uniquely associated to each member. Respondents exposed to mu ltiple sponsorships should process the sponsors with similar brand-concepts in a category-based manner. Therefore, a given image-trait characterizing a brand will be less uniquely linked to that brand if entitativity is high. As a consequence, it should be mo re difficult for respondents to learn the association of an image-trait a nd a brand that are consistent ( Brand Image Reinforcement trials, BIR) in the case of multiple sponsorsh ips because the image-trait should be less strongly associated with the brand th an in the case of no sponsorship. When the sponsors have dissimilar brand-concepts, however, they will not be categorized together and entitativity will re main low in both conditions. As a result, when memorizing a consistent paired-associ ation for sponsors with dissimilar brandconcepts, respondents should have the same amount of savings (recall) in multiple sponsorships as in no sponsorship. Savings (recall) should be lower in multiple sponsorships compared with no sponsorshi p when the sponsors have similar brandconcepts. In addition, in multiple sponsors hips, respondents should show lower savings (recall) when memorizing consistent paired-a ssociations for sponsors with similar brandconcepts than for sponsors with dissimilar br and-concepts. Therefore, the following is hypothesized:
55 H3a: The recall of Brand Image Reinforcement (BIR) trials for sponsors with dissimilar brand-concepts (e.g., no sport brands) will not be significantly different in the multiple versus the no sponsorship condition. H3b: The recall of Brand Image Reinforcement (BIR) trials for sponsors with similar brand-concepts (e.g., sp ort brand) will be lower in the multiple sponsorships versus the no sponsorship condition. H4: In the multiple sponsors hips condition, the recall of Brand Image Reinforcement (BIR) trials will be greater for sponsors with dissimilar brand-concepts (e.g., no sport brand) than for sponsors with similar brandconcepts (e.g., sport brand). High group entitativity should lead group me mbers to be associated with all the image-traits characterizing the group due to category-based processing. Low entitativity should lead individual members to be associated uniquely with their own image-traits due to individuating processing. In high entitativi ty, the association of an image-trait and a brand that do not match should be as easy to memorize as when they match. As a result, savings (recall) on consistent paired-associ ations should not be greater than savings (recall) on inconsistent paired-associations when entitativity is high, but they should be greater for sponsors with dissimilar brand-co ncepts when entitativity is low. As suggested before, high entitativity is only warranted for sponsors with similar brand-
56 concepts in multiple sponsorships. Dissi milar brand-concepts or the absence of sponsorship will lead to lower levels of entitativity. Thus the following is hypothesized: H5a: In the case of no sponsorship, the recall of Brand Image Reinforcement (BIR) trials will be greater than the recall of Brand Image Transfer (BIT) trials for sponsors with similar brand-concepts (e.g., sport brand). H5b: In the case of multiple sponsorships, the recall of Brand Image Reinforcement (BIR) trials will not be significantly different from the recall of Brand Image Transfer (BIT) trials for sponsors with similar brand-concepts (e.g., sport brand). The pattern of effects po sited above should be the same when comparing Event Image Transfer (EIT) trials to Brand Image Reinforcement (BIR) trials; therefore, the following is hypothesized: H5c: In the case of no sponsorship, the recall of Brand Image Reinforcement (BIR) trials compared to Event Image Transfer (EIT) trials will be greater for sponsors with simila r brand-concepts (e.g., sport brand). H5d: In the case of multiple sponsorships, the recall of Brand Image Reinforcement (BIR) trials compared to Event Image Transfer (EIT) trials
57 will not be significantly different for sponsors with similar brand-concepts (e.g., sport brand). As seen before, for low entitativity, sa vings on consistent paired-associations should be higher than savings on inconsiste nt paired-associations. Sponsors with dissimilar brand-concepts will not be categorized together and, therefore, will not form a group, which will keep entitativity low even in the case of multiple sponsorships. Therefore, the following is hypothesized: H6a: In the case of no sponsorship, the recall of Brand Image Reinforcement (BIR) trials compared to Brand Image Transfer (BIT) trials will be greater for sponsors with dissimilar brand-concepts (e.g., no sport brand). H6b: In the case of multiple sponsorships, the recall of Brand Image Reinforcement (BIR) trials compared to Brand Image Transfer (BIT) trials will be greater for sponsors with dissimilar brand-concepts (e.g., no sport brand). As before, the pattern of effects pos ited above should be the same for Event Image Transfer trials; thus:
58 H6c: In the case of no sponsorship, the recall of Brand Image Reinforcement (BIR) trials compared to Event Image Transfer (EIT) trials will be greater for sponsors with dissimilar brand-concepts (e.g., no sport brand). H6d: In the case of multiple sponsorships, the recall of Brand Image Reinforcement (BIR) trials compared to Event Image Transfer (EIT) trials will be greater for sponsors with dissimilar brand-concepts (e.g., no sport brand). Higher entitativity leads information to be associated with each group member and the group as a whole due to category-b ased processing (Fiske and Neuberg 1990). This results in a loss of individuality for group members, which will be perceived as interchangeable (Crawford et al. 2002). As a consequence, the higher the perceived entitativity, the more weakly the respondents should associate the brands with their tagline and the more difficult it should be fo r the respondents to recognize a sponsoring brands correct tag-line. As we saw earlier, sponsors with a brand-concept similar to that of the event will already be perceived as a group through a categorization phenomenon (Murphy and Medin 1985). Sponsors with simila r brand-concepts will be perceived as somewhat entitative whether or not they sponsor the event. Therefore, it is expected that the brands will be more uniquely associated with their respective tag-lines in the case of no sponsorship compared to multiple sponsorships only when the sponsors brandconcepts are dissimilar. When the sponsors brand concepts are similar, they should
59 already form an entitative group and the tag-lines should be weakly associated with their brands for both multiple sponsorships and no sponsorship. Therefore, the following is hypothesized: H7a: Tag-line recognition for sponsors with similar brand-concepts (e.g., sport brand) will not be significantly di fferent in the no sponsorship versus the multiple sponsorships condition. H7b: Tag-line recognition for spons ors with dissimilar brand-concepts (e.g., no sport brand) will be greater in the no sponsorship versus the multiple sponsorships condition. In this chapter, the relevant litera ture on sponsorship, brand associations, experimental psychology, and social psychology was reviewed. It pr ovided a rationale for the conceptualizatio n of multiple sponsorships as a highly entitative situation, which leads the sponsors and the event to be grouped together when they have the same brandconcepts. Based on the previous elements, Brand Image Transfer (BIT) and Event Image Transfer (EIT) phenomena were predicted in the case of multiple sponsorships and formulated in a hypothesis format. In the ne xt section, a design will be developed which allowed testing these hypothe ses through an experiment.
60 CHAPTER 3 Methodology Overview The objectives of this disser tation are threefold. First, to provide evidence of image transfer among concomitant sponsors of the same event, as well as from the event to its sponsors. Second, to show the implic it aspect of these image transfer phenomena by investigating them with the savings in relearning paradigm. And finally, to demonstrate that these effects are due to respondents processing of information on a category-based manner as opposed to an individuating manner. The respondents were split between a multiple sponsorships condition and a no sponsorship condition. All the respondents were first exposed to eight sponsoring brands; four of these brands had an exciting image and the other four had a sincere image. Respondents in the multiple sponsorships condition were also exposed to a sophisticated sporting event. This allo wed respondents to associate specific imagetraits with the brands and the event and to abstract a group impression (i.e., in the multiple sponsorships condition). Respondents we re then presented with four consistent paired-associations that constituted the Brand Image Reinforcement (BIR) trials and four inconsistent paired-associations. For half of the respondents, the four inconsistent pairedassociations were between one of the sponsor ing brands and either the word sincerity or exciting, that constituted the Brand Image Transfer (BIT) trials. For the other respondents, the inconsistent paired-associations were be tween one of the sponsoring
61 brands and the word sophisticated that constituted the Event Image Transfer (EIT) trials. After this pa ired-associations task, respondents were presented with a cued-recall task during which the sponsoring brands previously seen were shown again. Respondents had to indicate the image-trait that was paired with each brand during the previous paired-association task. Eventually, respondents had to identify the correct tagline of each brand. Respondents and Experimental Design One hundred and seventy-one students from a large state university received extra-credit for participating in this experiment. They were randomly assigned to a 2 (sponsorship: multiple vs. no) x 2 (image-trait: brand related vs. event related) x 8 (counterbalancing factor) x 2 (paired-associatio ns: consistent vs. inconsistent) x 2 (brandconcept: similar vs. dissimilar) mixed-measures design with repeated measures on the last 2 factors (see Appendix 4 for a graphical representation of the design). Sixteen respondents were dropped from the analysis du e to obvious misunderstanding of the task or for missing data on the cued-recall task. This resulted in a usable sample of 155 respondents with cell sizes ranging from n = 36 to n = 41. Respondents were told they were participating in a study concerning th e quality of the web sites of several new brands. In addition, they were told that they might not be aw are of the brands they were exposed to because they were not available nationwide yet and were in a phase of test market in the northeast of the country.
62 Materials Brand personality is an important co mponent of brand image (Aaker 1997; Gwinner and Eaton 1999) and is defined as the set of human characteristics associated with a brand (Aaker 1997, p. 347). As a consequence, Brand Image Reinforcement (BIR), Brand Image Transfer (BIT), and Event Image Transfer (EIT) trials were assessed through image-traits operationalized by adjec tives describing specific brand personality dimensions. In order to avoid confounding e ffects due to the preex isting image of real brands and of real events, the eight sponsoring brands and the event were fictitious. Aakers (1997) brand personality scale was used to operationalize the specific image-traits assigned to the experimental stim uli (i.e., the event and the brands). Aakers (1997) brand personality scale is a 42-item inst rument that captures the 5 dimensions of the brand personality construct: sincerity, excitement, competence, sophistication, and ruggedness. It was decided to rely on the dime nsions of sincerity and excitement in order to create the brands used as experimental stimuli. Sincerity and excitement were chosen because they represent most of the varian ce in personality ratings of brands across individuals, product categorie s, or cultural contexts (A aker, Benet-Martnez, and Garolera 2001; Aaker, Four nier, and Brasel 2004; Caprar a, Barbaranelli, and Guido 2001). Four of the sponsoring brands conveye d a sincere image a nd four conveyed an exciting image. Images of excitement and sincerity were generated through four different venues on the web pages: 1) brand name, 2) product information, 3) logo, and 4) tag-line (see Appendix 2 for the se lected experimental stimuli).
63 Pretest A pretest ensured that these web page s appropriately created sincerity or excitement. This pretest was conducted by prov iding 30 students with fictitious brands associated with sincere avenues on their web pages and others associated with exciting avenues. Respondents rated the degree to which each brand could be described by sincere traits (e.g., sincere, w holesome, sentimental, family-oriented), exciting traits (e.g., exciting, unique, young, trendy), as well as s ophisticated traits (e.g., upper class, glamorous, good looking, charming) using items from Aakers (1997) 7-point Likert-type brand personality scale (1 = strongly disagree, 7 = strong ly agree) (see Appendix 3 for items). The validity of the experimental stimuli was established by making sure that: 1) the exciting brands yielded higher ratings on the exciting traits than on the sincere traits, 2) the sincere brands yielded higher ratings on the sincere traits than on the exciting traits, 3) the sincere brands obtained higher sincerity ratings than the exci ting brands, and 4) the exciting brands obtained higher excitement ratin gs than the sincere brands, 5) the event received higher sophistication ratings than the exciting and sincere brands, 6) the event received lower excitement ratings than the exciting brands and lower sincerity ratings than the sincere brands. In addition, the sponsors had to satisfy the criterion of similari ty/dissimilarity for their underlying brand-concept: half of the spons ors had to have the same brand-concepts as the event (i.e., sport) and the other half had to have a dissimilar brand-concept (i.e., no sport). Furthermore, it was important to verify that the manipulations of brand personality did not differ on vari ables that could threaten the validity of the experimental treatments. The personal relevance of the bra nds (the brand image is relevant to me,
64 the brand image makes sense to me) as well as the relevance of the category (relevant in [product/service category], makes se nse in [product/service category]) were measured on a 7 points Likert-type scale (1 = strongly disagree, 7 = strongly agree) (Aaker et al. 2004) (see Appendix 3 for items). More than 30 brands and several versions of the event were pretested using different samples of undergradu ate students (due to responden ts fatigue and the iterative nature of this selection process, the same sample could not rate all the brands and versions of the event). Significance levels we re used to make sure that the brands and event selected within each sample satisfied the six criteria listed above. It was decided to select the brands and the event based on the individual samples significance tests and to include a manipulation check of the brands and the events images in the main experiment. 2 As a result, eight sponsoring brands and one event were se lected that satisfied the six criteria established above had the appropriate brandconcept, and had brand and category relevance ratings significantly higher than the neutral point of the scale (i.e., > 4) (Table 5-1 in Appendix 5 presents the pr etest results for the ei ght sponsoring brands and the event chosen). The four sincere br ands were: Fitness and Health, which are two fitness centers, Massachusetts Bank, which is a financial institution, and Aunt Marys Gourmet Treat, which is a brand of cookie. Two of these sincere sponsors had sport as an underlying brand-concept simila r to the events (i.e., Fitness and Health), and the other two had a brand-concept differe nt from each other and dissimilar to the 2 In the experiment, in order to limit respondents fatigue due to the length of the brand personality scale (i.e., 42 items for each brand or event), a factor an alysis was performed on the pretest data and the two adjectives that explained the most variance for each personality dimension (i.e., sincere, exciting and sophisticated) were included.
65 events (i.e., Massachusetts Bank and Aunt Mary). The four exciting brands were: Glissade, which is a sportswear manufacturer Energetic, which is an energy drink, Urbane, which is a consumers electronics brand, and Night-Club, which is a club. Two of these exciting brands had sport has an underlying brand-concept (i.e., Glissade and Energetic), similar to the events, and the other two had a brand-concept different from each other and dissimilar to the events (i .e., Night-Club and Urbane). This allowed a balance between brand-concept and brand image (see Table 1). In addition, the version of the 2005 Boston Golf Tournament with the highest sophistication ratings was selected as the event. Table 1. Concept Similarity and Images of the Brands Used in the Experiment Sincere Image Exciting Image Sport Brand-concept Health Fitness Glissade Energetic No Sport Brandconcept Massachusetts Aunt Marys Gourmet Treat Urbane Night-Club Procedure An overview of each step of the pro cedure is provided in Figure 4. Each respondent received a package comprised of four booklets that corresponded to a different phase of the study (i.e., exposure to target brands, exposure to foil brands, memorization task, and cued-recall/recogniti on tasks). On each page, the booklets presented the respondents with a one-page excerpt of the web site of th e fictitious brands. It was important that all respondents fo llowed the same pace. Respondents were
66 instructed to wait when finish ed with a booklet and not to ge t started on another one. The entitativity manipulation th rough sponsorships was received at the beginning of the experimental session before respond ents started the exposure phase. Figure 4. Methodological Steps of the Experimental Procedure Step Number Description Variables Involved 1-Sponsorship Manipulation The entitativity treatment is given to the respondents through a multiple sponsorships (high entitativity) vs. no sponsorship manipulation (low entitativity). Independent Variable 1 (between-subject factor): Sponsorship (multiple vs. no) 2-Target Brand Exposure Subj ects are educated about the images of 8 fictitious brands plus the event for those in the multiple sponsorships condition. Independent Variable 2 (within-subject factor): Brandconcept (similar vs. dissimilar). 3-Foil Brand Exposure Subjects are provided with an additional set of 6 fictitious brands. 4-Paired-Associations Task Subjects are asked to memorize the association between each of the previous brands and an image-trait. Independent Variable 3 (within-subject factor): Pairing (consistent vs. inconsistent). Independent Variable 4 (between-subject factor): Imagetrait (brand related vs. event related). 5-Filler Task Subjects have to search in a letter matrix the names of famous Fortune 500 companies CEOs. 6-Cued-recall Task Subjects are provided with the 8 target brands and have to write down with which image-trait each one was previously associated. Dependent Measure 1 : Cuedrecall. 7-Tag-line Recognition Task Subjects are provided with the 8 target brands and have to identify the correct tag-line for each brand out of a choice of 4 tag-lines. Dependent Measure 2 : Tagline Recognition. 8-Manipulation Checks and Covariate Subjects are administered scales to asses the success of the manipulation of entitativity, brand-concept, and the images of the experimental stimuli. Need for Closure is also measured. Manipulation checks: Brandconcept; Exciting, Sincere and Sophisticated brand image; Entitativity. Covariate : Need for Closure
67 Sponsorships Manipulation Respondents were given one of the two sponsorship manipulations before the exposure task began (Independent Variable 1). Respondents in the multiple sponsorships condition were provided with a page that co ntained a two-paragraph statement about the eight brands they were about to see. The page mentioned that th e brands would all be sponsors of the 2005 Boston Golf Tournament, support the event, and participate in activities before, during, and after the event. In additi on, at the bottom of the page appeared an illustration of the different signa ges at the event locati on where logos of the event and of the sponsoring brands could be seen together (see Appendix 6 for the multiple sponsorships manipulation). In the multiple sponsorships condition, respondents were presented with a one-page excerpt of the web s ite of the 2005 Boston Golf Tournament. Furthermore, the web page s of each sponsoring brand mentioned the sponsorship of the 2005 Boston Golf Tourname nt. Respondents in the no sponsorship condition were given a written statement mentioning that th e brands were arbitrarily chosen from a pool of brands conducting test markets in the Northeast area. They were not provided with a one-page excerpt of the we b site of the event and the web pages of the brands did not mention any sponsorship agreement concerning the 2005 Boston Golf Tournament. Target Brands Exposure Every respondent was provided with a copy of a one-page excerpt of the web site of each of the four exciting and four sincer e sponsoring brands. Half of these sponsors had a sport brand-concept while the othe r half had a no sport brand-concept
68 (Independent Variable 2). Respondents were asked to familiarize themselves with the material presented in order to make sure they developed some knowledge about these fictitious brands to test for image transfer effects; respondents read the materials at their own pace. Respondents in the multiple sponsorships condition were also provided with an excerpt of the web site of the 2005 Boston Golf Tournament, which had a sophisticated image. The sequence of presen tation of the eight br ands was alternated within each condition in order to avoid order effects. Foil Brands Exposure The first exposure phase was followed by a task aimed at promoting the forgetting of the specific information contained in the we b pages so that, during the cued-recall task, the respondents had to rely on their remembra nce of the general impression of the brand rather than on their remembrance of the we b site (e.g., Crawford et al. 2002). Six new fictitious brands were presented to the re spondents through a one-page excerpt of their web site. The six brands matched the image-traits of the brands generated in the first exposure phase. Two of the brands had a sin cere image, two had an exciting image, and the other two had a sophisticated image (these images were determined through a pretest. See Table 5-2 in Appendix 5). Paired-Associations Task Participants were presented successively w ith pairs of brands and image-traits. Each pair of brand and image-traits was pr esented on a different page. The sequence of presentation was alternated across respondents and differed from the sequence of the
69 exposure phase to avoid order effect and vi carious learning of the sequencing of imagetraits. Eight of these pairings were composed of one of the target brands presented during the exposure task (i.e., through its one-page we b site excerpt) and one image-trait. Each target brand was paired with either the word sincere, exciting, or sophisticated (Aaker 1997). In addition, paired-associations between foil br ands and image-traits (i.e., competent and rugged) were provided to the respondents. Participants were given 15 seconds to me morize paired-associations. Studies in experimental and social psychology usually give respondents be tween six and eight seconds to memorize pairs of words or pi ctures (e.g., Carlston and Skowronski 1994; Crawford et al. 2002; MacLeod 1988). Since, in this study, respondents had to memorize pairings between more complex stimuli (i.e ., an entire web page and a word), it was decided to double the memorization time. Ther efore, the investigat ors indicated to the respondents to go to the next pa ge every 15 seconds. Out of the eight brands presented, four brands were associated with an image-trait that matched their image during the exposure task (i.e., consistent paired-associatio ns) (see Figure 4). For example, if a brand was portrayed as sincere through its web page it was now paired with the image-trait sincere. If respondents perc eived a brand as being sincer e after the exposure phase, they were now relearning something they al ready knew. These combinations constituted the Brand Image Reinforcement (BIR) trials. The four other brands were paired with an image-trait that did not match their image during the exposure phase (i.e., inc onsistent paired-associations, Independent Variable 3). For half of the respondents the inconsistent image-traits matched the image of other target brands (i.e., brand image related). For example, if a brand was portrayed
70 as sincere by its web page, it was now pair ed with the adjective exciting. These combinations constituted the Brand Image Transfer (BIT) trials as the transfer of imagetrait initially associated with other brands was now assessed. For the other respondents, these four brands were paired with an incons istent image-trait that matched the image of the event (i.e., event related image, Independent Variable 4). For ex ample, if a brand was portrayed as sincere by its web page, it was now paired with the adjective sophisticated. These combinations constituted the Event Image Transfer (EIT) trials as the transfer of image-trait initially asso ciated with the event was now assessed. In the case of high entitativ ity, the brands and the even t should be perceived as being interchangeable. Therefore, the res pondents should have already had associated the other brands and the events traits with the focal brand an d they are not asked to learn something new. They should exhibit more s avings than in the case of low entitativity. The BIT and EIT take place if the image-traits originally associated with the sponsoring brands and the event are abstracted into a group stereotype and become associated with all the concomitant sponsoring brands. For sponsors with similar brand concepts, it is expected that BIT and EIT will be greater. BIR will be smaller in the case of multiple sponsorships as compared to no sponsorship due to an increased perc eived entitativity. The pairing of the brands with the words sincere, exciting, and sophisticated was alternated so that a given brand was equally us ed as a BIR, BIT or EIT trial. This was done to ensure that savings effects we re not confounded by paired-associations of brands and words that could be easier to remember than others.
71 Filler Task Following the memorization task, a five minute filler task was given to the participants during which time they had to search a letter matrix for the names of wellknown Fortune 500 CEOs. This task was desi gned to clean up respondents short-term memory and served to both increase forgetting of the particularities of the web pages seen during the memorization task and as a filler be fore the cued-recall task (e.g., Crawford et al. 2002). Cued-Recall Task The eight target brands were presented one at a time on a separate page in an order different from during the exposure and learning phase to avoid recall from vicarious learning of the seque ncing. On each page appeared a blank text box in which participants were asked to write the exact im age-trait that was paired with that brand during the paired-associa tion task. Participants responses were recorded as correct if they were the same or a close synonym of the image-trait paired with the brand. The percentage of correctly recalled responses constituted the main dependent measure (see Figure 5). Tag-Line Recognition Task Respondents were provided with each of th e eight target brands on a separate page along with four tag-lines. In each case, one tag-line wa s the one presented with the web page of that brand during the exposure task (the correct tag-line ), two were tag-lines from other target brands, and one tag-line was from a foil brand. Respondents had to
72 indicate which tag-line was actually associ ated with each brand during the exposure phase (Dependent Measure 2). Figure 5: Examples of Exposure, Memorizat ion and Cued-Recall Tasks in the Brand Image Related and the Event Image Related Conditions Brand Image Related Event Image Related Brand Exposure Phase Paired Association Task Recall (DV) Exposure Phase Paired Association Task Recall (DV) Aunt Marys Treat Sincere Exciti ng BIT Sincere Sophisticated EIT Night-Club Exciting Exciting BIR Exciting Exciting BIR Fitness Sincere Sincere BIR Sincere Sincere BIR Glissade Exciting Sincere BIT Exciting Sophisticated EIT Massachusetts Bank Sincere Exciti ng BIT Sincere Sophisticated EIT Health Sincere Sincere BIR Sincere Sincere BIR Energetic Exciting Sincere BI T Exciting Sophisticated EIT Urbane Exciting Exciting BIR Exciting Exciting BIR BIR: Brand Image Reinforcement (% of correctly recalled paired associations) BIT: Brand Image Transfer (% of co rrectly recalled paired associations) EIT: Event Image Transfer (% of co rrectly recalled paired associations) Manipulation Checks and Covariate Respondents rated their perception of the entitativity of the brands presented during the exposure phase by indicating to what extent they considered that these brands qualified as a group by using a Likert-type sc ale adapted from Lickel et al.s (2000) manipulation check (1 = the companies do not form a group; 9 = the companies form a group). In addition, respondents rated the extent to which the eight target brands were sincere, exciting, and sophisticated usi ng items from the Aakers (1997) brand
73 personality scale. Furthermore, respondents indicated the extent to which each brand was related to the notion of spor t on a 7-point Likert-type sc ale (e.g., I associate [brand X] with the idea of sports). Finally, respondents filled out a sc ale purported to measure their need for closure (Webster and Kruglanski 1994). As previously explained, Need for Closure is the desire to obtai n any answer about a topic ra ther than to remain in a situation of ambiguity. People high in Need for Closure have been shown to rely more systematically on stereotypic processing (K ruglanski and Freund 1983). Therefore, the effect of Need for Closure on information pr ocessing is similar to perceived entitativity and might be confounded with our sponsorship manipulation. As a result, it was decided to include Need for Closure as a covariate during the analysis. Respondents ratings were made on a 42-item 6-point Likert-type scale developed by Webster and Kruglanski (1994). They had to agree or disagree with st atements such as Id rather know bad news than stay in a state of uncertainty or I di slike questions that could be answered in many different ways (see Appendix 3). This chapter laid out the design of th e experiment undertaken which is an adaptation of the general procedure used in the spontaneous trai t-inference and traittransference paradigm. Four factors were manipulated (multiple sponsorships, brandconcept, pairing consistency, and brand or event related image) in a mixed design and two dependent measures (cued-recall and tagline recognition) were collected. The next chapter will develop the analyt ic procedure used to test for the hypotheses previously developed. Some additional analyses will be performed as well.
74 CHAPTER 4 Results The cued-recall data yielded four dependent variables. The percentage of correct responses was tabulated separately for consis tent and inconsistent paired-associations. For both consistent and incons istent paired-associations, th is percentage of recall was then tabulated separately for sponsors with similar brand-concepts and for sponsors with dissimilar brand-concepts. In addition, the pe rcentage of correctly recognized tag-lines was tabulated separately for sponsors with si milar brand-concepts and for sponsors with dissimilar brand-concepts, which resulted in the two other dependent variables. A preliminary analysis revealed th at the three counterbalancing factors (i.e., reversed order of the brands during the first exposure phase, re versed order of the paired-associations, as well as alternated consistent/inconsistent paired-associations dur ing the memorization task) did not significantly impact any of the dependent variables. The main effects of these factors were not significant and neither we re their interactions with the factors of interest (i.e., sponsorship brand concept, and event/brand related image) and are therefore not discussed further. Manipulation Checks (see Appendix 3 for items) The sponsorship manipulation successfully increased the perceived entitativity of the group of brands and the event. The enti tativity ratings were higher in the multiple sponsorships condition than in the no sponsorship condition ( M multiple = 4.42 vs. M no =
75 3.44, t 153 = 3.28, p < .01). In addition, the checks s how that the manipulation of brandconcept was successful. All the brands that were expected to be perceived as sports brands (i.e., Energetic, Gli ssade, Health, and Fitness) had a mean score significantly greater than the mid-point (i.e., > 4) of the scale measuring th e extent to which a brand is related to the notion of sport. This confirmed that thei r underlying brand-concept was similar to the one of the event. Also, all the brands that were expect ed to be perceived as no sport brands (i.e., Massachusetts, Urba ne, Night-Club, and Aunt Mary) had a mean score significantly lower than the mid-point of the scale, which confirmed they had an underlying brand-concept different from the one of the event (see Table 7-1 in Appendix 7 for detailed results). Finally, the brand personality ratings showed that each brand had the image intended. As shown in Table 7-2 (Appendix 7) all the sincere brands had significantly greater sincerity ratings than the exciting brands and were rated significantly lower on sophistication than sin cerity. All the exciting brands had higher excitement ratings than the sincere brands and were rated significantly lower on sophistication than on excitement. Furthermore, all but two brands had lower sophistication ratings than the Boston Golf Tournament (i.e., Night-Club was significantly greater while Urbane was not significantly different from the event). A lthough Night-Club and Urbane were selected because they showed significantly lower sophistication ratings than the event during the pretest, these results did not re plicate. Contrary to the pret est, only two out of six items from the sophistication measure (Aaker 1997) were used for the manipulation checks in the study in order to limit questionnaire length and could explain this result.
Outcome Hypotheses: Implicit Brand and Event Image Transfer H1a: The recall of Brand Image Transfer (BIT) trials for sponsors with similar brand-concepts (e.g., sport brand) will be greater in the multiple sponsorships versus the no sponsorship condition (supported). H1b: The recall of Brand Image Transfer (BIT) trials for sponsors with dissimilar brand-concepts (e.g., no sport brand) will not be significantly different in the multiple sponsorships versus the no sponsorship condition (failed to reject). Figure 5a. Implicit Brand Image Transfer 67.163.447.661010203040506070SportNo Sport% of Correctly Recalled Inconsistent Paired-Associations No Sponsorship Multiple Sponsorships H1a predicted that implicit image transfer would take place among sponsors with similar brand-concepts and H1b predicted that implicit image transfer would not occur among sponsors with dissimilar brand-concepts. In Figure 5a, it can be seen that the recall of inconsistent paired-associations was greater in the multiple sponsorships 76
condition as compared to the no sponsorship condition only for sponsors with similar brand-concepts. H2a: The recall of Event Image Transfer (EIT) trials for sponsors with similar brand-concepts (e.g., sport brand) will be greater in the multiple sponsorships versus the no sponsorship condition (not supported). H2b: The recall of Event Image Transfer (EIT) trials for sponsors with dissimilar brand-concepts (e.g., sport brand) will not be significantly different in the multiple sponsorships versus the no sponsorship condition (not supported). Figure 5b. Implicit Event Image Transfer 7383.351.459.701020304050607080SportNo Sport% of Correctly Recalled Inconsistent Paired-Associations No Sponsorship Multiple Sponsorships H2a predicted that implicit image transfer would take place from the event to the sponsors with similar brand-concepts and H2b predicted that implicit image transfer 77
78 would not occur from the event to the sponsor s with dissimilar bra nd-concepts. Figure 5b shows that the recall of inc onsistent paired associations was not greater in the multiple sponsorships condition compared to the no sponsorship condition for sponsors with similar brand-concepts; however, it was greater in the multiple sponsorships condition for sponsors with dissimilar brand-concepts. Although an EIT effect was found, it was significant for the no sport brands but not for the sport brands. Hypotheses H1a to H2b were tested using an ANCOVA (see Table 8-1 in Appendix 8) followed by a-priori contrasts of specific ce ll means (see Table 9-1 in Appendix 9). The recall data for inconsistent paired-associations (BIT and EIT trials) were submitted to a 2 (sponsorship: multiple sponsorships vs. no sponsorship) x 2 (image: brand related vs. even t related) x 2 (brand-concept: similar vs. dissimilar) mixed measures ANCOVA with repeated measure on the third factor. The covariate Need for Closure was included. As expected, the s ponsorship by brand-conc ept interaction was significant ( F (1,146) = 3.03, p < .09). However, a 3-way sponsorship by brand-concept by image interaction was also significant ( F (1, 146) = 7.71, p < .01). The interaction between sponsorship and brand-concept was different depending on whether the image traits of the inconsistent paired-associati ons were related to the other concomitant sponsoring brands or to the event. Consistent with H1a and H1b, planned c ontrasts showed that the recall of BIT trials was significantly greater in the multip le sponsorships condition than in the no sponsorship condition for sponsors with similar brand-concepts (M multiple = 60.98 vs. M no = 47.56, t 80 = 1.70, p < .05), but it was not significan tly different across sponsorship conditions for sponsors with dissimilar brand-concepts ( M multiple = 63.42 vs. M no = 67.07,
79 t 80 = -0.46, ns). As suggested by the 3-way inte raction, this pattern of results differed for EIT trials. The recall of EIT trials for s ponsors with similar brand-concepts was not significantly different across multiple and no sponsorship ( M multiple = 59.72 vs. M no = 51.35, t 71 = 1.00, ns), but it was significantly greater in the multiple sponsorships condition for sponsors with dissimilar brand-concepts ( M multiple = 83.33 vs. M no = 72.97, t 71 = 1.43, p < .08). Therefore, H2a and H2b were not supported. Overall, these results indicated the existence of both implicit Brand Image Transfer and implicit Event Image Transfer due to multiple sponsorships. As expected, the BIT effect held true only for the sports brands. Howeve r, the EIT effect held true only for the no sport brands (see Figure 5a and 5b). The EIT effect for the sport brand was in the hypothesized direction although not significant. It could be that two entitative groups were formed with image transfer occurring with in the boundaries of each group: one group composed of the s port brands uniquely and another group composed of the event and the no sport brands. This point will be developed further in the discussion section. Additional Analyses The 3-way sponsorship by brand-concept by image interaction was qualified by a significant 4-way sponsorship by brand-c oncept by image by Need for Closure interaction ( F (1, 146) = 9.02, p < .01). Need for Closure allowed further testing of the entitativity model of image tran sfer developed earlier. If transfer of image effects in multiple sponsorships are caused by stereo typic processing, individuals with a high Need for Closure should ha ve intensified Brand and Event Image Transfer effects
because they more strongly rely on stereotypic processing as compared to respondents with a low Need for Closure. Results indicate that these post-hoc predictions are warranted concerning both Brand Image Transfer (Figure 6a and 6b) and Event Image Transfer (Figure 7a and 7b). Brand Image Transfer for: Figure 6a. Individuals with a Low Need for Closure 61.952.7857.1461.110102030405060SportNo Sport% of Correctly Recalled Inconsistent Paired-Associations No Sponsorship Multiple Sponsorships Figure 6b. Individuals with a High Need for Closure 7538.8969.0557.14010203040506070SportNo Sport% of Correctly Recalled Inconsistent Paired-Associations No Sponsorship Multiple Sponsorships 80
Event Image Transfer for: Figure 7a. Individuals with a Low Need for Closure 43.338077.2872.7201020304050607080SportNo Sport% of Correctly Rrecalled Inconsistent PairedAssociations No Sponsorship Multiple Sponsorships Figure 7b. Individuals with a High Need for Closure 92.8656.8268.1839.290102030405060708090SportNo Sport% of Correctly Recalled Inconsistent Paired-Associations No Sponsorship Multiple Sponsorships 81
82 Brand image transfer analysis (see Table 9-2 in Appendix 9). Respondents low in Need for Closure did not have a significantly better recall of BIT trials in the multiple sponsorships condition as compared to the no sponsorship condition for both the sport brands ( M multiple = 61.11 vs. M no = 57.14, t 37 = 0.34, ns) and the no sport brands ( M multiple = 52.78 vs. M no = 61.91, t 37 = -0.84, ns). For respondents high in Need for Closure, however, the recall of BIT trials fo r the sport brands was significantly greater in the multiple sponsorships condition than in the no sponsorship condition ( M multiple = 57.14 vs. M no = 38.89, t 37 = 1.64, p = .05). Whereas, for the no sport brands, the recall of BIT trials was not influenced by multiple sponsorships ( M multiple = 69.05 vs. M no = 75.00, t 37 = -0.49, ns). When Need for Closure was low there was no BIT effect either for the no sport brands or for the sport brands. The BIT effect was observed for the sport brands, but was not observed for the no s port brands when Need for Closure was high similarly to what was found at an aggreg ate level. This s upports the notion that Brand Image Transfer is conti ngent on stereotypic processing. Event image transfer analysis (s ee Table 9-2 in Appendix 9). As shown earlier, event image transfer took place for the no spor t brands but not for the sport brands. Therefore, if this is due to stereotypic pro cessing, one would expect event image transfer for the no sport brands to depend on N eed for Closure. For respondents low in Need for Closure, the EIT trials on the no sport brands were not affected by multiple sponsorships ( M multiple = 80.00 vs. M no = 77.28, t 35 = 0.29, ns); whereas, for respondents high in Need for Closure, recall of EIT trials on the no sport brands was significantly
83 greater in the case of multiple sponsorships ( M multiple = 92.86 vs. M no = 68.18, t 35 = 2.19, p < .02). Similarly to BIT, stereotyping seems to be the process responsible for Event Image Transfer effects. Note that the BI T effect took place am ong the sport brands, whereas the EIT effects took place from the ev ent to the no sport brands. For the sport brands, however, respondents low in Need for Closure had a better recall of EIT trials in the multiple sponsorships condi tion than in the no sponsorship condition ( M multipl e = 72.73 vs. M no = 43.33, t 35 = 2.67, p < .01). Respondents high in Need for Closure had a significantly lower recall of EIT trials in the multiple sponsorships condition than in the no sponsorship condition ( M multiple = 39.29 vs. M no = 56.82, t 35 = 1.45, p < .08). Since EIT occurred for the no sport brands, the abstracted group impression of the sport brands does not in clude the event. Based on these premises, one could expect no significant differences across sponsorship conditions for the sport brands on EIT trials; however, this was not the case. Differences in the ease with which low and high Need for Closure respondents could learn the inconsistent pairedassociations due to their processing styles could explain these results (this will be further addressed in the discussion section). Processes Hypotheses: Brand Image Reinforcement (BIR) H3a: The recall of Brand Image Reinforcement (BIR) trials for sponsors with dissimilar brand-concepts (e.g., no sport brand) will not be significantly different in the multiple versus the no sponsorship condition ( failed to reject ).
H3b: The recall of Brand Image Reinforcement (BIR) trials for sponsors with similar brand-concepts (e.g., sport brand) will be lower in the multiple sponsorships versus the no sponsorship condition (not supported). H4: In the multiple sponsorships condition, the recall of Brand Image Reinforcement (BIR) trials will be greater for sponsors with dissimilar brand-concepts (e.g., no sport brand) than for sponsors with similar brand-concepts (e.g., sport brand) (supported). Figure 8. Brand Image Reinforcement 60.2673.7279.8765.5801020304050607080SportNo Sport% of Correctly Recalled ConsistentPaired-Associations No sponsorship Multiple Sponsorship Hypothesis 3a predicted that performance on BIR trials for sponsors with dissimilar brand-concepts would not be significantly different across sponsorship conditions while H3b predicted that BIR would be smaller in the case of multiple sponsorships for sponsors with similar brand-concepts. Hypothesis 4 predicted that, in the case of multiple sponsorships, the recall of BIR trials would be greater for sponsors 84
85 with dissimilar brand-concepts. The results show that higher entitativity tend to impede the memorization of brand-consiste nt information (see Figure 8). These hypotheses were tested with an ANOVA (Table 8-2 in Appendix 8) followed by contrast tests (Table 9-3 in A ppendix 9). The recall data for consistent paired-associations (BIR trials) were submitted to a 2 (sponsorship: multiple vs. no) x 2 (brand-concept: similar vs. dissimilar) mixe d measures ANOVA with repeated measure on the second factor. 3 Although not hypothesized, the main effect of brand-concept was significant: the recall of BIR tria ls was greater for sponsors with dissimilar versus similar brand-concepts ( M dissimilar = 76.77 vs. M similar = 62.91, F (1,153) = 17.05, p < .001). This was consistent with our conceptual view becau se BIR trials should be easier to remember for brands forming a lower entitativity group. Contrary to what was expected, the bra nd-concept by sponsorship interaction was not significant (F (1,153) = 0.2, ns). However, the recall of BIR trials, as predicted, was not significantly different across the sponsor ship conditions for sponsors with dissimilar brand-concepts, ( M multiple = 79.87 vs. M no = 73.72, t 153 = 1.19, ns); therefore, H3a could not be rejected. For sponsors with similar br and-concepts, the recall of BIR trials was not significantly different across s ponsorship conditions either ( M multiple = 65.58 vs. M no = 60.26, t 153 = 0.90, ns), which did not support H3b. In addition, as predicted by H4, the recall of BIR trials in the multiple sponsor ships condition was greater for sponsors with dissimilar brand-concepts than with similar brand-concepts ( M dissimilar = 79.87 vs. M similar = 65.58, t 76 = 3.04, p < .01). This indicated that entita tivity affected how strongly each brand was associated with its image. Overall, this analysis supports the notion that higher 3 The image factor (brand related vs. event related) on ly applies for inconsistent paired-associations. BIR trials are based on consistent paired-associations; therefore, the image factor was not used in this analysis.
86 entitativity weakens the link between the brands and their own image in respondents minds. Process Hypotheses: BIR vs. BIT and EIT Trials for Sponsors with Similar BrandConcepts H5a: In the case of no sponsorship, the recall of Brand Image Reinforcement (BIR) trials will be greater than the recall of Brand Image Transfer (BIT) trials for sponsors with similar brand-concepts (e.g., sport brand) ( supported ). H5b: In the case of multiple sponsorships, the recall of Brand Image Reinforcement (BIR) trials will not be significantly different from the recall of Brand Image Transfer (BIT) trials for sponsors with similar brand-concepts (e.g., sport brand) ( failed to reject ). H5c: In the case of no sponsorship, the recall of Brand Image Reinforcement (BIR) trials compared to Event Image Transfer (EIT) trials will be greater for sponsors with simila r brand-concepts (e.g., sport brand) ( not supported ). H5d: In the case of multiple sponsorships, the recall of Brand Image Reinforcement (BIR) trials compared to Event Image Transfer (EIT) trials
will not be significantly different for sponsors with similar brand-concepts (e.g., sport brand) (failed to reject). These hypotheses posit that the levels of image transfer (either for BIT or EIT) and image reinforcement will not be different when entitativity is high, but that the level of image transfer will be greater when entitativity is low. Figure 9a shows that multiple sponsorships facilitated the memorization of BIT trials to a greater extent than the memorization of BIR trials as compared to no sponsorship, while Figure 9b shows the same phenomena for EIT trials. Consistent vs. Inconsistent Paired-Associations for Brands with Similar Concepts: Figure 9a. BIR vs. BIT 62.2706047.56010203040506070No SponsoMulti Sponso% of Correctly Recalled Paired-Associations BIR BIT 87
Figure 9b. BIR vs. EIT 61.1158.1159.7251.350102030405060No SponsoMulti Sponso% of Correclty recalled Paired-Associations BIR EIT These hypotheses were tested through an ANCOVA (see Table 8-3 in Appendix 8) followed by contrast tests (see Table 9-4 in Appendix 9). The percentage of correctly recalled paired-associations for sponsors with similar brand-concepts was submitted to a 2 (sponsorship: multiple vs. no) x 2 (image: brand related vs. event related) x 2 (paired-associations: consistent vs. inconsistent) mixed measures ANCOVA with repeated measures on the last factor; Need for Closure was included as a covariate. As expected, the paired-association by sponsorship interaction was significant (F(1,146) = 9.61, p < .01), which indicated that the recall differential between consistent and inconsistent paired-associations varies with sponsorship conditions. A-priori contrasts revealed that the recall of BIR trials versus BIT trials was greater in the case of no sponsorship (M BIR = 62.20 vs. M BIT = 47.56, t 40 = 1.82, p < .05), but was not significantly different in the case of multiple sponsorships (M BIR = 69.51 vs. M BIT = 60.98, t 40 = 1.1, ns), which was in line with H5a and H5b. The recall of BIR trials versus EIT trials, in the case of no sponsorship, was in a direction consistent with H5c, 88
89 but it did not reach statistical significance (M BIR = 58.11 vs. M EIT = 51.35, t 35 = 0.87, ns). In accordance with H5d, the recall of BIR versus EIT trials was not significantly different in the case of multiple sponsorships (M BIR = 61.11 vs. M EIT = 59.72, t 35 = 0.17, ns). For sponsors with similar brand-concepts, this anal ysis confirms that in the case of multiple sponsorships consistent paired -associations were as easy to remember as inconsistent paired-associations. In the case of no sponsor ship, they were easier to remember than inconsistent paired-associ ations. In addition, the 3-way paired-association by sponsorship by image interaction was significant (F (1,146) = 2.94, p < .09). This shows that the differential impact on the memorizati on of consistent vs. inconsistent pairedassociations was greater for EIT trials than for BIT trials. Additional Analysis (see Table 9-5 in Appendix 9) The 3-way paired-associations by sponsorsh ip by image interaction was qualified by a 4-way paired association by sponsorship by image by Need for Closure interaction ( F (1,146) = 2.93, p < .09). Contrast tests revealed that, in the multiple sponsorships condition, respondents high in Need for Closure exhibited a greater recall on BIR trials than BIT or EIT trials (M BIR = 78.57 vs. M BIT = 57.14, t 20 = 2.10, p < .03 and M BIR = 60.71 vs. M EIT = 39.29, t 13 = 1.39, p < .09) whereas, respondents low in Need for Closure did not have a significan t greater recall of BIR tria ls compared to BIT or EIT trials (M BIR = 61.11 vs. M BIT = 61.11, t 17 = 0.00, ns and M BIR = 61.36 vs. M EIT = 72.73, t 21 = 1.31, ns). This indicates that low Need for Closure respondents exhibited effects consistent with our hypotheses (see Figure 10a and 10b). Respondents high in Need for Closure exhibited a reversed pattern of results (see Figure 11a and 11b). Although it
was shown earlier that high Need for Closure respondents had a stronger stereotypic processing in multiple sponsorship conditions on BIT trials, this analysis revealed that their memorization performance on the BIR trials was higher than on the BIT trials in the case of multiple sponsorships. BIR vs. BIT for Sponsors with Similar Brand-Concepts: Figure 10a. Low Need for Closure 90 61.9161.1157.1461.110102030405060NoMulti% of Correctly Recalled Paired-Association s BIR BIT
Figure 10b. High Need for Closure 57.1463.8978.5732.3401020304050607080NoMulti% of Correclty Recalled Paired-Associations BIR BIT BIR vs. EIT for Sponsors with Similar Brand-Concepts: Figure 11a. Low Need for Closure 43.3372.7361.3770010203040506070NoMulti% of Correctly recalled Paired-Associations BIR EIT 91
Figure 11b. High Need for Closure 60.715056.8239.290102030405060NoMulti% of Correctly recalled Paired-Associations BIR EIT Process Hypotheses: BIR vs. BIT and EIT Trials for Sponsors with Dissimilar Brand-Concepts H6a: In the case of no sponsorship, the recall of Brand Image Reinforcement (BIR) trials compared to Brand Image Transfer (BIT) trials will be greater for sponsors with dissimilar brand-concepts (e.g., no sport brand) (supported). H6b: In the case of multiple sponsorships, the recall of Brand Image Reinforcement (BIR) trials compared to Brand Image Transfer (BIT) trials will be greater for sponsors with dissimilar brand-concepts (e.g., no sport brand) (supported). H6c: In the case of no sponsorship, the recall of Brand Image Reinforcement (BIR) trials compared to Event Image Transfer (EIT) trials 92
93 will be greater for sponsors with dissimilar brand-concepts (e.g., no sport brand) ( not supported ). H6d: In the case of multiple sponsorships, the recall of Brand Image Reinforcement (BIR) trials compared to Event Image Transfer (EIT) trials will be greater for sponsors with dissimilar brand-concepts (e.g., no sport brand) ( not supported ). These hypotheses posit that the level of im age reinforcement (B IR) will be greater than the level of image transfer (BIT and EIT) when entitativity is low (i.e., due to either no sponsorship or dissimilar brand-concepts). Consistent with these predictions, Figure 12a shows that consistent paired-associations were easier to memorize than inconsistent ones independently from the multiple sponsorships manipulation. Figure 12b reveals that this prediction did not hold true for EIT trials.
Consistent versus Inconsistent Paired-Associations for Brands with Dissimilar Concepts: Figure 12a. BIR vs. BIT 8087.8167.0762.50102030405060708090No SponsoMulti Sponso% of Correclty Recalled Paired-Associations BIR BIT Figure 12b. BIR vs. EIT 79.1758.1183.3372.970102030405060708090No SponsoMulti Sponso% of Corectly Recalled Paired-Associations BIR EIT These hypotheses were tested through an ANCOVA (see Table 8-4 in Appendix 8) followed by contrast tests (see Table 9-6 in Appendix 9). The percentage of correctly recalled paired-associations for sponsors with dissimilar brand-concepts was submitted to 94
95 a 2 (sponsorship: multiple vs. no) x 2 (image: brand related vs. event related) x 2 (pairedassociations: consistent vs. inconsistent ) mixed measures ANCOVA with repeated measures on the third factor and Need for Clos ure as a covariate. Contrary to what was expected, the paired-association by sponsor ship interaction wa s not significant ( F (1,146) = 0.03, ns). However, the 3-way paired-asso ciation by sponsorship by image interaction was significant (F (1,146) = 4.05, p < .05). For sponsors with dissimilar brand-concepts, the impact of multiple sponsorships on the differential recall between consistent and inconsistent paired-associati ons differed depending on whether the inconsistent pairedassociations were BIT or EIT trials. A-priori contrasts re vealed that the recall of BIR trials versus BIT trials was grea ter in the case of no sponsorship ( M BIR = 87.81 vs. M BIT = 67.07, t 40 = 3.76, p < .001) as well as in the case of multiple sponsorships ( M BIR = 80.49 vs. M BIT = 63.41, t 40 = 3.33, p = .001) in support of H6a and H6b. However, contrary to H6c and H6d, the recall of BIR trials versus EIT trials wa s lower in the case of no sponsorship ( M BIR = 58.11 vs. M EIT = 72.97, t 35 = 2.06, p < .05) and not significantly different in the case of multiple sponsorships ( M BIR = 79.17 vs. M EIT = 83.33, t 35 = 0.72, ns). In line with the absence of BIT e ffect found for sponsors with dissimilar brandconcepts, it appears that respondents failed to group the no sport brands together. Therefore, they did not engage in stereotypi c processing, which did no t impair their recall of consistent paired-associati ons. As revealed by the 3-way interaction, however, this pattern did not hold for event image transf er. For EIT, stereotypic processing was evident in the multiple sponsorships condition. This is consistent with the results from the outcome hypotheses (i.e., event image tran sfer occurred from the event to the no
96 sport brands) and shows that high entitativ ity rendered difficult th e association between the no sport brands and their own images. Additional Analysis (see T able 9-7 in Appendix 9) The 3-way paired-associations by sponsorsh ip by image interaction was qualified by a 4-way paired association by sponsorship by image by Need for Closure interaction ( F (1,146) = 3.67, p < .06). In line with the theori zation that sponsors with dissimilar brand-concepts would not be subjected to stereotypic processing, high Need for Closure did not increase brand image transfer for the no sport brands (see Figure 13a and 13b). Respondents either low or high on th at variable exhibited a greater recall for BIR than BIT trials when exposed to the multiple sponsorships treatment ( M BIR = 72.22 vs. M BIT = 52.78, t 17 = 2.72, p < .01 and M BIR = 85.71 vs. M BIT = 69.05, t 20 = 2.09, p < .05, respectively). In the case of multiple sponsorships, respondents low in Need for Closure did not exhibit a significant difference betw een BIR and EIT trials ( M BIR = 86.36 vs. M EIT = 77.37, t 21 = 1.31, ns) while respondents high in Need for Closure had a lower recall of BIR trials than EIT trials ( M BIR = 67.86 vs. M EIT = 92.86, t 13 = -2.88, p < .01). This EIT analysis showed that Need for Closure had the same moderating effect on the no sport brands as on the sport brands previously investigated. For multiple sponsorships, the performance of high N eed for Closure respondents on consistent paired-associations was the same as on inc onsistent paired-associations, whereas the performance of low Need for Closure res pondents was greater on inconsistent pairedassociations (see Figure 14a and 14b).
BIR vs. BIT for Sponsors with Dissimilar Brand-Concepts: Figure 13a. Low Need for Closure 52.7872.2290.4861.910102030405060708090NoMulti% of Correctly Recalled Paired-Associations BIR BIT Figure 13b. High Need for Closure 85.7183.337569.050102030405060708090NoMulti% of Correclty Recalled Paired-Associations BIR BIT 97
BIR vs. EIT for Sponsors with Dissimilar Brand-Concepts: Figure 14a. Low Need for Closure 86.3663.338077.270102030405060708090NoMulti% of Correclty Recalled Paired-Associations BIR EIT Figure 14b. High Need for Closure 54.5567.8668.1892.860102030405060708090NoMulti% of Correclty Recalled Paired-Associations BIR EIT 98
Process Hypotheses: Tag-Line Recognition Data H7a: Tag-line recognition for sponsors with similar brand-concepts (e.g., sport brand) will not be significantly different in the no sponsorship versus the multiple sponsorships condition (failed to reject). H7b: Tag-line recognition for sponsors with dissimilar brand-concepts (e.g., no sport brand) will be greater in the no sponsorship versus the multiple sponsorships condition (supported). These hypotheses state that the strength of the association between the tag-lines and their respective sponsors will be different across multiple sponsorships and no sponsorship conditions when entitativity is low (i.e., only for the no sport brands). Figure 15 shows that the percentage of recognition is systematically greater for no sponsorship than for multiple sponsorships. However, the difference proved statistically significant only for sponsors with dissimilar brand-concepts in line with H7a and H7b. Figure 15. Tag-Line Recognition 74.166.3238.7833.77010203040506070SportNo Sport% of Correctly Recognized Tag-Lines No Sponsorship Multiple Sponsorships 99
100 These hypotheses were tested through an ANOVA (see Table 8-5 in Appendix 8) followed by contrast tests (see Table 9-8 in Appendix 9). The per centage of correctly recognized tag-lines was submitted to a 2 (sponsorship: multiple vs. no) x 2 (brandconcept: similar vs. dissimilar) mixed measures ANCOVA with repeated measures on the second factor and Need for Closure as a c ovariate. The sponsorship by brand-concept interaction was not significant ( F (1,150) = 0.15, ns), however, the cell means exhibited a pattern of results consiste nt with H7a and H7b. For sponsors with similar brandconcepts, tag-lines recognition was not signi ficantly impacted by multiple sponsorships ( M multiple = 33.77 vs. M no = 38.78, t 153 = -1.20, ns) while tag-lines recognition was greater in the no sponsorship than in the multiple sponsorships condition for sponsors with dissimilar brand-concepts ( M multiple = 66.32 vs. M no = 74.10, t 153 = -2.17, p < .05). Taglines were less uniquely associated with their brands in the case of multiple sponsorships especially for the no sport brands. For the sport br ands, although the difference was not significant, it also sugge sts that lower recognition is a function of entitativity. Additional Analysis The results presented were qualified by tw o 2-way interactions in which Need for Closure was involved. The interaction between sponsorship and Need for Closure was significant (F(1,150) = 3.10, p = .08) as well as the interaction between brandconcept and Need for Closure (F(1,150) = 5.26, p < .03). The impact of multiple sponsorships on tag-line recognition was gr eater for respondents high in Need for Closure than for respondents low in Need for Closure. Contrasts tests (see Table 9-9 in Appendix 9) showed that for low Need for Closure, recognition of the sport
101 brands tag-lines in the no sponsorship condi tion was not significantl y different from the multiple sponsorships condition ( M multiple = 37.50 vs. M no = 40.28, t 74 = -0.46, ns) while recognition of the no sport brands taglines was greater in the no sponsorship condition than in the multiple sponsorships condition ( M multiple = 65.18 vs. M no = 72.92, t 74 = -1.44, p < .08). This pattern of results was the same across the sport brands and the no sport brands for high Need for Closure respondents ( M multiple = 30.00 vs. M no = 37.5, t 73 = -1.23, ns and M multiple = 65.71 vs. M no = 75.75, t 73 = -2.06, p < .05, respectively). The significant interaction between spons orship and Need for Closure indicates that the observed decrease in tag-line rec ognition in the multiple sponsorships condition compared to the no sponsorship condition is greater for respondents high in Need for Closure (see Figure 16a and 16b). In addition, the significant interaction between brand-concept and Need for Closure indicat es that, independently of sponsorship condition, the level of tag-lines recognition fo r sport brands is weaker for respondents with a high Need for Closure compared to th ose with a low Need for Closure. Need for Closure does not impact recognition of the no sport brands tag-lines (see Figure 16a and 16b). This is consistent with th e theoretical view th at a higher Need for Closure should lead to intensifie d effects. In this analysis greater entitativity led to a lower level of tag-lines recognition. For high versus low Need for Closure respondents, this phenomenon was larger for th e sport brands vers us the no sport brands (i.e., high vs. low entitativity) as well as for multiple sponsorships versus no sponsorship (i.e., high vs. low entitativity).
Tag-Line Recognition By Level of Need for Closure: Figure 16a. Low Need for Closure 72.9265.1740.2837.501020304050607080SportNo Sport% of Correctly Recognized Tag-Lines No sponsorship Multiple Sponsorships Figure 16b. High Need for Closure 75.7565.7137.53001020304050607080SportNo Sport% of Correctly Recognized Tag-Lines No sponsorship Multiple Sponsorships Mediation Analysis Although each hypothesis tested was derived from the entitativity model of image transfer, the models mechanism has not been tested in and of itself. This model implies that a group impression is abstracted and then influences the perception of each brand. Accordingly, the transfer of image does not take place from one brand to another or from 102
103 the event to the sponsoring brand directly bu t from that group impression to each brand (see Figure 2). As a result, in a multiple s ponsorships situation, an observer first abstracts an entitative group and then images are transferred from that group to each brand. This is a view different from the brand leverage model (Keller 2003), according to which entities are directly associated with each other (e.g., brands, event). The entitativity model of image transfer would be supported against th e brand leverage model if the perceived entitativity of the group mediated the impact of multiple sponsorships on image transfer (recall of inconsistent paired associations; BIT or EIT). This w ould demonstrate that image transfer does not take place dir ectly between entities, but through a group impression. The type of mistakes on the cued-recall ta sk was used to operationalize the impact of entitativity on image transfer. If a group of brands and event was formed and perceived as being entitative, mistakes co mmitted on the cued-recall task should not be constituted of image-traits th at did not characterize the spons oring brands or the event. During the memorization task, re spondents were exposed to a set of fourteen brands, including the eight target bra nds and the six foil brands. The latter were paired with either the word Competent or Rugged, wh ich should not have been associated with the abstracted group in the case of multiple sponsorships and, therefore, should not have been associated with the sponsor ing brands (i.e., the target bra nds). On the other hand, in the case of no sponsorship, since no group impr ession should have been abstracted, the words Competent or Rugged were more likely to constitute recall errors when respondents could not retrieve the word they memorized with the target brands when presented with the paired-associations (i.e., sincere and exciting). As a result, if a
104 group impression was abstracted due to percei ved entitativity, wrongl y recalling that a target brand was paired with the words Competent or Rugged should mediate the relationship between multiple spons orships and image transfer. Mistakes associated with the words C ompetent or Rugged constituted the variable out-group mistake (e.g., a target brand is recalled as being associated with the word Competent). Image tran sfer was operationalized as the percentage of correctly recalled inconsistent paired-associations at an aggregate level. The causal step procedure of mediation analysis from Baron and Kenny (1986) was followed (see Table 2) in which the independent variable was multiple sponsorships, the dependent variable was image transfer, and the mediating variable was out-g roup mistakes. A first regression analysis revealed a significant negative impact of multiple sponsorships on out-group mistakes ( = -.41, t = 1.71, p < .1). As expected, in the case of multiple sponsorships, high entitativity rendered out-group mistakes le ss likely due to stronger associations among group members and group level information (i .e., sincere and ex citing). A second regression analysis showed that multiple s ponsorships had a significant positive impact on image transfer (i.e., = 9.53, t = 2.17, p < .05). When image transfer was regressed on both multiple sponsorships and out-group mistakes, the beta coefficient associated with multiple sponsorships became insignificant (i.e., = 5.01, t = 1.41, ns). This supports the notion that a group impression fully mediated the impact of multiple sponsorships on image transfer and is consiste nt with the entitati vity model of image transfer as opposed to th e brand leverage model.
Table 2. Mediation of the Impact of Multiple Sponsorships on Image Transfer by a Group Impression First Step Second Step Third Step Dependent Variable Out-group mistake (y) Image Transfer (y) Image Transfer (y) Independent Variable Multiple sponsorships (x 1 ) Multiple sponsorships (x 1 ) Multiple sponsorships (x 1 ) Out-group mistake (x 2 ) Model y = b 1 x 1 + y = b 1 x 1 + y = b 1 x 1 + b 2 x 2 + Parameters 1 = -0.41, SD = 0.24, t = -1.71, p = .09. 1 = 9.53, SD = 4.31, t = 2.17, p = .032. 1 = 5.01, SD = 3.55, t = 1.41, p = .16. 2 = -10.90, SD = 1.17, t = -9.31, p < .001. Strength of BIT vs. Strength of EIT An important question remains to determine which of the concomitant sponsoring brands or the event has a stronger impact on image transfer. In other words: is the information associated with the abstracted group impression more strongly influenced by the sponsoring brands or the event? In order to address this question, it is necessary to compare the effect size of BIT trials and EIT trials across sponsorship conditions. The results obtained earlier showed that a Brand Image Transfer phenomenon occurred for the sport brands while an Event Image Transfer phenomenon occurred for the no sport brands. Therefore, the effect size relative to the difference between multiple sponsorships and no sponsorship on the recall of inconsistent paired-associations of the sport brands for BIT will be compared with the effect size relative to the differences between multiple sponsorships and no sponsorship on the recall of inconsistent paired associations of the no sport brands for EIT. According to meta-analytic thinking, two effect sizes can be deemed different if their confidence intervals do not overlap (Hunter and Schmidt 2004; Kline 2004). BIT 105
and EIT effect sizes result from the comparison of two groups (i.e., multiple sponsorships versus no sponsorship) on the mean of a continuous variable (cued-recall of inconsistent paired-associations). Therefore, Hedges g, which is an estimator of the population parameter is appropriate for this purpose. Kline (2004) proposed a method to compute exact confidence intervals for based on Hedges g when comparing means from two independent samples. These exact confidence intervals are based on the assumption that if the null hypothesis (i.e., the two groups are equal) tested is false, one should not rely on a central t distribution since it is only appropriate when the null hypothesis is true. Rather, a noncentral t distribution needs to be used when the null hypothesis is false. The noncentral t distribution has two parameters: the degree of freedom and a noncentrality parameter, which indicates to what extent the null hypothesis is false. Since we found significant effects in previous analyses for BIT and EIT, we can assume that the null hypotheses are false. In order to rely on a noncentral t distribution, the TNONCT function in SAS was used in order to generate exact confidence intervals for BIT and EIT. The TNONCT function computes a noncentrality parameter, which is then transformed to obtain the exact confidence interval for the population effect sizes The estimator Hedges g of the population for BIT was .38; EIT was slightly lower with .33. 4 However, the computation of exact C.I for BIT showed that the observed effect size g = .38 was as consistent with a population effect size as low as = .0073 as it was with one as high as = .75 with 90% confidence. The computation of exact C.I for EIT showed that the observed effect size of .33 was as consistent with a population effect size 106 4 Hedges g = 2111nnt
107 as low as = -.05 as it was with one as high as = .68 with 90 % confidence. In spite of the observed differences between the Hedges g for BIT versus EIT (.38 vs. .33), the large overlap of the exact confiden ce intervals of their corresponding leads to an inconclusive assessment of which effect size is stronger. At this point, all cases are possible. BIT could be stronger or EIT could be stronger or they could be of even strength. Further research a nd replications studies would be needed in order to reduce sampling error (Hunter and Schmidt 2004) a nd narrow the exact confidence intervals around in order to firmly establish how BIT and EIT compare. This chapter brought empirical evidence of a Brand Image Transfer phenomenon taking place among the sponsors with a similar brand-concept as well as of a phenomenon of Event Image Transfer taking pl ace from the event to the sponsors with dissimilar brand-concepts. Further analyses supported the notion that category-based processing triggered by the entitativity of multiple sponsorships situations was responsible for these image tran sfer phenomena. The next se ction will discuss further the theoretical implications of thes e findings as well as the mana gerial recommendations that can be derived from them. Limitations and guidelines for further research will be provided as well.
108 CHAPTER 5 Discussion and Conclusion As a medium of marketing communi cation, sponsorship is going through a tremendous phase of growth encompassing firms marketing activities as well as scholarly research. Although the early approa ches to sponsorships by practitioners and academics were mostly speculative and descriptive, today both management and research conceive sponsorships in a more sophisticat ed manner. Kellers (2003) brand leverage model provides a good basis for understandi ng sponsorships effects. This model describes sponsorships as marketing/promotion tools and conceptualizes a direct transfer of knowledge from the event spons ored to the sponsoring brand. This dissertation focused on a very important factor that is not addressed either by sponsorship research in general or by the bra nd leverage model in pa rticular: most of the sponsorship agreements involve multiple bra nds with multiple possible sources of brand associations. In addition, many practitioners and scholars still question the positive effect of sponsorships and sometimes point out the mixed support that sponsorships have received in terms of the beneficial impact on the sponsoring brand. Some researchers have suggested that the best way to fully e xplore sponsorship is to measure its influence on consumers implicit memory (Johar and Pham 1999; Pham and Vanhuele 1997). To the authors knowledge, however, all of the studies conducted so far have relied on explicit measures of attitudes, awarene ss, intention, preferences, or evaluations.
109 In view of the previous points, this disse rtation contributes to the expansion of knowledge regarding sponsorship effects in three ways: (1) by conceptualizing and empirically testing an entitativity model of image transfer, (2) by unraveling the psychological processes underpinning imag e transfer, and (3) by measuring image transfer effects at the implicit level rather th an at the explicit level. This chapter is organized as follows. First, a detailed discus sion of the theoretical implications of the results is provided. Then, the practical imp lications of the main findings are developed. Finally, areas of further research, as well as limitations, are discussed before a concluding statement. Discussion of Experimental Results Outcome Hypotheses This dissertation addresses several questions that are critical for evaluating the impact of multiple sponsorships on consumers perceptions of sponsoring brands: is there evidence of Brand Image Transfer in situations of multiple sponsorships? Is there is also evidence of Event Image Transf er? Are these imag e transfer phenomena occurring at the implicit level? Do they have boundary conditions? Evidence of brand image and event image transfer. This dissertation demonstrates the existence of a Brand Image Tr ansfer (BIT) effect as well as an Event Image Transfer (EIT) effect due to multiple sponsorships. In other words, meaning associated with one sponsoring brand becomes associated with the other sponsors due to multiple sponsorship agreements. In the same vein, meaning associated with the event
110 becomes associated with the sponsors. Due to the particularities of the paradigm used (i.e., savings in relearning), it was shown that both BIT and EIT occurred in an implicit manner. Consistent with the definition of implicit memory, the recall of inconsistent paired-associations was improved by multiple sponsorships (Roediger 1990). This demonstrated that respondents had already implicitly associated the brand with the images of the other concomitant sponsors at the time they were learning these inconsistent paired-associations, which generated th e savings. The impact of brandconcept similarity Importantly, BIT as well as EIT effects were contingent on the spons ors brand-concept. As expected, brand image was transferred among the sponsors that shared the same brand-concept as the event. Although event image transfer was also expected to take place for the sponsors with a sport brand-concept, the im age of the event was transfer red to the sponsors with a brand-concept different from the one of the ev ent. Subjects associated the images of the sport brands together but di d not do so for the no sport brands. Rather, the no sport brands were associated with the imag e of the event. A better understanding of the cognitive process underlying image transfer (B IT or EIT) will serve as the basis for discussing this unexpected result concerning EIT. The individual difference variable Need for Closure is utilized to discuss this process. The role of individual differences An investigation of the psychological process underlying BIT and EIT confirmed that the obse rved effects were due to category-based processing (Fiske and Neuberg 1990). Re spondents high in Need for Closure, a variable that assesses the propensity of an individual to process information stereotypically (i.e., to rely on category-based pr ocessing), showed image transfer effects.
111 Respondents low in Need for Closure did not show such effects. Specifically, low Need for Closure respondents did not exhibit a better performance on inconsistent paired-associations in the case of multiple s ponsorships, which showed that they did not associate the images of the concomitant sponsor s together (i.e., individuating processing). High Need for Closure respondents did exhib it BIT. They had a better performance on inconsistent-paired associations, revealing that they associated the images of the concomitant sponsors together due to categ ory-based processing. This supports the notion that BIT is due to stereotypic proces sing because BIT effect did not occur when that processing was suppressed but it did o ccur when that processing was intensified. Results concerning Event Image Transfer followed exactly the same pattern. Respondents with a low Need for Closure di d not show EIT while respondents with a high Need for Closure did show EIT. It is important to note that BIT occurred for the sport brands and EIT occurred for the no sport brands. Ca tegory-based processing was evident for both effects, which implies th at a group was formed in each case (Sujan 1985). It was expected that respondents w ould categorize the spor t brands together with the event and that EIT, as well as BIT, would take place for these brands while the no sport brands would not be subjected to any transfer effects. According to the results, however, respondents abstracted two di fferent groups that both were the basis of category-based processing and image transfer. The sport brands were abst racted into a group that resulted in the observed BIT effect while the no sport brands and the ev ent were abstracted in to another group that resulted in the observed EIT effect. Therefor e, the transfer of the events image to the no sport brands was also due to a high perc eived entitativity for this group of brands
112 and the event. But why were the no sport brands and the event categorized together despite the fact that these sponsors did not ha ve the same brand-concept as the event? Although respondents did not categor ize the no sport brands with the event based on a sport concept, they may have relied on a more abstract leve l of categorization in order to classify together these stimuli. According to Rosch, Gray, Johnson, and Boyes-Barem (1976), cognitive categories for stimuli range along an abstract ness continuum. Studies have shown that when objects failed to be categorized togeth er, perceivers switch to a category with a higher level of abstractness in order to succes sfully classify these objects (Johnson 1984). Classification into a basic cognitive category, the most abstract category, is an automatic process that takes place below th e conscious level (Mervis and Rosch 1981). Therefore, respondents might have relied on a basic category define d by both the notions of sport and brand, which resulted in gr ouping the event and the no sport brands together since they did not belong to that category. A lthough the event was a sporting event (i.e., 2005 Boston Golf Tournament), b ecause it was not a brand (a characteristic that might be at a higher hierarchical level for categorization purposes in that context) respondents may not have grouped it with the sport brands but with the no sport brands that could not be categ orized in to the basic category sport brands. The event and the no sport brands may have constitute d an out-group, which was entitative due to multiple sponsorships. This is consiste nt with accentuation theory (Tajfel 1959) according to which the goal of categorization is to maximize within-group similarity and to maximize between group dissimilarity (Mervis and Rosch 1981). Grouping the sport brands together in a category, as well as grouping the event and the no sport
113 brands in another category, fulfills that objec tive. This would explain why there was an image transfer among the sport brands but not from the event to the sport brands; however, there was a transfer of image from the event to the no sport brands. Other unexpected results. Given that the EIT effect was found for the no sport brands only there should have been no di fference across sponsorship conditions on the recall of EIT trials for the sport brands, independently of Need for Closure, because these brands and the event do not form a gr oup. Surprisingly, it was found that the EIT trials for the sport brands were more easily remembered in the case of multiple sponsorships for respondents w ith a low Need for Closure wh ile they were more easily remembered in the case of no sponsorship for respondents high in Need for Closure. These results could be explained by the di fferent cognitive styles of respondents that could have altered how easily they le arned the EIT trials. For high Need for Closure, because the sport brands were already a group, entitativity was even stronger in the case of multiple sponsorships, which made the association between the brands and the images sincere and exciting very strong. It could also ha ve rendered difficult learning pairings with the image sophisti cated. On the other hand, no sponsorship triggered less entitativity and the previous association with sincere and exciting were weaker, which could have fac ilitated the learning of paired-associations with the image sophisticated. Low Need for Closure resp ondents did not make strong associations between the sport brands and the images exciting and sincere due to their lack of stereotypic processing. Therefore, in th e multiple sponsorships condition, respondents could make the association between sophist icated and the brands more easily than respondents in the no sponsorships du e to a higher entitativity.
114 Process Hypotheses This dissertation contributes to the b ody of knowledge on sponsorship not only by showing evidence of implicit brand and even t image transfer in the case of multiple sponsorships, but also by unraveling the information processing m echanism causing these transfer effects. Since the ma nipulation check of entitativity supports the fact that brands in a multiple sponsorships arrangement form an entitative group with the event, one needs to focus on the information processing implications of a hi ghly entitative group. Particularly, this dissertation examines if the entitativity triggered by multiple sponsorships leads to category-based proces sing as opposed to i ndividuating processing, which causes the image transfer phenomena. Brand Image Reinforcement Brand Image Reinforcement (BIR) is a measure of how strongly a brand is associated with its own image. BIR trials represent the performance of respondents on consistent paired-associations This dissertation addressed the following question: is brand reinforcement weaker in the case of high entitativity? The impact of multiple sponsorships on BIR. High entitativity leads respondents to associate information about a group member with all the other members of the group (i.e., category-based processing). Because the no sport brands constituted a lower entitativity group than the sport brands, they were more uniquely associated with their respective images, which resulted in enhan ced respondents performance on the recall of consistent paired-associati ons (BIR). The recall of BIR trials, however, was not significantly different across sponsorship c onditions. This shows that the multiple
115 sponsorships manipulation did not diminish the recall of BIR trials, although the previous analysis showed that it facilitated the recall of inconsistent paired-associations (i.e., BIT and EIT). This indicates that multiple sponsorships triggered more category-based processing than no sponsorship whereas no sponsorship did not trigger more individuating processing than multiple sponsorships (Fiske and Neuberg 1990). It could be that the savings differentia l across sponsorship conditions was larger for inconsistent than for consistent paired-a ssociations. This is perfectly possible when one compares the posited mechanisms underlying image transfer and image reinforcement. BIT and EIT effects are based on how learning is f acilitated whereas BIR effect is based on how learning is impaired. In the case of multiple sponsorships, all the image-traits are implicitly associated w ith the target brands. Memorization of inconsistent paired-associations is then f acilitated because respondents do not have to learn these associations but to relearn them, which generate some savings. Memorization of consistent paired-associations is impaired in multiple sponsorships. Learning is made more difficult because the as sociation with all the image-traits makes the brands own image less salient. Therefor e, it could be that category-based processing due to entitativity was intense enough to trigger image transf er but not to hinder learning of consistent paired-associations. Res pondents in multiple sponsorships exhibited savings on inconsistent paired associati ons, but their learning of consistent pairedassociations was not impeded in spite of category-based pr ocessing. Therefore, it is possible that these two mechanisms have to reach different threshold levels before making a significant difference on implicit memory.
116 Comparison of BIRT Trials versus BIT and EIT Trials If higher entitativity implies category-based processing, then all the images composing the group (from the sponsors and the event) should be associated with every brand. This dissertation addr esses the following question: does high entit ativity lead inconsistent paired-associati ons to be as easy to memorize as consistent pairedassociation? Results indicated that when entitativity wa s the highest (for the sport brands in multiple sponsorships), the notion of consis tent and inconsistent paired-associations became almost irrelevant due to category-ba sed processing. For the sport brands BIR was greater than BIT and EIT in no sponsor ship, but it was not significantly different from BIT and EIT in multiple sponsorships. This result tends to indicate that th e sport brands were processed in a stereotypic way in the multiple sponsorships condition, which did not lead to significant differences between consistent and inconsistent paired-associ ations. For the no sport brands, entitativity was lower and consistent pa ired-associations were easier to remember than inconsistent paired-associations in both the multiple and no sponsorship conditions for BIT but not for EIT. As shown earlier, th e event and the no sport brands formed an entitative group and therefore, for EIT, entitativity was the highest in that condition and the recall performance on incons istent and consistent paired-associations was the same in multiple sponsorships. This impact of entita tivity on information processing is consistent with what has been found in social psychology where respondents performed similarly on paired-association that matched or did not match with the traits of members of an entitative social group (Crawford et al. 2002).
117 The impact of indi vidual differences. As before, investigating the obtained results through the Need for Closure variables support ed further the role of entitativity in information processing in multiple sponsorships. For the no sport brands, BIR recall was greater than BIT for both sponsorship conditions for both levels of Need for Closure. Respondents failed to categorize these brands into a group which, in turn, precluded entitativity and image transfer. Consistently, high Need for Closure increased the predominance of EIT trials compar ed to BIR trials for the no sport brands because that group was higher on entitativity (i .e., as shown earlier the EIT effect took place for the no sport brands). In the case of multiple sponsorships, the performance of low Need for Closure respondents on EIT trials was not different from their BIR trials performance whereas the performance of high Need for Closure respondents on EIT trials rose above their pe rformance on BIR trials. This further supported the notion that enti tativity drives st ereotypic processing since the greater the propensity to process in formation stereotypically, the greater are the consequences of entitativity. For the sport brands, however, performance on the consistent paired-associations was greater th an performance on the inconsistent paired association in multiple sponsorships for respondents with a high Need for Closure although it should have been the same (for bot h BIT and EIT). As explained before, the mechanism of BIR trial is based on the difficulty of learning new associations due to interference with associations already held in mind; it is different from the mechanism of inconsistent trials that is based on easin ess of memorization due to savings in relearning. Therefore, even though savings occu rred in the case of multiple sponsorships, this does not necessarily mean that the categ ory-based processing orig inating this effect
118 generated enough interference to significantly impede memorization on BIR trials compared to no sponsorship. Tag-Line Recognition The recognition of the correct tag-line for each sponsoring brand allowed the investigation of the impact of entitativity on another aspect of information processing. Members of entitative groups tend to lose their uniqueness as they become assimilated with the identity of the group. Therefore, this dissertation addresses the question whether or not entitativity lowers th e ability to recognize a sponsor ing brands correct tag-line, which would be an indication that the brand is not strongly associat ed with its own tagline. The tag-lines recognition results confirm that, in the case of high entitativity, sponsoring brands tend to be seen as in terchangeable. Tag-lines recognition was diminished for the sport brands and for multiple sponsorships in general. Once a group of sponsoring brands has been abstracte d, the brands lose their individuality and respondents are less able to identify brand specific info rmation. Need for Closure offers further support for this interchangeabi lity interpretation. Tag-lines recognition in low entitativity condition outperformed taglines recognition in the high entitativity condition to a greater extent for respondents w ith a high versus a low Need for Closure. This is in line with findings from attit ude research and social group perception. Lingle and Ostrom (1979) found that once a global attitude is formed, specific items that generated that attitude are of ten forgotten. Similarly, Craw ord et al. (2002) found that once a group impression is abstracted, indivi dual-level information is lost. More
119 generally, this dissertations findings show that abstra ction is accompanied by the forgetting of concrete features. It is as if people had to forg et the particular in order to conceive the general. This supports the Gestalt versus the st ructuralist view of categorization research. Gestalt psychol ogists assert that categories cannot be decomposed into its elements whereas stru cturalist psychologist s assert the opposite (Murphy and Medin 1985). The Entitativity Model of Image Tr ansfer in Multiple Sponsorships This dissertation also makes a contributi on by demonstrating th e validity of the entitativity model of image transfer in accounting for multiple sponsorships effects as compared to the brand leverage model (Kelle r 2003). The brand leve rage model seems to be most valid when only two entities ar e tied together (i.e., co-branding, single sponsorship), but not when multiple simultaneous entities are associated such as in multiple sponsorships. Not only strong support was found in favor of category-based processing as the mechanism of image transf er effects, but the formation of a group impression was also demonstrated. Res pondents wrong recall of image-traits nonassociated with the group impression mediated the impact of multiple sponsorships on image transfer. Therefore, as conceptuali zed, entitativity leads to the formation of a group impression in light of which inform ation about group members was perceived. This resulted in image transfer because indi vidual members were then perceived in light of that group impression due to category-based processing. The entitativity model of image transfer adds nicely to the traditional models of categorization-based transf er (e.g., Gregan-Paxton and Roeder John 1997; Moreau,
120 Markman and Lehmann 2001). In marketing, these models are used when knowledge from a familiar domain is transferred to an unfamiliar element (e.g., when product labels are used to transfer knowledge from the label category to a new product). The model proposed in this dissertation, however, does not depict how knowledge is transferred from a category to a new member of that ca tegory, but rather de picts how the category itself is constructed. Image transfer among me mbers of a category is a by-product of the categorization process due to entitativity a nd results in a loss of individuality for the members of that category. Comparing the Importance of BIT with the Importance of EIT This dissertation contributes to the literat ure not only within the sponsorships area of inquiry, but also within the broader domai n of the brand leverage process by providing direct evidence of image transfer. As de scribed earlier, studie s revolving around the concept of associating a bra nd with another entity such as brand extension, brand alliances, co-branding, or sponsorship (i.e., brand leverage techniques) only assume knowledge transfer rather than measuring it (e.g., Park, Milberg and Lawson 1991; Roy and Cornwell 2002, 2003; Ruth and Simonin 2003). In other studies, the design does not allow to test for image transfer (e.g., Gr hos, Wagner, and Svetecka 2004; Martin and Stewart 2001; Simonin and Ruth 1998). In fact, the only study that directly assessed knowledge transfer, whether in a sponsorshi ps setting or in another domain of brand leverage, is the experiment conducte d by Gwinner and Eaton (1999). They operationalized image transfer as the differe nce in image congruency between the event and the sponsoring brand across a sponsors hip treatment group and a control group.
121 Other works such as the one by Grhos, Wagner, and Svetecka (2004) relied on a correlational study in which the correlation be tween the events image and the sponsors image was used to assert image transfer. It seems difficult to claim that sponsorship triggers image transfer outside of the experime ntal realm due to the impossibility to test for causal links. This dissert ation contributed to the lite rature by providing direct evidence of a causal re lationship between multiple sponsorsh ips and image transfer at the implicit level not only from the event to the sponsors, but also among the sponsors themselves. This result is consistent w ith the findings from Crawfo rd et al. (2002) in social psychology and tends to show th at the cognitive processes in volved in the perception of social groups are applicable to the perception of brands as well. Some researchers have claimed that the anthropomorphic view of bra nds initiated by Aaker (1997) is unrealistic due to the higher level of complexity of hu man beings (i.e., Azoulay and Kapferer 2003; Caprara, Barbaranelli, and Guido 2001). This dissertation suggests, however, that the literature on group perception is adapted to the study of bra nds. Probably the degree of complexity of the observer is as important as the degree of complexity of the observed when determining which concepts or tools ar e appropriate in a re search domain. In addition, this dissertation investigated th e impact of entitativity within the boundary conditions of categorization. Image transfer effects were contingent on respondents clustering of the sponsoring brands according to the categories s port and no sport.
122 Contribution of the Savings in Relearning Paradigm to the Marketing Literature Finally, this dissertation is the first one to use Ebbinghaus (1885/1964) savings in relearning paradigm in a marketing context. Although a few adjustments had to be made compared to experimental or social ps ychology (e.g., giving respondents more time to memorize paired-associations du e to the higher complexity level of the experimental stimuli used), this paradigm proved itself effective in investigating implicit memory. Marketing researchers are becoming more inte rested in studying the implicit aspect of consumer knowledge (e.g., Brunel, Tieje a nd Greenwald 2004; Lee 2002). Johar and Pham (1999) suggested the investigation of the implicit impact of sponsorships on consumer perception. This dissertation conf irms and extends Johar and Phams (1999) findings that sponsor identification is biased toward market prominent brands and brands semantically related to the event at the explicit memory level. The tag-lines recognition was a measure of explicit memory because respondents were previously explicitly exposed to these tag-lines. It was found, in the case of multiple sponsorships, that tag-lines recogni tion was impaired, not necessarily because respondents relied on heuristics like J ohar and Pham (1999) found, but because entitativity made the brands appear interc hangeable. Therefore, similar to Johar and Phams (1999) results, it was found that sponsor ships adversely impact explicit memory. Of course, their focus was on sponsor identif ication. This dissertations focus was on knowledge about the sponsors image, but these results extend the negative influence of sponsorships to a different aspect of e xplicit memory (i.e., sponsor awareness vs. sponsors tag-line recognition). In addition, their inquiry about the implicit impact of sponsorships was followed and the savings in relearning paradigm showed that
123 potentially beneficial conseque nces could be reaped from sponsorships due to implicit image transfer effects. Managerial Implication The most important conclusion that can be drawn from this dissertations findings for marketing practitioners is the necessity to adopt a different approach concerning sponsorships and especially multiple sponsorsh ips. It is vital to consider multiple sponsorships situations from a holistic standpo int and to take into account all the possible sources of brands associations that result from such operatio ns. First and foremost, the decision of participating in multiple sponsorships agreement should not be taken without considering the other pot ential sponsoring brands that will also be present. As shown, peoples implicit memory about the sponsori ng brands images might be impacted not only by the events image but also by the imag es of the other concomitant sponsors. The likely influence of the other sponsors can potentially be beneficial or harmful for the brand depending on the other sponsors imag es. Lets assume that a brand has the objective of improving its image as a corporat e citizen. If the event sponsored has an image of high corporate citizen ship and if the target brand is sponsoring that event, the extent to which other concomitant spons ors are perceived as being good corporate citizens would have to be taken into account as well. A more in-depth look at this dissertat ions results suggests more specific recommendations. When brand managers c onsider event sponso rship, it is not only necessary to identify the other sponsors of th e event, but also to consider how they fit together with the target brand (i.e., are the underlying brand-concepts similar?). If a
124 subgroup among the concomitant sponsors share th e same brand-concepts, they are likely to constitute an entitative gr oup and to have their respective images implicitly transferred to each other. They will not get, however, image transferred from the event. On the other hand, if a subgroup of brands does not fit together (i.e., their underlying brandconcepts are all different), then these brands are more likely to get some image transferred from the event, but not from each other. This implies that a brand manager could still be willing to sponsor an event whose image does not correspond to the brand as long as the brand fits well with the images of the subgrou p of concomitant sponsors. In that case, because the subgroup will be entita tive, image transfer w ill take place within that group, but not from the event to the gr oup. Conversely, a brand manager could also decide to engage in multiple sponsorships in spite of the presence of concomitant sponsors with an image ill-adapted to his goals as long as his brand does not fit with any of these other brands and that the event does fu lfill his objectives in terms of image. For example, lets assume that Red Bull has an energetic image and wants to make its image evolve toward a more established brand, a re al sport drink. A c oncomitant sponsor of that event is Gatorade that has an authentic image but wants to veer toward an image more related to power and ener gy; another concomitant sponsor is a brand of watch that wants to cultivate an image of durability. If these three br ands sponsor a boat race that carries an image of endurance and durability, th ey could all achieve their goals. Gatorade and Red Bull because they will be forming an entitative group (i.e., drink as a brandconcept) and will transfer image to each other. The watch brand and the event will form an entitative group and some image will be transferred from the event to the brand.
Figure 17 summarizes the decision rules that marketing managers could adopt when deciding about whether or not to get involved in multiple sponsorships activity. Figure 17. Multiple Sponsorships Decision Rules Concomitant Sponsoring Brands Is the target brand similar to the concomitant sponsoring brands? Yes N o 125 Are the concomitant sponsors images adapted to the target brands image objectives? Sponsored Event Is the events image adapted to the target brands image objectives? Yes N o N o Yes Engage in multiple sponsorships Do not engage in multiple sponsorships Engage in Multiple sponsorships Do not engage in multiple sponsorships It is evident that marketing managers need to rethink multiple sponsorships as a possibility to cooperate with other brands for win-win situations. It is quite possible that two brands might be direct competitors but have different, yet mutually interesting, images as well as similar underlying brand-concepts. In that case, cooperation as
126 concomitant sponsors could be considered. Ma nagers of events could also use some of these findings for strengthening their value proposition in search of sponsors. Some sponsors might be reluctant to become associ ated with an event because they consider there is little fit between thei r brand and the image of the event. Events managers could object that this low fit might prevent them from being associated with other concomitant sponsors (given that these bra nds fit with the event). As a result, their brand might be influenced exclusively by the events image. Limitations and Further Research Although this dissertation sheds light on some of the effects of multiple sponsorships, it has several limita tions that need to be considered when interpreting the results. First, the generalizability of the fi ndings to different cont exts should be made cautiously and any applications of this dissertations results need to be made with some perspective. The use of student subjects is acceptable at the early stage of theory development (Ferber 1977) and efforts were ma de to ensure that the stimuli used were relevant to the respondents. However, further research should be undertaken to replicate these findings with non-student samples. Th e generalizability of the findings is also limited by the particular experimental stimuli used. Clearly, the researcher decided to rely on brands that were either since re or exciting and an event that was sophisticated. Other dimensions from the brand personality scale (Aaker 1997) could be used in further studies. In addition, different ways to operationalize brand image could be used. The decision to rely on bra nd personality was motivated by the existence of a well-established scale, but other possibili ties exist like the creation of ad-hoc scales
127 for the experimental stimuli (i.e., Gwinner and Eaton 1999). Furthermore, due to internal validity concerns, fictitious e xperimental stimuli, were us ed. However, now that the mechanisms of multiple sponsorship are be ginning to be better understood, further replications might involve the use of real brands and real events, which would imply extensive pre-testing. Specifi cally, because the main depe ndent variables are based on memory (i.e., cued-recall and recognition), the le vel of saliency of the brand associations and their accessibility would have to be contro lled for. Brand familiarity is an important variable that impact thes e two factors (Low and Lamb 2000; Simonin and Ruth 1998) that is not involved when using fictitious experimental stimu li but that would need to be taken into account otherwise. Another important issue concerns the exte nt to which the manipulations can be compared to real exposure to multiple sponsorships stimuli. In the paradigm used, respondents were exposed for a few minutes to black and white copies of a web page mentioning that some brands would be spons oring the 2005 Boston Golf Tournament. In reality, consumers are exposed to a much larger amount of stim uli through on site signage exposure and/or media coverage for an extended period of time. This might lead to much more perceived entitativity than what was obtained through the manipulation performed. As a consequence, this dissert ations findings are po ssibly a conservative estimate of the real impact of multiple sponsorships on image transfer. Field experiments could alleviate this limitation in the future by replicating this study in a natural setting. Note, however, that given the likely sma ll cause size produced by manipulations in a laboratory experiment (Kline 2004), the impor tance of the results should not be underestimated.
128 Although considerable support for the entitativity mode l of image transfer was found, more work is needed to definitely as certain the validity of this conceptual framework. The data strongly supported the notion that entitativity triggers categorybased processing in multiple sponsorships (Fiske and Neuberg 1990). The analysis undertaken did a good job for pr oviding evidence that a group impression is formed due to entitativity and mediates the impact of multiple sponsorships on brand image transfer, however, some more work is needed to defini tely support the entitati vity model of image transfer developed (Keller 2003). The mediation analysis does not completely ensure that the group impression is formed thr ough an abstraction process resulting from entitativity. Further research could try to replicate these findings by using a different timing for the manipulation of entitativity. In this dissertation, entitativity was manipulated by giving the respondents the multiple sponsorships information before presenting the brands. This means that if the respondents abstracted a group, they encoded the brand information in light of that group, which allowed category-based processing and image transfer. If the info rmation about the multiple sponsorships is given after the brands are presented (i.e., re trieval stage) then re spondents should not be able to process the brands in light of a group impression when they are exposed to the brands. Therefore, it should preclude category-based processing and image transfer. Another possible alternative explanation for these results is that respondents stereotyped the sponsoring brands not because they sponsored the same event, but because they had the same images. In th e study, brands were either sincere or exciting. Therefore, stereotypic proce ssing could have been triggered by image similarity rather than entitativity. Further research could rule out that explanation by
129 including some sponsoring brands with a neut ral image. If the en titativity model of image transfer is correct, these neutral br ands should still be included in the group impression despite the fact that due to thei r image neutrality they do not fit the brand stereotype characterist ics sincere-exciting. Two important factors that were not incl uded in this dissertation could impact image transfer. First, because fictitious brands and a fictitious event were used, familiarity was not manipulated although it ha s been acknowledged as a major factor influencing the processing of brand inform ation as well as brand evaluation (e.g., Alba and Hutchinson 1987; Fazio 1989). Specifically consumer knowledge structure about familiar brands is more rigid (Fazio 1989); therefore, image transfer for familiar brands could be attenuated. Conversely, if an unfamiliar brand is subsumed under a group impression with other more familiar brands, it is likely that transfer effects for this brand would be stronger. Stronger image transfer effects are also likely from the event to unfamiliar brands compared to familiar brands. A second important factor concerns the way multiple sponsorships have been assumed to increase perceived entitativity. Remember that it was posited that multiple sponsorships would positively influence ent itativity through two ante cedents: proximity and interdependence (Campbell 1958 ). Campbell established that another very important antecedent of entitativity, in fact the most important according to him, was common fate: the extent to which different elements are f ound together at different places in the same situations at the same time. Although common fate is not relevant when one considers multiple sponsorships as a one time experience, it can become relevant when one starts to consider cases in which multiple sponsors ar e tied to the same event several times over
130 the years. This might potentially increase their entitativity even more and moderate image transfer process. In the future, l ongitudinal experimental studies could examine these issues. The impact of common fate mi ght trigger numerous questions. If the same group of brands sponsors a different event ev ery time, does that strengthen Brand Image Transfer as opposed to Event Image Transfer ? If the brands that sponsored the same event together over the years form a very highly entitative group, can a new sponsor easily integrate that group? If a new sponsor joins the existing concomitant sponsors will it also be subsumed under the same group impr ession? If two brands are linked together through an operation of co-branding and one of these brand has also been involved in multiple sponsorships with other concomitant sponsors several times, is there an image transfer not only betwee n the co-brands but also from th e entitative group to the co-brand not involved in the multiple sponsorships? In other words, could co-branding be a link between a group impression resulting from multiple sponsorships and a brand not involved with these sponsors? Also, it is important to noti ce that some results obtaine d could not be definitely explained, such as when the moderating effects of high Need for Closure facilitated learning for BIR trials in multiple sponsorships. Further research is needed for a more indepth understanding of the mechanisms of imp licit memorization. At this point, it is a problem difficult to handle since implicit me mory is an indirectly observed phenomenon. Implicit memory is measured through its obs erved impact on performance (Reber 1989). More progress in experimental and cognitive psychology needs to be accomplished before the impact of multiple sponsorsh ips on implicit knowledge can be more completely addressed.
131 Conclusion This dissertation is the fi rst study to demonstrate that sponsorships can impact a brand in two ways: (1) through the event sponsored and (2) th rough the other brands that sponsor that same event. Specifically, when the sponsor has a brand-concept different from the events brand-concept and different from the other sponsors brand-concepts, it is subjected to image transfer from the event. In the case the sponso rs brand-concept fits with the other sponsors brand-concepts, it is subjected to image transfer from the concomitant sponsors. The entitativity mode l of image transfer proposed was supported. Stereotypic processing is responsible for imag e transfer due to the formation of a group impression. The event and the sponsoring bran ds tend to be perceived as one entity and each element forming that group becomes assi milated into a collective identity which leads the brands to lose their uniqueness. These results indicate that multiple spons orships are formidable avenues for brand management. The leverage process seems to be multidimensional and iterative rather than a two-way linear process. This dissert ation shows that multiple sponsorships can provide numerous sources of brand associations and could constitute a very effective marketing promotion tool for brand manage rs. Finally, by proposing a framework of multiple sponsorships based on entitativity, th is dissertation lays the ground for a more comprehensive study of sponsorship effects in the future.
132 References Aaker, David A. (1991), Managing Brand Equity New York, NY: The Free Press. ---and Kevin Lane Keller (1990), "Consum er Evaluations of Brand Extensions," Journal of Marketing 54 (1), 27-41. Aaker, Jennifer L. (1997), "Dimension of Brand Personality," Journal of Marketing Research 34 (3), 347-56 ----, Veronica Benet-Martnez, and Jordi Garolera (2001), "Consumption Symbols as Carriers of Culture: A Study of Japanese a nd Spanish Brand Personality Constructs," Journal of Personality & Social Psychology, 81 (3), 492-508. ----, Susan Fournier, and S. Adam Bras el (2004), "When Good Brand Do Bad," Journal of Consumer Research 31 (1). Abratt, Russell, Brian C. Clayton, and Leyla nd F. Pitt (1987), "Corporate Objectives in Sports Sponsorship," International Journal of Advertising 6 (4), 299. Alba, Joseph W. and J. Wesley Hutchi nson (1987), "Dimensions of Consumer Expertise," Journal of Consumer Research 13 (4), 411-54. Allport, Gordon W. (1954), The Nature of Prejudice Reading, MA: Addison-Wesley. Anderson, John R. and Gordon H. Bower (1973), Human Associative Memory. New York: John Wiley & Sons. Asch, S. E. (1946), "Forming Impressions of Personality," Journal of Abnormal & Social Psychology, 41, 258-90. Atkins, Charles and Martin Block (1983), "Effectiveness of Ce lebrity Endorsers," Journal of Advertising Research, 23 (February/March), 57-61. Azoulay, Audrey and Jean-Noel Kapferer (2003), "Do Brand Personal ity Scales Really Measure Brand Personality?" Journal of Brand Management, 11 (2), 143-55. Baron, Reuben M. and David A. Kenny (1986), "The Moderator-Media tor Distinction in Social Psychological Research: Con ceptual, Strategic, and Statistical Considerations", Journal of Personality and Social Psychology 51 (6), 1173-82.
133 Barsalou, Lawrence W. (1983), "Ad Hoc Categories", Memory and Cognition, 11 (May), 211-27. Bartlett, Fredefic C. (1932), Remembering: A Study in Experimental and Social Psychology. Cambridge: Cambridge University Press. Becker-Olsen, Karen and Caro lyn J. Simmons (2002), "Whe n Do Social Sponsorships Enhance or Dilute Equity? Fit, Message S ource, and the Persistence of Effects," in Advances in Consumer Research, Susan Broniarczyk and Kent Nakamoto (Eds.) Vol. 29. Valdosta, GA: Association fo r Consumer Research, 287-89. Benett, R. (1999), "Sport Sponsorship, Spectators Recall and False Consensus," European Journal of Marketing 34 (3/4), 291-313. Berger, Ida E., Peggy H. Cunningham, and Robert V. Kozinets (1999), "Consumer Persuasion through Cause Related Marketing," in Advances in Consumer Research Vol. 26: Association for Consumer Research. Berry, Leonard L. (2000), "Culti vating Service Brand Equity", Journal of the Academy of Marketing Science, 28 (1), 128-37. Boush, David M. and Barbara Loken (1991), "A Process-Tracing Study of Brand Extension Evaluation," Journal of Marketing Research 38 (February), 16-28. ----, Shannon Shipp, Barbara Loken, Esra Gentrk, Susan Crockett, Ellen Kennedy, Bettie Minshall, Dennis Misurell, Linda Ro chford, and Jon Strobel (1987), "Affect Generalizations to Similar and Dissimilar Brand Extensions," Psychology and Marketing 4 (3), 225-37. Brewer, Marilynn B. (1988), "A Dual Process Model of Impression Formation," in Advances in Social Cognition, Thomas K. Srull and Robert S. Jr. Wyer, Eds. Hillsdale, NJ: Herlbaum. ----, and Amy S. Harasty (1996), "Seeing Gr oups as Entities: The Role of Perceiver Motivation," in Handbook of Motivation and Cognition, Richard M. Sorrentino and E. Tory Higgins, Eds. Vol. 3. New York: Guilford Press. ----, Joseph G. Weber, and Barbara Carini (1995), "Person Memory in Intergroup Contexts: Categorization Versus Individuation," Journal of Personality and Social Psychology, 69 (1), 29-40. Broniarczyk, Susan M. and Joseph W. Alba (1994), "The Importance of the Brand in Brand Extension," Journal of Marketing Research 31 (2), 214-28.
134 Brown, Thomas J. and Peter Dacin (1997), "The Company and the Product: Corporate Associations and Consumer Product Responses," Journal of Marketing 61 (January), 68-84. Brunel, Frdric F., Brian C. Tietje, a nd Anthony Greenwald (2004), "Is the Implicit Association Test a Valid and Valuable Measure of Implicit Consumer Social Cognition," Journal of Consumer Psychology 14 (1), 385-404. Campbell, Donald T. (1958), "Common Fate, Si milarity, and Other I ndices of the Status of Aggregates of Persons as Social Entities," Behavioral Sciences 3, 14-25. Caprara, Gian Vittorio, Claudio Barbaran elli, and Gianluigi Guido (2001), "Brand Personality: How to Make the Metaphor Fit?," Journal of Economic Psychology, 22 (June), 377-95. Carlston, Donald E. and John J. Skowronski (1994), "Savings in th e Relearning of Trait Information as Evidence for Spont aneous Inference Generation," Journal of Personality and Social Psychology 66 (5), 840-56. ---(1999), "Spontaneous Trait Transference to Familiar Communicators: Is a Little Knowledge a Dangerous Thing?," Journal of Personality & Social Psychology 77 (2), 233-46. Carrillat, Francois A., Eric G. Harris, and Barbara A. Lafferty (2004a), "The Moderating Role of Sponsor Familiarity on the Effect of Single and Multiple Sponsorships on Brand Attitude and Purchase Intention," in AMA Winter Educator Conference, Charles Chandler (Ed.). Scottsdale, AZ. ---, ----, and ---(2004b), "Investigati ng Fortuitous Image Transfer in Multiple Sponsorships," Working Paper, University of South Florida, Ma rketing Department. Cartwright, Dorwin a nd Alvin Zander (1960), Group Dynamics: Research and Theory Evanston, Ill: Peterson Row. Chapman, Kenneth J. and Andrew Ayleswor th (1999), "Riding th e Coat-Tails of a Positive Review: Rave Reviews and Attitude Transfer," The Journal of Consumer Marketing 16 (5), 418-37. Cohen, Joel B. and Kunal Basu (1987), "Alternative Models of Categorization: Toward a Contingent Processing Framework," Journal of Consumer Research 13 (4) (Mars), 455-72. Cohen, Neal J. and Larry R. Squire (1980) "Preserved Learning and Retention of Pattern-Analyzing Skill in Amnesia: Di ssociation of Knowing How and Knowing That," Science 210 (4466), 207-10.
135 Cornwell, Bettina and Isabel Maignan (1998) "International Revi ew of Sponsorship Research," Journal of Advertising 27 (1), 1-21. Crawford, Matthew T., Steven J. Sherman, and David L. Hamilton (2002), "Perceived Entitativity, Stereotype Formation, and th e Interchangeability of Group Members," Journal of Personality and Social Psychology 83 (5), 1076-94. Crimmins, James and Martin Horn (1996), "Spo nsorship: From Management Ego Trip to Marketing Success," Journal of Advertising Research 36 (July-August), 11-21. Crowley, Martin G. (1991), "Priorit ising the Sponsorship Audience," European Journal of Marketing 25 (11), 11-21. Dambron, Patrick (1993), Mcnat et Sponsoring, La Commmunication+ Paris: Editions Organizations. Dean, Dwane Hal (2002), "Associating the Corp oration with a Charitable Event through Sponsorship: Measuring the Effects on Corporate Community Relations," Journal of Advertising, 31 (1), 77-87. ---(1999), "Brand Endorsement, Popularity, and Event Sponsorsh ip as Advertising Cues Affecting Consumer Pre-Purchase Attitudes," Journal of Advertising 28 (3), 1-12. Desai, Kalpesh Kausik and Kevin Lane Ke ller (2002), "The Eff ects of Ingredient Branding Strategies on Host Brand Extendibility," Journal of Marketing 66 (1), 7393. Durgee, Jeffrey F. and Robert W. Stuart (1987), "Advertising Sy mbols and Brand Names That Best Represent Key Product Meanings," Journal of Consumer Marketing 4 (3), 15-24. Ebbinghaus, Hermann (1885/1964), Memory: A Contribution to Experimental Psychology. New York: Dover. Farquhar, Peter (1989), "Managing Brand Equity," Marketing Research 1 (September), 24-33. Fazio, Russell (1989), "How Do Attitude Guide Behaviors?," in The Handbook of Motivation and Cognition: Founda tions of Social Behavior, Robert M. Sorrentino and Edward T. Higgins, Eds. Mahwah: NJ: Lawrence Erlbaum Associates. Ferber, Robert (1977), "Research by Convenience," Journal of Consumer Research 4 (1), 57-58.
136 Ferrand, Alain. and Monique Pages (1999), "Ima ge Management in Sport Organizations: The Creation of Value," European Journal of Marketing 33 (3-4), 387-401. Fiske, Susan T. and Shelly E. Taylor (1991), Social Cognition Reading, MA: AddisonWestley. ---and S. L. Neuberg (1990), "A Continuum of Impression Formation, from CategoryBased to Individuating Processes: Influe nces of Information and Motivation on Attention and Interpretation," in Advances in Experimental Social Psychology, M. P. Zana, Ed. New York: Academic Press. Friedman, Hershey H. and Linda Friedman (1979), "Endorser Effectiveness by Product Type," Journal of Advertising Research 19 (October), 63-71. Gartner, Lowell and John Schopler (1998), "P erceived Group Entitativity and Intergroup Bias: An Interconnection of Self and Others," European Journal of Social Psychology, 28 (6), 963-80. Gillies, Caroline (1991), Business Sponsorship: The Sponsor's Guide Boston, MA: Butterworth-Heinemann. Goodman, Nelson (1972), "Seven Strictures on Similarity," in Problems and Projects, Nelson Goodman, Ed. New York: Bobbs-Merrill. Greenwald, Anthony G. (1992), "Unc onscious Recognition Reclaimed," American Psychologist 47 (6), 766-79. ----, Debbie E. McGhee, and Jordan L. K. Schwartz (1998), "Measuring Individual Differences in Implicit Cognition: The Implicit Association Test," Journal of Personality & Social Psychology 74 (6), 1464-80. Gregan-Paxton, Jennifer and Deborah Ro edder John (1997), "Consumer Learning by Analogy: A Model of Internal Model Transfer," Journal of Consumer Research 24 (3), 266-84. Grohs, Reinhard, Udo Wagner, and Sa bine Vsetecka (2004 ), "Assessing the Effectiveness of Sport Sponsorship san Empirical Examination", Schmalenbach Business Review 56 (April), 119-38 Gwinner, Kevin (1997), "A Model of Image Creation and Image Transfer in Event Sponsorship," International Marketing Review 14 (3), 145-58. ---and John Eaton (1999), "Building Br and Image through Event Sponsorship: The Role of Image Transfer," Journal of Advertising 28 (4), 47-57.
137 Hamilton, David L. and Steven J. Sherma n (1996), "Perceiving Persons and Groups," Psychological Review 103 (2), 336-55. Harvey, Bill (2001), "Measuring the Effect of Sponsorship," Journal of Advertising Research 41 (1), 59-65. Hastie, Reid and Bernadette Park (1986) "The Relationship between Memory and Judgment Depends on Whether the Judgment Task Is Memory-Based or on-Line," Psychological Review 93 (3), 258-68. Hayman, C. A. Gordon and Endel Tulving (1989), "Is Priming Fragment Completion Based on a "Traceless" Memory System?," Journal of Experimental Psychology, Learning, Memory, and Cognition 15 (5), 941-56. Heath, Timothy B., Michael S. McCart hy, and David L. Motherbaugh (1994), "Spokeperson Fame and Vividness Effects in the Context of IssueRelevant Thinking: The Moderating Role of Competitive Setting," Journal of Consumer Research 20 (March), 520-34. Hoek, Janet, Gendall Philip, Michelle Jeffc oat, and David Orsman (1997), "Sponsorship and Advertising: A Comparison of Their Effects," Journal of Marketing Communications, 3 (1), 21-32. IEG (2003), "IEG Sponsorship Repo rt." Chicago, IL: IEG Inc. IEG, Glossary and Lexicon (2003), "Interna tional Event Group Web Site, Available At: www. Sponsorship.Com/Learn/ Glossary.Asp," Vol. 2004. Hunter, John E. and Frank L. Schmidt (2004), Methods of Meta-Analysis: Correcting Error and Bias in Research Findings (Second ed.). Thousand Oaks, CA: Sage Publications. Jalleh, Geoffrey, Robert J. Donovan, Billie Gile s-Corti, and C. D'Arcy J. Holman (2002), "Sponsorship: Impact on Brand Awareness and Brand Attitudes," Social Marketing Quarterly 3 (1), 35-45. Javalgi, Rajshekhar G., Mark B. Traylor, Andrew C. Gross, and Edward Lampman (1994), "Awareness of Sponsorship and Corporate Image: An Empirical Investigation," Journal of Advertising 23 (4), 47-58. Johar, Gita Venkataramani and Michel Tu an Pham (1999), "Relatedness, Prominence, and Constructive Sponsor Identification," Journal of Marketing Research 36 (3), 299-312.
138 Johnson, Michael D. (1984), "Consumer Choice Strategies for Comparing Noncomparable Alternatives," Journal of Consumer Research 11 (3), 741-53. Kamins, Michael A. (1990), "An Investiga tion into the "Match-up" Hypothesis in Celebrity Advertising; When Beauty May Be Only Skin Deep," Journal of Advertising, 19 (1), 4-13. ----, Meribeth J. Brand, and Stuwart A. Hoeke (1989), "Two-Sided Versus One-Sided Celebrity Endorsements: The Impact on Advertising Effectiveness and Credibility," Journal of Advertising 18 (2), 4-10. Keller, Kevin Lane (2003), "Brand Synthesis: The Multidimensionality of Brand Knowledge," Journal of Consumer Research 29 (4), 595-600. ---(1993), "Conceptualizing, M easuring, and Managing Customer-Based Brand Equity," Journal of Marketing 57 (1), 1-22. ---and David A. Aaker (1992), "The Eff ects of Sequential In troduction of Brand Extensions," Journal of Marketing Research 29 (1), 35-50. Kelley, Harold H. (1973), "The Proc esses of Causal Attribution," American Psychologist 28 (February), 107-28. Khale, Lynn B. and Pamela M. Homer (1985) "Physical Attractiveness of Celebrity Endorsers: A Social Adaptation Perspective," Journal of Consumer Research 11 (March), 954-61. Kline, Rex B. (2004), Beyond Significance Testing: Reform ing Data Analysis Methods in Behavioral Research Washington, DC: American Psychological Association. Krueger, Joachim and Russel W. Clemen t (1994), "Memory-Based Judgments About Multiple Categories: A Revision and Extens ion of Tajfel's Accentuation Theory," Journal of Personality and Social Psychology 67 (1), 35-47. Kruglanski, Arie W. and Tallie. Freund (1983), "The Freezing and Unfreezing of LayInferences: Effects on Impr essional Primacy, Ethnic St ereotyping, and Numerical Anchoring," Journal of Experime ntal Social Psychology 19 (5), 448-68. ----, Donna M Webster, and Adena Klem ( 1993), "Motivated Resistance and Openness to Persuasion in the Presence or Absence of Prior Information," Journal of Personality & Social Psychology 65 (6), 861-76. Langmeyer, Lynn and Mary Walker (1991), "A First Step to Identify the Meaning in Celebrity Endorsers," in Advances in Consumer Research, Rebecca Holman and Michael Solomon (Eds.) Vol. 18: Association for Consumer Research.
139 Lardinoit, Thierry and Christian Derbaix ( 2001), "Sponsorship and Recall of Sponsors," Psychology & Marketing 18 (2), 167-90. ---and Pascale G. Quester (2001), "Attitudina l Effects of Combined Sponsorship and Sponsor's Prominence on Basketball in Europe," Journal of Advertising Research, 41 (1), 48-58. Lee, Myung-Soo, Dennis M. Sandler, and Davi d Shani (1997), "Attitudinal Constructs Towards Sponsorship: Scale Development Using Three Global Sporting Events," International Marketing Review 14 (3), 159-69. Leibniz, Gottfried W. (1916), New Essays Concerning Human Understanding Chicago, IL: Open Court. Lee, Angela Y. (2002), "Effects of Implic it Memory on Memory-Based Versus StimulusBased Brand Choice," Journal of Marketing Research 39 (4) (November), 440-54. Levy, Sydney J. (1958), "Symbols by Whic h We Buy," in Advancing Marketing Efficiency, Lynn H. Stockman (Ed.) Vol. December. Chicago: American Marketing Association. Levy, Joan Meyers and Alice M. Tybout ( 1989), "Schema Congruity as a Basis for Product Evaluation," Journal of Consumer Research 16 (June), 39-54. Lickel, Brina, David L. Hamilton, Amy Lewis, Steven J. Sherman, Grazyna Wieczorkowska, and A. Neville Uhles (2000), "Varieties of Groups and the Perception of Group Entitativity," Journal of Personality and Social Behavior 78 (2), 223-46. Lingle, John H. and Thomas M. Ostrom (1979) "Retrieval Selectiv ity in Memory-Based Impression Judgments," Journal of Personality & Social Psychology, 180-94. Loken, Barbara and Deborah Roedder John ( 1993), "Diluting Brand Beliefs: When Do Brand Extensions Have a Negative Impact?," Journal of Marketing Research 57 (3), 71-84. Low, George S. and Charles W. Jr. Lamb (2000), "The Measurement and Dimensionality of Brand Associations," Journal of Product & Brand Management 9 (6), 350-68. Lynch, James and Drue Schuler (1994), "T he Matchup Effect of Spokesperson and Product Congruency: A Schema Theory Interpretation," Psychology & Marketing 11 (5), 417-45. MacDonald, Collin (1991), Sponsorship and the Image of the Sponsor," European Journal of Marketing 25 (1991), 31-38.
140 MacLeod, Colin M. (1988), "Forgotten but Not Gone: Savings for Pictures and Words in Long-Term Memory," Journal of Experimental Psychology: Learning, Memory, and Cognition 14 (2), 195-212. Macrae, C. Neil, M. Hewstone, and R. J. Griffiths (1993), "Processing Load and Memory for Stereotype-Based Information," European Journal of Social Psychology 23, 7787. ----, Alan B. Milne, and Galen V. Bodenha usen (1994), "Stereotypes as Energy-Saving Devices: A Peek Inside the Cognitive Toolbox," Journal of Personality & Social Psychology, 66 (1), 37-47. Madrigal, Robert (2000), "The Influence of Social Alliances with Sports Teams on Intentions to Purchase Co rporate Sponsors' Products," Journal of Advertising 29 (4), 13-24. ---(2001), "Social Identity Effects in a BeliefAttitude-Intention Hierarchy: Implications for Corporate Sponsorships," Psychology & Marketing 18 (2), 145-65. Marshall, D.W and G Cook (1992), "T he Corporate (Sports) Sponsor," International Journal of Advertising, 11 (4), 307. Martin, Ingrid M. and David W. Stewart (2001), "The Differential Impact of Goal Congruency on Attitudes, Intentions, and the Transfer of Brand Equity," Journal of Marketing Research 38 (4), 471-84. Matthews, Robert C., Ray R. Russ, William B. Stanley, Fredda Blanchard-Fields, Jeung Ryeul Cho, and Barry Druhan (1997), "Role of Implicit and Exp licit Processes in Learning from Examples: A Synergistic Effect," Journal of Experimental Psychology: Learning, Memory, and Cognition 15 (6), 1083-100. McConnell, Allen R., Steven J. Sherman, a nd David L. Hamilton (1994), "On-Line and Memory Based Aspects of Individual and Group Target Judgments," Journal of Personality & Social Psychology 67 (2), 173-85. ----, ----, and ---(1997), "Tar get Entitativity: Implications for Information Processing About Individual and Group Targets," Journal of Personality and Social Psychology, 72 (4), 750-62. McCracken, Grant (1989), "Who Is the Celebr ity Endorser? Cultural Foundations of the Endorsement Process," Journal of Consumer research 16 (3), 310-22.
141 McDaniel, Stephen R. (1999), "An Investig ation of Match-up Effects in Sport Sponsorship Advertising: The Implicati ons of Consumer Advertising Schemas," Psychology & Marketing 16 (2), 163-84. McDaniel, Stephen R. and Lance Kinney ( 1998), "The Implications of Recency and Gender Effects in Consumer Response to Ambush Marketing," Psychology & Marketing 15 (4), 385-403. McDaniel, Stephen R. and Daniel S. Mason (1999), "An Exploratory Study of Influence on Public Opinion toward Alcohol and T obacco Sponsorship of Sporting Event," Journal of Services Marketing, 13 (6), 481-94. Medin, Douglas L. (1989), "Concepts and Conceptual Structure," American Psychologist 44 (12), 1469-81. ----, Robert L. Goldstone, and Dedre Gent ner (1993), "Respects for Similarity," Psychological Review 100 (2), 254-78. ---and Marguerite M. Schaffer (1978), "Context Theory of Classification," Psychological Review 85 (May), 207-38. Meyvis, Tom and Chris Janiszewski (2004), "Why Are Broader Brands Stronger? An Accessibility Perspective on the Success of Brand Extensions," Journal of Consumer Research 31, 2 (September), 346-57. Mervis, Carolyn B. and Eleanor Rosch (1981), "Categorization of Natural Objects," Annual Review of Psychology, 32, 89-115. Meenaghan, Tony (1983), "Commercial Sponsorship," European Journal of Marketing 25 (11), 5-10. ---(1991), "Sponsorship: Legitimising the Medium," European Journal of Marketing 25 (11), 5-10. ---(1998), "Current Development and Fu ture Directions in Sponsorships," International Journal of Advertising, 17 (1), 3-28. ---(2001), "Sponsorship and Advertising: A Comparison of Consumer Perceptions," Psychology & Marketing 2 (18), 191-215. ---and David Shipley (1999), "Media E ffect in Commerci al Sponsorship," European Journal of Marketing 33 (3/4), 328-47.
142 Miyazaki, Anthony D. and Angela G. Morgan (2001), "Assessing Market Value of Event Sponsoring: Corporate Ol ympic Sponsorships," Journal of advertising Research 41 (1), 9-15. Moore, Elizabeth S., William L. Wilkie, and Richard J. Lutz (2002), "Passing the Torch: Intergenerational Influences as a Source of Brand Equity," Journal of Marketing 66 (2), 17. Moreau, C. Page, Arthur B. Markman, and D onald R. Lehmann (2001), ""What Is It?" Categorization Flexibility and Consumers' Responses to Really New Products," Journal of Consumer Research 27 (4) (March), 489-98. Mowen, John C. and Stephen W. Brown ( 1981), "On Explaining and Predicting the Effectiveness of Celebrity Endorsers," in Advances in Consumer Research, Kent B. Monroe (Ed.) Vol. 8. Ann Arbor, Mi: Association for Consumer Research. Murphy, Gregory L. and Douglas L. Medin (1985) "The Role of Theories in Conceptual Coherence," Psychological Review 92 (July), 289-316. Nelson, T. O. (1985), "Ebbinghaus' Contributio n to the Measurement of Retention: Savings During Relearning," Journal of Experimental Psychology, Learning, Memory, and Cognition, 11, 472-79. ----, ----, and H. A. Laskey (1994), "Sports Event Sponsorship for Brand Promotion," Journal of Applied Business Research 10 (4), 35-40. Nichols, J.A.F, Sydney Roslow, and Sandipa Dublish (1999), "B rand Recall and Brand Preference at Sponsored Golf and Tennis Tournaments," European Journal of Marketing 33 (3/4), 365-86. Ohnian, Roobina (1991), "The Impact of Ce lebrity Spokespersons' Perceived Image on Consumers' Intention to Purchase," Journal of Advertising Research 31 (February/March), 46-54. Park, C. Whan, Sandra Milberg, and Robert Lawson (1991), "The Evaluation of Brand Extension: The Role of Product Features Si milarity and Brand Concept Consistency," Journal of Consumer Research 18 (2), 185-93. ----, Sung Youl Jung, and Allan D. Schocker (1996), "Composite Brand Alliances: An Investigation of Extensi on and Feedback Effects," Journal of Marketing Research 33 (November), 453-66. Petty, Richard E., John T. Cacioppo, and David Schumann (1983), "Central and Peripheral Routes to Advertising Eff ectiveness: The Moderating Role of Involvement," Journal of Consumer Research 10 (September), 135-46.
143 Pham, Michel Tuan and Gita Venkataramani Johar (1997), "Contingent Processes of Source Identification," Journal of Consumer Research 24 (3), 249-65. ---and Gita Venkataramani Johar (2001), "Market Prominence Biases in Sponsor Identification: Processes and Consequentiality," Psychology & Marketing 18 (2), 123-43. ---and Marc Vanhuele (1997), "Analyzing the Memory Impact of Advertising Fragments," Marketing Letters, 8 (4), 407-17. Pope, Nigel (1998), "Consumption Values, Sponsorship Awareness, Brand and Product Use," Journal of Product & Brand Management 7 (2), 124-36. Pracejus, John Walter (1998), "An Inference-Based Model of Building Brand Equity through Sponsorship," Dissertati on, University of Florida. Quester, Pascale G. and Francis Farrelly (1998), "Brand Association and Memory Decay Effects of Sponsorship: The Case of the Australian Formula One Grand Prix," Journal of Product & Brand Management 7 (6), 539-56. ---and Beverley Thompson (2001), "Advert ising and Promotion Leverage on Arts Sponsorship Effectiveness," Journal of Advertising Research 41 (January/February). Rao, Akshay R., Lu Qu, and Robert W. Ruekert (1999), "Signaling Unobservable Product Quality through a Brand Ally," Journal of Marketing Research 36 (2), 25868. Reber, Arthur S. (1989), "Implic it Learning and Tacit Knowledge," Journal of Experimental Psychology, Learning, Memory, and Cognition 118 (3), 219-35. Roediger, Henry L. III. (1990), "Retention without Remembering," American Psychologist 45 (9), 1043-56. ----, M. S. Weldon, and B. H. Challis (1989) "Explaining Dissociati ons between Implicit and Explicit Measures of Retention: A Pro cessing Account," in Varieties of Memory and Consciousness: Essays in Honour of Endel Tulving, Henry L. III. Roediger and F. I. M. Craik, Eds. Hillsdale, NJ: Erlbaum. Rogier, Anouk and Vincent Yzerbyt (1999), "Social Attribution, Correspondence Bias, and the Emergence of Stereotypes," Swiss Journal of Psychology 58 (4), 233-40. Rosch, Eleanor and Carolyn B. Mervis (1975) "Family Resemblances: Studies in the Internal Structures of Categories," Cognitive Psychology 7 (4), 382-439.
144 ----, ----, W. D. Gray, D. M. Johnson, and P. Boyes-Braem (1976), "Basic Objects in Natural Categories", Cognitive Psychology 8, 382-439. Rothbart, Myron and Marjorie Taylor (1992), Language, Interaction and Social Cognition Newbury Park, CA: Sage. Roy, Donald P. and T. Bettina Cornwell (2003), "Brand Equity Influence on Responses to Event Sponsorships," Journal of Product & Brand Management 12 (6), 377-93. ---(2004), "The Effects of Consumer Knowledge on Respons es to Event Sponsorships," Psychology & Marketing 21 (3), 185-207. ---(2002), "The Influence of Brand Equ ity on Consumer Responses to Sports Sponsorships," in American Marketing Association Educator's Conference Proceedings: Enhancing Knowledge Developm ent, S. J. Crove and G. W. Marshall (Eds.): American Marketi ng Association, Chicago, IL. Ruth, Julie A. and Bernard L. Simonin ( 2003), "'Brought to You by Brand a and Brand B': Investigating Multiple Sponsors' Influence on Consumers' Attitudes toward Sponsored Events," Journal of Advertising 32 (3), 19-30. Shani, David and Dennis M. Sandler (1998), "Ambush Marketing: Is Confusion to Blame for the Flickering of the Flame?," Psychology & Marketing 15 (4), 367-83. Schacter, Daniel L. (1987), "Implicit Memory: History and Current Status," Journal of Experimental Psychology; Learning, Memory, and Cognition 13 (3), 501-18. Seger, Carol Augart (1994), "Implicit Learning," Psychological Bulletin 115 (2), 163-96. Sherman, Steven J., Luigi Castelli, and Davi d L. Hamilton (2002), "The Spontaneous Use of a Group Typology as an Organizing Principle in Memory," Journal of Personality & Social Psychology 82 (3), 328-42. Simonin, Bernard L. and Julie A. Ruth ( 1998), "Is a Company Known by the Company It Keeps? Assessing the Spillover Effects of Brand Alliances on Consumer Brand Attitudes," Journal of Marketing Research 35 (1), 30-42. Skowronski, John J., Donal E. Carlston, Lynda Mae, and Matthew T. Crawford (1998), "Spontaneous Trait Transference: Comm unicators Take on the Qualities They Describe in Others," Journal of Personality & Social Psychology 74 (4), 837-48. Smith, Eliot R. and N. Branscombe (1988), "C ategory Accessibility as Implicit Memory," Journal of Experime ntal Social Psychology 24, 490-504. Smith, Edward and Douglas L. Medin (1981), Categories and Concepts Cambridge: MA: Harvard University Press.
145 Speed, Richard and Peter Thompson (2000), "Determinants of Sports Sponsorship Response," Journal of the Academy of Marketing Science 28 (2), 226-38. Srull, Thomas K. and Robert S. Wyer (1989), "Persons Memory and Judgment," Psychological Review 96 (1), 58-83. Stangor, Charles and David McMillan (1992) "Memory for Expectancy-Congruent and Expectancy-Incongruent Information: A Review of the Social and Social Developmental Literatures," Psychological Bulletin 111 (1), 42-61. Stipp, Horst (1998), "The Impact of Olym pic Sponsorship on Corporate Image," International Journal of Advertising, 17 (1), 75-88. ---and Nicholas P. Schiavone (1996), "Mode ling the Impact of Olympic Sponsorship on Corporate Image," Journal of Advertising Research 36 (July-August), 22-27. Sujan, Mita (1985), "Consumer Knowledge: Effects on Evaluation Strategies Mediating Consumer Judgments," Journal of Consumer Research 12 (1), 31-46. Tajfel, Henri (1959), "Quantitative J udgment in Social Perspective", British Journal of Psychology, 62, 207-18. ---and John C. Turner (1985), "The Social Identity Theory of Group Behavior," in Psychology of Intergroup Relations, Stephe n Worchel and William G. Austin, Eds. Chicago, IL: Nelson-Hall Publishers. Till, Brian D. and Michael Busler (2000), "The Match-up Hypothesis: Physical Attractiveness, Expertise, a nd the Role of Fit on Brand Attitude, Purchase Intent and Brand Beliefs," Journal of Advertising 39 (3), 1-13. ---and ----` (1998), "Matching Products with Endorsers: Attr activeness Versus Expertise," Journal of Consumer Marketing 15 (6), 576-86. Titchener, E. B. (1923), "R elearning after 46 Years," American Journal of Psychology, 34, 468-69. Tripp, Carolyn, Thomas D. Jensen, and Les Carlson (1994), "The Effects of Multiple Prodtc Endorsements by Celebrities on Consumers' Attitudes and Intentions," Journal of Consumer Research 20 (1), 535-47. Tulving, Endel and Daniel L. Schacter ( 1990), "Priming and Human Memory Systems," Science 247 (4940), 301-06. Tversky, Amos (1977), "Features of Similarity," Psychological review 84 (4), 327-52.
146 Ulibarri, Carlos A. (2000), "Rationa l Philanthropy and Cultural Capital," Journal of Cultural Economics 23, 135-46. Webster, Donna M. and Arie W. Kruglanski ( 1994), "Individual Differences in Need for Cognitive Closure," Journal of Personality & Social Psychology 67 (6), 1049-62. Welbourne, Jennifer L. (1999), "The Impact of Perceived Entitativity on Inconsistency Resolution for Groups and Individuals," Journal of Experimental Social Psychology 35. Whittlesea, Bruce W. A. and Richard L. Wright (1997), "Implicit (and Explicit) Learning: Acting Adaptively withou t Knowing the Consequences," Journal of Experimental Psychology: Learning, Memory, and Cognition 23 (1), 181-200. Yzerbyt, Vincent Y., Anouk Rogier, and Susa n T. Fiske (1997), "Group Entitativity and Social Attribution: On Translating Situ ational Constraints into Stereotypes," Personality and Social Psychology Bulletin 24, 1090-104. ----, Steve Rocher, and Georges Schadron (1997), "Stereotypes as Explanations: A Subjective Essentialistic View of Group Perception," in The Social Psychology of Stereotyping and Group Life, Russell Sp ears and Penelope J. Oakes, Eds.
Appendix 1 Summary of Empirical Studies on Sponsorship Author (Year) Journal ConceptualBackground Study (Type, Subjects, Sample size) Stimuli (Brand, Event) Main Findings Roy and Cornwell (2004) Psychology & Marketing Schema Theory Experiment, Students, 402 Real, real Expert consumers engage in more negative thoughts about the sponsor-event combination than novices (piecemeal processing); both experts and novices perceive a high equity sponsor and an event to be congruent whereas only experts perceive a low equity sponsor and an event to be incongruent. Ruth and Simonin (2003) Journal of Advertising Schema Theory Experiment, Students, 219 Real, fictitious In the case of multiple sponsorships (2 brands) attitude toward both sponsors is positively related to attitude toward the event. Events associated with controversial sponsors have a less favorable attitude than when associated with a non controversial sponsor when the sponsor is a domestic brand. Grohs, Wagner, and Vsetecka (2004) Schmalenbach Business Review Heuristics, Human Associative Memory Experiments, General Population, 132 Real, real Ambush marketing seems to be minimal with only 2 sponsors out of 6 being misidentified. The closer the fit between the event and the sponsors the higher the sponsors recall. Event involvement and event exposure are positively related with unaided recall of sponsors. Image transfer depends on post-event sponsors image, event image and sponsors awareness. 148
Appendix 1 Continued Madrigal (2000) Journal of Advertising Social Identity Theory Survey, adults attending college football game, 678 Fictitious, real The positive impact of team identification on purchase intention is greater when the group norm is not important for the fans than when it is important. Dean (1999) Journal of Advertising Balance Theory Experiment, Students, 185 Fictitious, real Sponsorship positively impacts perceived corporate citizenship Dean (2002) Journal of Advertising Balance Theory, Attribution Theory SEM, Students, 272 Real, real Sponsorships positively impacts consumer-perceived-corporate-community of sponsors. Sponsorship leads to the formation of both positive and negative consumer attribution. Bennet (1999) European Journal of Marketing Mere exposure Effects, False consensus effect Soccer game spectators, survey, 789 Real, real Unprompted and aided recall of event sponsor is positively related to the frequency of game attendance; it does not follow the inverted U shape usually found with mere exposure effects. In addition, respondents tended to believe that spectators attending the game were more inclined to buy products of the sponsor than the general population. Nicholls, Roslow, and Dublish (1999) European Journal of Marketing N/A Golfandtennis tournament spectators, survey, 762 Real, real Sponsorship leads to unaided sponsor recall and brand preference even more so for a tennis than for a golf tournament. 149
Appendix 1 Continued Brown and Dacin (1997) Journal of Marketing Consumer inference Social judgment theory Study 1: students, survey, 148 Study 2: students, survey, 127 Study 3: mall intercepts surveys, shoppers, 229 Study 1: Fictitious, fictitious Study 2: Real, real Study 3: Fictitious, fictitious Study 1: Corporate ability (CA) associations and corporate social responsibility (CSR) association both affect product response but in a different manner. CA through product attribute and CSR through the overall corporate reputation. Study 2: the impact of corporate ability is stronger with real brands than the impact of corporate social responsibility. Surprisingly, product social responsibility negatively impacts product evaluation. Study 3: There is evidence of a context effect that differs according to the type of corporate association: evaluation of a good product is greater in the case the CA is negative than when it is positive. Evaluation of a good product is greater when the CSR is positive than when it is negative. Becker-Olsen and Simmons (2002) Advances in Consumer Research Congruency Real, real Low fit between sponsor and a social cause generates less favorable thoughts, attitudes, affective, and behavioral responses to the firm than high fit. These result hold when the sponsor artificially creates the fit; also the effects of fit persist one year after the manipulation. When the sponsorships is announced by the sponsored social cause rather then the sponsor, the benefits for the sponsor are greater (this finding does not hold over time). 150
Appendix 1 Continued Johar and Pham (1999) Journal of Marketing Research Heuristics Study1:Students, experiment, 44 Study 2: Students, experiment, 65 Study 3: Students, experiment, 78 Real, real When consumers cannot directly retrieve the sponsor of a given event they rely on heuristics. Brands that are semantically related to the event (relatedness bias) and that dominate the market (prominence bias) are more likely to be falsely identified as sponsors of a given event. Roy and Cornwell (2003) Journal of Product & Brand Management Schema Theory Experiment, Students, 402 Real, real High equity sponsors are perceived as being more congruent with the sponsored event; sponsor-event congruence is positively related to attitude toward the sponsor. Javalgi et al. (1994) Journal of Advertising N/A Phone survey,200 Real, real Different dimension of corporate image are impacted differently by sponsorship. Overall, corporate sponsorship improves corporate image independently of the type of sponsorship considered (charitable, cultural, and sporting events). Madrigal (2001) Psychology & Marketing Social Identity Theory Telephone survey, Households, 368 Fictitious, real Beliefs and identification level are positively related to attitude toward corporate sponsors. Attitudes predict purchase intention of sponsors more strongly for fans with a low level of identification. The difference in purchase intention between low and high identification is greater in the case of unfavorable than favorable attitude. 151
Appendix 1 Continued Pham and Johar (2001) Psychology & Marketing Heuristic Students,experiment, 34 Real, real Replication of the prominence bias found in Johar and Pham (1999): consumers relying on an hypothesis testing procedure for identifying sponsors. In addition, they showed that more prominent sponsors benefit more in terms of image relevant to the sponsored activity. Stipp and Schiavone (1996) International journal of Advertising N/A Phone surveys,General population, 479 Real, real The corporate image of a particular sponsor of the Olympics is impacted by attitude toward the general Olympic Games corporate sponsors, the recall of advertisements of the sponsor, how much liked these ads are, and how strong is the linkage between the sponsor and the event. Quester and Farrelly (1998) The Journal of Product & Brand Management N/A Pre/Postlongitudinal phone survey, General population, 250 every year for 4 years Real, real Unprompted recall is positively impacted by the involvement of the sponsor in the core activity of the event, the congruency between the sponsor domain of activity and the event, and especially by the local factor (consumer living close to where the area takes place exhibit higher unprompted recall scores). Quester and Thompson (2001) The Journal of Advertising Research N/A Experiment,general population, 248 Real, real Sponsorship of art festival increases attitude toward sponsor. Lardinoit and Quester (2001) The journal of Advertising Research N/A Experiment,240 Real, real On site and television sponsorships both improve attitude toward the sponsor for non-prominent brands only. Miyazaki and Morgan (2001) The Journal of Advertising Research N/A Event studyanalysis, firms, 27 Real, real The announcement of sponsorship agreement with the Olympic Games positively impacts sponsors stock value. 152
Appendix 1 Continued Shani and Sandler (1998) Psychology & Marketing N/A Survey, generalpopulation, 1500 Real, real Consumers are confused about sponsorship right, which leads to ambush marketing. A third of the consumers think that advertisers during the Olympic telecast are official sponsors. In addition, consumers do not distinguish among the different degrees of commitment of different sponsor category independently of their level of involvement with the Olympics. In addition, consumers do not exhibit negative attitudes toward ambush marketers. McDaniel (1999) Psychology & Marketing Schema Experiment,students, 216 Real, real Consumers degree of involvement with the sponsored event and degree of relevance of the media used are positively related to the impact of a sponsorship advertisement on attitude toward the ad for males. For females there was an impact on purchase intention, also the impact on attitude toward the ad was stronger than for males. Hoek et al. (1997) Journal of Marketing Communications Awareness-trial-Reinforcement Model from advertising Survey, Students, Real, real Contrary to advertising, sponsorship impact on recall is stronger for non-users than users. The impact of advertising on brand beliefs is greater for users than non-users whereas this difference of impact is not consistent for sponsorships. However, neither sponsorship nor advertising increased the probability of purchase of users and non users. Jalleh, Giles-Corti, and Holman (2002) Social Marketing Quarterly N/A Questionnaires,spectators of the event, 355 Real, real The impact of health sponsorship on awareness and attitude is greater than the impact of commercial sponsorship. 153
Appendix 1 Continued McDaniel and Mason (1999) Journal of Services marketing N/A Phone survey,general population, 248 Real, real Consumers judge more acceptable the sponsorship of the Olympics by a beer company than by a cigarette company. In addition, the acceptance toward the use of Olympic sponsorship to promote beer is related to ones use of beer/alcohol, attitude toward advertising and beer expectancies. Finally, acceptance toward the use of Olympic sponsorships to promote tobacco products is related to attitude toward advertising and tobacco product expectancies. McDaniel and Kinney (1998) Psychology & Marketing N/A Experiment,Students, 215 Real, real Ambush sponsors can be recalled as being the official sponsor of an event (no difference across males and females). In addition, males and females do not differ in terms of their response to sponsorship and ambush sponsorship stimuli. Gwinner and Eaton (1999) Journal of Advertising Schema Experiment,Students, 360 Real, real Image congruency between a brand and an event is greater when that brand sponsors the event (i.e., image transfer). This effect is stronger when the brand and the event present a functionally based or an image-based similarity. In the case image or function based similarity is very low, image congruency was found to be greater in the case of no sponsorship than in the case of sponsorship (i.e., contrast of image). Nicholls, Roslow, and Laskey (1994) Journal of Applied Business Research N/A Golftournament spectators, survey, 276 Real, real Mixed results concerning the positive link between number of days attending the event and preference for the sponsoring brand. 154
Appendix 1 Continued Speed and Thompson (2000) Journal of the Academy of Marketing Science Classical conditioning Experiment, Students, 195 Real, real Attitude toward the sponsor, perceived sincerity and sponsor-event fit positively impact interest in the sponsor, favorability toward the sponsor and intention to use the sponsors products. The status of the sponsored event is positively related to interest and favorability. Perceived ubiquity of the sponsor was negatively related to interest and favorability. In addition, the positive impact of personal liking on interest, favorability and use was stronger when there was a fit between the event and the sponsors. Also, the impact of the event status on interest, favorability and use was weaker in the case of fit between the sponsor and the event. Lardinoit and Derbaix (2001) Psychology & Marketing N/A Experiment.Students, 240 Real, real Television sponsorship is more effective than field sponsorship for improving unaided and aided recall. In addition, the impact of field sponsorship on unaided recall is greater for highly involved consumers whereas the impact of television sponsorship is not moderated by involvement. Carrillat, Harris, and Lafferty (2004a) Proceedings of the AMA Winter Educators Conference Human Associative Memory Experiment, Student, 158 Real, real The impact of sponsorship on attitude and purchase intention is greater for unfamiliar than familiar brands. Multiple sponsorships do not dilute the impact of sponsorship on attitude and purchase intention for both familiar and unfamiliar brands. Pope (1998) The Journal of Product & Brand Management Consumption value Survey, Students, 921 Real, real Sponsorship awareness is positively related to social, functional, epistemic, and emotional consumption value depending on the product category of the sponsor. 155
Appendix 2 Experimental Stimuli Used in the Study About Us Drinks Ingredients Sponsorship Athletes Culture Competition Contact Us Thanks to its unique formula, Energetic is the only energy drink that contains all the electrolytes necessary to keep your blood concentration of potassium, sodium, and glucose stable during your effort. Energetic is an energy drink developed for those who demand maximum energy from their body. Energetic is pure concentration of power and stamina that will allow you to take all the challenges. Explore Your Limits Explore Your Limits 156
Appendix 2 Continued Certified by ISSA the world leader in fitness certification 157
Appendix 2 Continued 158
Appendix 2 Continued 159
Appendix 2 Continued Boston Boston 160
Appendix 2 Continued Demonstrate Your Character Home History Press Contact Us N ews Products People Goodies Company Escape Lin k Presentation Ever since its creation in Europe in 1985, Glissade has been a unique brand for sportswear garments and accessories that has always stood out by its expression of a peerless attitude. Glissade has brought about a genuine revolution not only in the world of sportswearbut also in its communication. Features The introduction of "fun" clothes in their cut, prints and materials. Shorts and pullovers come alive. Ever since its beginnings, Glissade has always positioned itself as an avant-garde brand totally different from all its competitors. Demonstrate Your Character 161
Appendix 2 Continued Aunts Mary Company Products Locate us Contact Us Link 162
Appendix 2 Continued 100 % Adrenaline About Us Night-Club is one of the most famous and fashionable venue in Boston. Ever since a million dollar renovation, The Night Club has grown to be a hotspot in Boston's nightlife scene. With an amazing light system, an even better sound system, new, sleek style and first-rate house music, The Night Club has become the place to be. Don't take our word for it, as you get hazy at the bar, and crazy on the dancefloor. The Night Club is the most spectacular production around, an extreme club experience. This new club is steeped in incredible theatrics, featuring indoor fireworks, sheer sheets of silk bathed in live club video, and dancers above the crowd on a 100-foot catwalk. Tonight Events Calendar Clubvibes Events VIP Passes Guestlists The Scene Add Event/Club Link 100 % Adrenaline 163
Appendix 2 Continued 164
165 Appendix 3 Scale Items Pre-test Brand Personality Scale (Aaker 1997) Consider the brands and events that were presented to you previously and how well they are described by the adjectives below. For each brand, choose a number between 1 and 7 where means that the adjective does not describes at all brand[event] X and means that the adjective describes very well brand [event]X Dimension (Column not shown to respondents) Adjective 1= not at all describe 7 = describe very well Sincerity Down-to-Earth 1 2 3 4 5 6 7 Family-oriented 1 2 3 4 5 6 7 Small-town 1 2 3 4 5 6 7 Honest* 1 2 3 4 5 6 7 Sincere* 1 2 3 4 5 6 7 Real 1 2 3 4 5 6 7 Wholesome 1 2 3 4 5 6 7 Original 1 2 3 4 5 6 7 Cheerful 1 2 3 4 5 6 7 Sentimental 1 2 3 4 5 6 7 Friendly 1 2 3 4 5 6 7 Excitement Daring 1 2 3 4 5 6 7 Trendy 1 2 3 4 5 6 7 Exciting 1 2 3 4 5 6 7 Spirited 1 2 3 4 5 6 7 Cool* 1 2 3 4 5 6 7 Young* 1 2 3 4 5 6 7 Imaginative 1 2 3 4 5 6 7 Unique 1 2 3 4 5 6 7 Up-to-date 1 2 3 4 5 6 7 Independent 1 2 3 4 5 6 7 Contemporary 1 2 3 4 5 6 7 Sophistication Upper class 1 2 3 4 5 6 7 Glamorous* 1 2 3 4 5 6 7 Good looking* 1 2 3 4 5 6 7 Charming 1 2 3 4 5 6 7 Feminine 1 2 3 4 5 6 7 Smooth 1 2 3 4 5 6 7 *: Used for manipulation checks during the main experiment
Appendix 3 Continued Need for Closure Scale (Webster and Kruglanski 1994) Please read the following set of statements and indicate the extent to which you disagree or agree with these statements by selecting a number between 1 and 6 where 1 indicates that you strongly disagree and 6 indicates that you strongly agree . You may use any of the numbers in the middle as well to show how much you agree or disagree with these statements. 1=strongly disagree 6=strongly agree I think that having clear rules and order at work is essential for success. 1 2 3 4 5 6 Even after I've made up my mind about something, I am always eager to consider a different opinion. a 1 2 3 4 5 6 I don't like situations that are uncertain. 1 2 3 4 5 6 I dislike questions which could be answered in many different ways. 1 2 3 4 5 6 I like to have friends who are unpredictable. a 1 2 3 4 5 6 I find that a well ordered life with regular hours suits my temperament. 1 2 3 4 5 6 When dining out, I like to go to places where I have been before so that I know what to expect. 1 2 3 4 5 6 I feel uncomfortable when I don't understand the reason why an event occurred in my life. 1 2 3 4 5 6 I feel irritated when one person disagrees with what everyone else in a group believes. 1 2 3 4 5 6 I hate to change my plans at the last minute. 1 2 3 4 5 6 I don't like to go into a situation without knowing what I can expect from it. 1 2 3 4 5 6 When I go shopping, I have difficulty deciding exactly what it is that I want. a 1 2 3 4 5 6 When faced with a problem I usually see the one best solution very quickly 1 2 3 4 5 6 When I am confused about an important issue, I feel very upset. 1 2 3 4 5 6 I tend to put off making important decisions until the last possible moment. a 1 2 3 4 5 6 167
Appendix 3 Continued I usually make important decisions quickly and confidently. 1 2 3 4 5 6 I would describe myself as indecisive. a 1 2 3 4 5 6 I think it is fun to change my plans at the last moment. a 1 2 3 4 5 6 I enjoy the uncertainty of going into a new situation without knowing what might happen. a 1 2 3 4 5 6 My personal space is usually messy and disorganized. a 1 2 3 4 5 6 In most social conflicts, I can easily see which side is right and which is wrong. 1 2 3 4 5 6 I tend to struggle with most decisions. a 1 2 3 4 5 6 I believe that orderliness and organization are among the most important characteristics of a good student. 1 2 3 4 5 6 When considering most conflict situations, I can usually see how both sides could be right. a 1 2 3 4 5 6 I don't like to be with people who are capable of unexpected actions. 1 2 3 4 5 6 I prefer to socialize with familiar friends because I know what to expect from them. 1 2 3 4 5 6 I think that I would learn best in a class that lacks clearly stated objectives and requirements. a 1 2 3 4 5 6 When thinking about a problem, I consider as many different opinions on the issue as possible. a 1 2 3 4 5 6 I like to know what people are thinking all the time. 1 2 3 4 5 6 I dislike it when a person's statement could mean many different things. 1 2 3 4 5 6 It's annoying to listen to someone who cannot seem to make up his or her mind. 1 2 3 4 5 6 I find that establishing a consistent routine enables me to enjoy life more. 1 2 3 4 5 6 I enjoy having a clear and structured mode of life. 1 2 3 4 5 6 168
Appendix 3 Continued I prefer interacting with people whose opinions are very different from my own. a 1 2 3 4 5 6 I like to have a place for everything and everything in its place. 1 2 3 4 5 6 I feel uncomfortable when someone's meaning or intention is unclear to me. 1 2 3 4 5 6 When trying to solve a problem I often see so many possible options that it's confusing. a 1 2 3 4 5 6 I always see many possible solutions to problems I face. a 1 2 3 4 5 6 I'd rather know bad news than stay in a state of uncertainty. 1 2 3 4 5 6 I do not usually consult many different opinions before forming my own view. 1 2 3 4 5 6 I dislike unpredictable situations. 1 2 3 4 5 6 I dislike the routine aspects of my work (studies). a 1 2 3 4 5 6 a: reversed items Manipulation Checks (Lickel et al. 2000) Brand-concept Indicate the extent you agree or disagree with the following statements by circling the number that best express your opinion. Statement Strongly Neutral Strongly Disagree Agree 1 2 3 4 5 6 7 The product category of Aunt Marys Gourmet Treat is related to sports 1 2 3 4 5 6 7 I associate Aunt Marys Gourmet Treat with the idea of sports 1 2 3 4 5 6 7 169
170 Appendix 3 Continued Entitativity 5 Consider the 8 companies that were fi rst introduced to you earlier as being sponsors of the 2005 Boston Golf Tournament The 8 sponsoring companies form a group. We would like to you to rate this group on some characteristics. In the space below, circ le a number that represents your opinion about the extent to which the companies: Night-Club, Aunt Marys Gourmet Treat, Urbane, Boston Health & Fitness, Glissade, Massachusetts Bank & Trust, Fitness and Energetic share characteristics. For example if the sponsoring companies do not share charac teristics you would circle a (1) if they share only a few characteristics you would circle a (4), and if they share many characteristics you would circle a (9). 1=companies do not form a group 9=companies form a group 1 2 3 4 5 6 7 8 9 Personal and Category Relevance Indicate the extent to which you agree with these statements using a number form 1 to 7 where 1 means that you strongly disagree wi th the statement and 7 mean that you strongly agree with the statement. 1= strongly disagree 7 = strongly agree The brand image of brand X is relevant to me 1 2 3 4 5 6 7 The brand image of brand X makes sense to me 1 2 3 4 5 6 7 Brand X is relevant in the Y product/service category 1 2 3 4 5 6 7 Brand X makes sense in the Y product/service category 1 2 3 4 5 6 7 5 This was the manipulation check for the multiple sponsorships condition; in the no sponsorship condition any reference to the sponsorship was omitted.
Appendix 4 Study Experimental Design Multiple Sponsorships No Sponsorship No Sport Brands Consistent Paired-Associations (BIR) No Sport Brands Inconsistent Paired-Associations (BIT) No Sport Brands Consistent Paired-Associations (BIR) No Sport Brands Inconsistent Paired-Associations (BIT) Brand related Image-Traits Sport Brands Consistent Paired-Associations (BIR) Sport Brands Inconsistent Paired-Associations (BIT) Sport Brands Consistent Paired-Associations (BIR) Sport Brands Inconsistent Paired-Associations (BIT) No Sport Brands Consistent Paired-Associations (BIR) No Sport Brands Inconsistent Paired-Associations (EIT) No Sport Brands Consistent Paired-Associations (BIR) No Sport Brands Inconsistent Paired-Associations (EIT) Event Related Image-Traits Sport Brands Consistent Paired-Associations (BIR) Sport Brands Inconsistent Paired-Associations (EIT) Sport Brands Consistent Paired-Associations (BIR) Sport Brands Inconsistent Paired-Associations (EIT) The bolded variables represent between subject factors, cells separated by the bolded straight lines represent repeated measures for a given subject. The No Sport and the Sport brands stand for the within subject factor of brand-concept similarity. Inconsistent paired-associations are BIT trials in the Brand Related Image-Traits condition and they are EIT trials in the Event Related Image-Traits condition. 171
Appendix 5 Results of Experimental Stimuli Pretest Table 5-1. Image and Relevance Ratings of the Target Brands and the Event Underl y in g Brand-Concept Sincerity (SD) Excitement (SD) Sophistication (SD) Usage During Experiment Brand Relevance (SD) Category Relevance (SD) Fitness Sport 5.06 (0.69) 4.08 (1.13) 3.79 (1.30) Sincere Sponsor 5.75 (1.05) 6.1 (0.91) Health Sport 4.87 (0.98) 4.03 (1.23) 3.50 (1.11) Sincere Sponsor 5.19 (1.28) 5.69 (1.21) Massachusetts Bank No Sport 5.05 (1.06) 3.30 (1.25) 3.48 (1.27) Sincere Sponsor 4.81 (1.45) 5.59 (1.36) Aunt Mary's Treat No Sport 4.81 (0.90) 3.66 (0.91) 3.69 (1.10) Sincere Sponsor 6.06 (1.01) 6.35 (0.79) Glissade Sport 4.36 (1.01) 5.68 (0.83) 3.74 (1.31) Exciting Sponsor 4.57 (1.21) 5.26 (1.42) Energetic Sport 3.37 (1.05) 4.89 (1.07) 3.06 (1.09) Exciting Sponsor 5.4 (0.93) 6.27 (0.65) Urbane No Sport 3.48 (1.16) 5.30 (1.08) 4.29 (1.36) Exciting Sponsor 5.43 (1.24) 6.1 (0.91) Night-Club No Sport 3.57 (0.92) 5.69 (0.79) 4.03 (1.27) Exciting Sponsor 5.86 (1.06) 6.38 (0.91) Boston Golf Tournament Sport 4.00 (1.00) 4.24 (0.87) 4.85 (1.03) Sophisticated Event 4.88 (0.98) 6.04 (0.71) 173
Appendix 5 Continued Table 5-2. Image and Relevance Ratings of the Foil Brands Sincerity (SD) Excitement (SD) Sophistication (SD) Usage During Experiment Brand Relevance (SD) Category Relevance (SD) Chic 3.55 (0.92) 4.12 (1.04) 5.26 (0.97) Sophisticated Foil 4.72 (1.09) 5.83 (1.18) Excellence 3.17 (1.11) 4.45 (1.39) 5.01 (1.08) Sophisticated Foil 5.56 (1.32) 5.87 (1.07) SuperRide 4.16 (1.16) 4.97 (1.10) 2.70 (1.43) Exciting Foil 4.92 (1.58) 5.86 (1.53) Radio Wave 3.61 (1.08) 4.27 (1.02) 3.14 (1.36) Exciting Foil 5.02 (1.44) 5.75 (1.33) Crystal 4.99 (0.71) 3.76 (1.15) 3.30 (1.39) Sincere Foil 5.46 (1.08) 4.42 (1.46) Healing 4.69 (1.15) 4.00 (1.09) 3.34 (1.36) Sincere Foil 5.50 (0.96) 6.26 (0.81) 174
Appendix 6 The Multiple Sponsorships Manipulation The companies on the next pages will be sponsors of the 2005 Boston Golf Tournament. You may not recognize the names of the companies as they are all from the Northeast part of the United States. It is difficult to become a sponsor of this locally popular event. The Boston Golf Tournament limits the number of companies as sponsors. The Boston Golf Tournament plans to have activities, programs and advertisements prior to the tournament, during the tournament as well as public relations campaign after the tournament. The sponsoring companies will appear together as a group on all signage for the event. In addition, the names and logos of the companies will be seen together on all television, newspaper, and magazine ads such as in the example below: Boston 175
Appendix 7 Manipulation Checks Results Table 7-1. Similarity Ratings of Target Brands with the Sport Concept Brand-Concept Mean (SD) T-value* Significance Fitness Sport 5.44 (1.44) 12.63 (1,154) < .01 Health Sport 5.14 (1.66) 8.56 (1,154) < .01 Energetic Sport 5.81 (1.29) 17.52 (1,154) < .01 Glissade Sport 6.21 (1.02) 26.88 (1,154) < .01 Night No Sport 1.81 (1.25) -21.89 (1,154) < .01 Massachusetts No Sport 1.26 (0.73) -44.68 (1,154) < .01 Urbane No Sport 1.81 (1.20) -22.79 (1,154) < .01 Aunt Mary No Sport 1.31 (0.83) -40.20 (1,154) < .01 *: t-values result from a comparison of the mean of that brand and the mid-point of the scale (i.e., 4). 176
Appendix 7 Continued Table 7-2. Image Ratings of the Target Brands *: Results of paired sample t-tests between the ratings of the brand on its expected image and its ratings on the image characterizing brands with an expected different image (e.g., the sincerity rating of a sincere brand versus its excitement rating) **: Results of paired sample t-tests between the ratings of the brand on its expected image and its ratings on the image characterizing the event (e.g., the sincerity rating of a sincere brand versus its sophistication rating) ***: Result of one-sample t-tests between the sophistication ratings of the brand and the sophistication rating of the event during the pretest, which is used as a test-value (i.e., 4.85). Image Sincerity (SD) Excitement (SD) Sophistication (SD) Sig 1* Sig 2** Sig 3*** Fitness Sincere 4.85 (1.51) 4.59 (1.19) 4.24 (1.40) t 75 = 1.74. p < .05 t 75 = 3.04, p < .01 t 76 = 3.99, p < .01 Health Sincere 4.68 (1.90) 4.32 (1.65) 3.79 (1.55) t 75 = 1.98, p < .05 t 75 = 4.59, p < .01 t 76 = 6.13, p < .01 Massachusetts Sincere 5.31 (1.33) 2.40 (1.21) 2.95 (1.60) t 76 = 16.15, p < .01 t 76 = 11.34, p < .01 t 76 = 10.45, p < .01 Aunt Mary Sincere 5.38 (1.52) 3.63 (1.61) 3.49 (1.81) t 76 = 8.55, p < .01 t 76 = 9.08, p < .01 t 76 = 6.61, p < .01 Glissade Exciting 3.69 (1.61) 5.89 (1.41) 4.31 (1.52) t 75 = 11.93, p < .01 t 75 = 9.64, p < .01 t 76 = 3.14, p < .01 Energetic Exciting 3.13 (1.43) 4.84 (1.47) 3.49 (1.74) t 75 = 8.19, p < .01 t 76 = 6.66, p < .01 t 76 = 6.83, p < .01 Urbane Exciting 4.05 (1.69) 5.46 (1.65) 4.94 (1.70) t 75 = 8.55, p < .01 t 76 = 3.59, p < .01 t 76 = 0.75, ns Night-Club Exciting 3.38 (1.69) 6.00 (1.45) 5.36 (1.61) t 75 = 12.15, p < .01 t 75 = 5.54, p < .01 t 76 = 2.74, p < .01 177
Appendix 8 ANOVA and ANCOVA Tables Table 8-1. ANCOVA Table for the Outcome Hypotheses H1a to H2b Source of Variation Effect Type DF Mean Square F-value P-value Brand-concept Within-Subjects 1 1098 1.25 .26 Brand-concept Sponsorships Within-Subjects 1 6 191 7.05 .009 Brand-concept Image Within-Subjects 1 2 558 2.91 .09 Brand-concept Closure Within-Subjects 1 2 670 3.04 .08 Brand-concept Sponsorships Image Within-Subjects 1 12 157 13.84 <.001 Brand-concept Sponsorships Closure Within-Subjects 1 5 617 6.40 .012 Brand-concept Image Closure Within-Subjects 1 1 920 2.19 .14 Brand-concept Sponsorships Image Closure Within-Subjects 1 13 479 15.35 <.001 Error Within-Subjects 146 878 Intercept Between-Subjects 1 12992 9.15 .003 Sponsorships Between-Subjects 1 5288 3.73 .056 Image Between-Subjects 1 221 .156 .694 Closure Between-Subjects 1 69 .049 .826 Sponsorships Image Between-Subjects 1 1 901 1.34 .249 Sponsorships Closure Between-Subjects 1 4 448 3.13 .079 Image Closure Between-Subjects 1 74 .052 .820 Sponsorships Image Closure Between-Subjects 1 1 687 1.188 .277 Error Between-Subjects 146 1419 178
Appendix 8 Continued Table 8-2. ANOVA Table for Brand Image Reinforcement (H3a to H4) Source of Variation Effect Type DF Mean Square F-value P-value Brand-concept Within-Subjects 1 14916 17.054 <.001 Brand-concept Sponsorships Within-Subjects 1 13.160 .015 .903 Error Within-Subjects 153 875 Intercept Between-Subjects 1 1 512 747 988.308 <.001 Sponsorships Between-Subjects 1 2553 1.668 .198 Error Between-Subjects 153 69 179
Appendix 8 Continued Table 8-3. ANCOVA Table for the Process Hypotheses H5a to H5d Source of Variation Effect Type DF Mean Square F-value P-value Paired-Association Within-Subjects 1 145 .124 .725 Paired-Association Sponsorships Within-Subjects 1 11 204 9.61 .002 Paired-Association Image Within-Subjects 1 2 388 2.048 .154 Paired-Association Closure Within-Subjects 1 9.59 .007 .932 Paired-Association Sponsorships Image Within-Subjects 1 3 432 2.944 .088 Paired-Association Sponsorships Closure Within-Subjects 1 10 663 9.146 .003 Paired-Association Image Closure Within-Subjects 1 2 662 2.283 .133 Paired-Association Sponsorships Image Closure Within-Subjects 1 3 421 2.934 .089 Error Within-Subjects 146 1166 Intercept Between-Subjects 1 25334 18.365 <.001 Sponsorships Between-Subjects 1 2076 1.505 .222 Image Between-Subjects 1 173 .125 .724 Closure Between-Subjects 1 2144 1.554 .215 Sponsorships Image Between-Subjects 1 9 079 6.581 .011 Sponsorships Closure Between-Subjects 1 1 473 1.068 .303 Image Closure Between-Subjects 1 268 .194 .660 Sponsorships Image Closure Between-Subjects 1 9 736 7.059 .009 Error Between-Subjects 146 1380 180
Appendix 8 Continued Table 8-4. ANCOVA Table for the Process Hypotheses H6a to H6d Source of Variation Effect Type DF Mean Square F-value P-value Paired-Association Within-Subjects 1 2637 3.966 .048 Paired-Association Sponsorships Within-Subjects 1 2.310 .003 .953 Paired-Association Image Within-Subjects 1 269.366 .405 .525 Paired-Association Closure Within-Subjects 1 2 269 3.411 .067 Paired-Association Sponsorships Image Within-Subjects 1 2 692 4.049 .046 Paired-Association Sponsorships Closure Within-Subjects 1 .047 .000 .993 Paired-Association Image Closure Within-Subjects 1 5.5 .008 .928 Paired-Association Sponsorships Image Closure Within-Subjects 1 2 443 3.673 .057 Error Within-Subjects 146 665 Intercept Between-Subjects 1 17477 12.363 .001 Sponsorships Between-Subjects 1 19.75 .014 .906 Image Between-Subjects 1 2403 1.700 .194 Closure Between-Subjects 1 153 .108 .743 Sponsorships Image Between-Subjects 1 218 .154 .695 Sponsorships Closure Between-Subjects 1 71.792 .051 .822 Image Closure Between-Subjects 1 2 505 1.772 .185 Sponsorships Image Closure Between-Subjects 1 656 .464 .009 Error Between-Subjects 146 1414 181
Appendix 8 Continued Table 8-5. ANCOVA Table for the Process Hypotheses H7a and H7b. Source of Variation Effect Type DF Mean Square F-value P-value Brand-concept Within-Subjects 1 246 .507 .478 Brand-concept Sponsorships Within-Subjects 1 73 .150 .699 Brand-concept Closure Within-Subjects 1 2 547 5.256 .023 Brand-concept Sponsorships Closure Within-Subjects 1 51 .106 .745 Error Within-Subjects 150 485 Intercept Between-Subjects 1 13 502 13 502 <.001 Sponsorships Between-Subjects 1 1493 2.205 .140 Closure Between-Subjects 1 69 .102 .750 Sponsorships Closure Between-Subjects 1 2 099 3.101 .080 Error Between-Subjects 150 677 182
Appendix 9 A-priori Contrast Tests Tables Table 9-1. Contrasts Tests for the Outcome Hypotheses H1a to H2b Hypothesis Brand-concept Factor Image Factor Means Compared (Multiple vs. No sponsorships) DF T-value P-value(1-tailed) H1a Sport Brand 60.98 vs. 47.56 80 1.70 .046 H1b No Sport Brand 63.42 vs. 67.07 80 -0.46 .646 H2a Sport Event 59.72 vs. 51.35 71 .996 .161 H2b No Sport Event 83.33 vs. 72.97 71 1.43 .077 Table 9-2. Contrast Test for the Need for Closure Analysis of the Outcome Hypotheses H1a to H2b Need for Closure Factor Brand-concept Factor Image Factor Means Compared (Multiple vs. No sponsorships) DF T-value P-value(1-tailed) Low Sport Brand 61.11 vs. 57.14 37 .339 .368 Low No Sport Brand 52.78 vs. 61.91 37 -.845 .202 High Sport Brand 57.14 vs. 38.89 37 1.644 .054 High No Sport Brand 69.05 vs. 75.00 37 -487 .314 Low Sport Event 72.73 vs. 43.33 35 2.665 .006 Low No Sport Event 80.00 vs. 77.28 35 -.29 .387 High Sport Event 39.29 vs. 56.82 34 -1.453 .077 High No Sport Event 92.86 vs. 68.18 34 2.188 .018 183
Appendix 9 Continued Table 9-3. Contrast Tests For Brand Image Reinforcement Analysis (H3a to H4) Hypothesis Brand-concept Factor Means Compared (Multiple vs. No sponsorships) DF T-value P-value(1-tailed) H3a No Sport 79.87 vs. 73.72 153 1.185 .119 H3b Sport 65.58 vs. 60.36 153 .899 .185 H4 MultipleSponsorships* 79.87 vs. 65.58 76 3.039 .001 *: These means were tested through an independent sample t-test. The means compared were dissimilar vs. similar brand-concept. Table 9-4. Contrast Tests for the Processes Hypotheses H5a to H5d Hypothesis Brand-concept Factor Sponsorship Factor Image Factor Means Compared (Consistent vs. Inconsistent paired -associations) DF T-value P-value (1-tailed) H5a Sport No Brand 62.20 vs. 47.56 40 1.818 .038 H5b Sport Multiple Brand 69.51 vs. 60.98 40 1.096 .140 H5c Sport No Event 58.11 vs. 51.35 36 .867 .195 H5d Sport Multiple Event 61.11 vs. 59.72 35 .167 .435 185
Appendix 9 Continued Table 9-5. Contrast Tests for the Need for Closure Analysis of the Process Hypotheses H5a to H5d. Need for Closure Brand-concept Factor Sponsorship Factor Image Factor Means Compared (Consistent vs. Inconsistent paired -associations) DF T-value P-value (1-tailed) High Sport Multiple Brand 78.57 vs. 57.14 20 2.007 .029 High Sport Multiple Event 60.71 vs. 39.29 13 1.385 .089 Low Sport Multiple Brand 61.11 vs. 61.11 17 0.000 1.000 Low Sport Multiple Event 61.37 vs. 72.73 21 -1.312 .102 High Sport No Sponsorship Brand 63.89 vs. 38.89 17 2.297 .017 High Sport No Sponsorship Event 50.00 vs. 56.82 21 -.680 .252 Low Sport No Sponsorship Brand 61.90 vs. 57.14 20 .384 .352 Low Sport No Sponsorship Event 70.00 vs. 43.33 14 2.477 .013 186
Appendix 9 Continued Table 9-6. Contrast Tests for Process Hypotheses H6a to H6d Hypothesis Brand-concept Factor Sponsorship Factor Image Factor Means Compared (Consistent vs. Inconsistent paired -associations) DF T-value P-value (1-tailed) H6a No Sport No Brand 87.80 vs. 67.07 40 3.759 < .001 H6b No Sport Multiple Brand 80.49 vs. 63.41 40 3.332 .001 H6c No Sport No Event 58.11 vs. 72.97 36 -2.060 .023 H6d No Sport Multiple Event 79.17 vs. 83.33 35 -.723 .237 187
Appendix 9 Continued Table 9-7. Contrast Tests for the Need for Closure Analysis of the Process Hypotheses H6a to H6d. Need for Closure Brand-concept Factor Sponsorship Factor Image Factor Means Compared (Consistent vs. Inconsistent paired -associations) DF T-value P-value (1-tailed) High No Sport Multiple Brand 85.71 vs. 69.05 20 2.092 .024 High No Sport Multiple Event 67.86 vs. 92.86 13 -2.876 .067 Low No Sport Multiple Brand 72.22 vs. 52.78 17 2.715 .007 Low No Sport Multiple Event 86.36 vs. 77.27 21 1.449 .081 High No Sport No Sponsorship Brand 83.33 vs. 75.00 17 .900 .190 High No Sport No Sponsorship Event 54.55 vs. 68.18 21 -1.449 .081 Low No Sport No Sponsorship Brand 90.48 vs. 61.90 20 4.382 < .001 Low No Sport No Sponsorship Event 63.33 vs. 80.00 14 -1.435 .085 188
Appendix 9 Continued Table 9-8. Contrast Tests for the Process Hypotheses H7a and H7b. Hypothesis Brand-concept Factor Means Compared (Multiple vs. No sponsorship) DF T-value P-value (1-tailed) H7a Sport 33.77 vs. 38.78 153 1.196 .117 H7b No Sport 66.32 vs. 74.10 153 2.170 .016 Table 9-9. Contrast Tests for the Need for Closure Analysis of the Process Hypotheses H7a and H7b. Need for Closure Factor Brand-concept Factor Means Compared (Multiple vs. No sponsorship) DF T-value P-value (1-tailed) Low Sport 37.50 vs. 40.28 74 .464 .322 Low No Sport 65.17 vs. 72.92 74 1.439 .077 High Sport 30.00 vs. 37.50 73 -1.226 .112 High No Sport 65.71 vs. 75.75 73 2.055 .021 189
About the Author Franois Anthony Carrillat received his undergraduate degree in Business Administration from the Universit de Savoie-Annecy (France) (1999) and his MS in Marketing from the IAE of Aix-en-Provence-Universit Aix/Marseille III (France) (2000). In 2001 he spent one semester in the exchange MBA program of the University of Florida before joining the University of South Florida where he received his Ph.D. in 2005. His research interests include multiple sponsorships, consumer cognitive structure and content, market oriented strategies, and salesforce performance measurement. His research has been published in several journals including the International Journal of Research in Marketing, the Journal of Personal Selling & Sales Management, and the Academy of Marketing Science Review. He has also published in the conference proceedings of the American Marketing Association, the Association for Historical Research in Marketing, and the Society for Marketing Advances. He will be starting as an assistant professor of marketing at HEC Montral in the summer of 2005.