Citation
An inference-based model of building brand equity through sponsorship

Material Information

Title:
An inference-based model of building brand equity through sponsorship
Creator:
Pracejus, John W
Publication Date:
Language:
English
Physical Description:
viii, 204 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Advertising research ( jstor )
Brands ( jstor )
Celebrities ( jstor )
Coffee ( jstor )
Commercials ( jstor )
Consumer research ( jstor )
Inference ( jstor )
Marketing ( jstor )
Sponsorship ( jstor )
Tennis ( jstor )
Dissertations, Academic -- Marketing -- UF ( lcsh )
Marketing thesis, Ph.D ( lcsh )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph.D.)--University of Florida, 1998.
Bibliography:
Includes bibliographical references (leaves 193-203).
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by John W. Pracejus.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
029546124 ( ALEPH )
40214281 ( OCLC )

Downloads

This item has the following downloads:


Full Text











AN INFERENCE-BASED MODEL OF BUILDING BRAND EQUITY THROUGH SPONSORSHIP













By

JOHN W. PRACEJUS













A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1998



























Copyright 1998

by

John W. Pracejus













ACKNOWLEDGEMENTS



I would first like to thank my chair, Rich Lutz, for guidance, support, and for allowing me to find my own way. I would also like to thank Chris Janiszewski for helping me to think more conceptually about the topic, Bart Weitz for helping me to focus on my strengths, and Mike Weigold for helpful comments and suggestions. Special thanks go to Joffre Swait for providing me with a new perspective on this project and research in general, and for tremendous help with development and analysis of the choice studies.

Others to whom I am grateful for contributing to my intellectual development include Alan Sawyer, Joel Cohen, Joe Alba, John Lynch, and especially Tom O'Guinn, without whom I would never have started this process.

I would also like to acknowledge all of the doctoral students who have made my life at Florida the wonderful experience it has been. Many thanks go to Luk Warlop, Michel Pham, Susan Fournier, and Prasad Niak for providing maps which allowed me to successfully follow the path, and to Norma Mednoza, Andre Menck, and Stijn Van Osselaer, for being so supportive while walking that path with me. Thanks also go to Corinne Faure, Francis Hollman Anne Stringfellow, Stacy Wood, Americus Reed, Lisa Bolton, Kevin Bradford, Tom Meyvis and Velitchka Kaltcheva for ideas, talk and coffee.

Finally, I am thankful to my parents, Walter and Rosemary, without whom none of this would have been possible.


ii














TABLE OF CONTENTS


pne

ACKNOWLEDGENIENTS ................................................................................... iii

A B STR A C T .......................................................................................................

IN TR O D U C TIO N .................................................................................................

CHAPTER ONE EXTANT SPONSORSHIP RESEARCH:
WHERE ARE WE NOW? ...................................................................... 4

CHAPTER TWO SPONSORSHIP WITHIN A BROADER
CATEGORY OF COMNIERCIAL ASSOCIATIONS ................................ 12

CHAPTER THREE TBE POTENTIAL RAPACT OF
SPONSORSHIP ON BRAND EQUITY AND POSITIONING .................. 21

CHAPTER FOUR SEVEN MECHANISMS THROUGH WHICH
SPONSORSHIP MAY FUNCTION .......................................................... 27
Sim ple A w areness .......................................................................... 28
A ff ect Transfer ............................................................................... 30
Im age T ransfer ............................................................................... 32
A ffiliation ...................................................................................... 36
Im plied Size ................................................................................... 38
Implied Endorsement .................................................................... 41
R eciprocity ..................................................................................... 43
C onclu sion ..................................................................................... 44

CHAPTER FIVE DRAENSIONS OF FIT .......................................................... 45
A ttribute F it .................................................................................... 49
A u dience F it ................................................................................... 65
P ositioning F it ................................................................................ 70

CHAPTER SIX SACRILEGE ........................................................................... 74

CHAPTER SEVEN AS SOCIATION-ENHANCING
COMMUNICATIONS (AEC ..................................................................... 80


iv








CHAPTER EIGHT THE IMPACT OF EVENT AND
ASSOCIATIONAL FACTORS ON INFERENCES
AB OU T TH E BRAN D .............................................................................. 86

C onceptual O verview ..................................................................... 89
Hypotheses About Specific Inferences ............................................ 95
S tudy la ......................................................................................... 99
S tu dy lb ......................................................................................... 1 15
S tudy 1c ......................................................................................... 122
A dditional Cross Study .................................................................. 127
Overall Discussion and Limitations of Studies I a- I c ...................... 128

CHAPTER NINE THE IMPACT OF AUDIENCE FACTORS
AND FIT ON BRAND E14FERENCES ...................................................... 131
Study 2a ......................................................................................... 133
Study ................ 139
O verall D iscussion ......................................................................... 141

CHAPTER TEN THE IMPACT OF SPONSORSHIP-BASED
INFERENCES ON BRAND CHOICE ....................................................... 144
Study 3a ......................................................................................... 14 5
Study 3b ......................................................................................... 14 9
Study 3c ......................................................................................... 152
Overall Discussion of Studies 3a, 3b, and 3c .................................. 157


CHAPTER ELEVEN CONCLUSION, IMPLICATIONS AND
FUTURE RE SEAR CH ............................................................................. 159
Im plications ................................................................................... 162
Future R esearch .............................................................................. 165
C onclu sion ..................................................................................... 169

APPENDIX A OPEN ENDED MEASURES AND MANIPULATION
CHECKS FOR EXPERIMENT 1A, 1B, AND 1 C ..................................... 170

APPENDIX B OPEN ENDED INFERENCE MEASURES AND
MANIPULATION CHECKS FOR EXPERIMENT 2A AND 2B .............. 173

APPENDIX C INVOLVEMENT AND KNOWLEDGE MEASURES
USED IN EXPERIMENT 2A .................................................................... 175

APPENDIX D OPEN ENDED THOUGHT LISTING BOXES .......................... 179

APPENDIX E MANTLE CHOICE PROFILE-USED IN
EX PERIM E N T 3 A .................................................................................... 180




V








APPENDIX F SAMPLE CHOICE PROFILE USED IN
EXPERIMENT 3B ........................................ I...................... 181

APPENDIX G SAMPLE CHOICE PROFILE USED IN
EXPERIMENT 3C............................................................... 182

APPENDIX H SAMPLE EXPOSURE STIULI FOR
EXPERIMENTS 1A, IB, 1C, 2A, AND 2B .................................. 183

APPENDIX I ONE POSSIBLE FORM OF A SPONSORSHIP
VALUATION FUNCTION FOR A SPECIFIC BRAND .................... 192

REFERENCES............................................................................ 193

BIOGRAPHICAL SKETCH ............................................................ 204





































vi








Abstract of Dissertation Presented to the Graduate School Of the University of Florida in Partial Fulfillment of the
Requirement for the Degree of Doctor of Philosophy

AN INFERENCE-BASED MODEL OF BUILDING BRAND EQUITY THROUGH SPONSORSHIP By

John W. Pracejus

August 1998

Chairperson: Richard J. Lutz
Major Department: Marketing

Sponsorship, or the intentional association of a brand or company with a sports teams or sporting event, entertainment tour or attraction, festival, fair or annual event charitable cause, the arts, or cultural institutions, is emerging as an important component of the marketing mix. While there is general agreement that sponsorship is effective at exposing consumers to a brand name, little is known about how sponsorship can build brand equity.

A two stage model of building brand equity through sponsorship is proposed. It is proposed that in the first stage, event, associational and audience factors lead consumers to draw different inferences about the sponsoring brand. These inferences include brand size (i.e. "this must be a big brand to be able to sponsor this event"), brand legitimacy (i.e. "this brand must be OK in order to be allowed to sponsor this event"), and event facilitation (i.e. "this brand is making this event possible, or at least making it better"). In the second stage, these inferences are proposed to differentially impact choice probabilities for different categories and brands.






Vii










A series of five studies investigates the role played by event, associational and audience factors on inference making about a sponsoring brand. Evidence is found that brand size inferences are best created through sponsoring large events for a long time, whereas event facilitation inferences are best created through being the title sponsor of smaller events. Three additional studies demonstrate that these two inferences (brand size and event facilitation) have a different impact on the choice probabilities of different categories and brands.




































Viii












INTRODUCTION

As a promotional tool, sponsorship is unique. It is, perhaps, the only means of communicating with customers without the benefit of a verbal message. Any message received must be derived from the association in the mind of a consumer. While traditional advertising often employs simple visual associations between objects and products, the headline, body copy and/or audio usually help to explain what the association means. With sponsorship, however, the association is the entire message. For this reason, the study of sponsorship involves the development and testing of constructs which are quite different from those employed in the study of traditional advertising.

Before going any further, it is necessary to define what is meant by sponsorship. Sponsorship is a commercial association of a brand with a sports teams or sporting event, entertainment tour or attraction, festival, fair or annual event, charitable cause, the arts, or cultural institutions (Andrews 1996). While purchasing advertising time on TV and radio has sometimes been referred to as "sponsoring," buying space in traditional media (i.e. TV, radio, print) is not what is meant by sponsorship here.

Spending on sponsorship (as defined above) has grown more than fivefold in the last ten years, and is expected to exceed $5 billion this year in North America alone (Andrews, 1996). Currently, the bulk of this sum (66%) is spent on sports sponsorship, but arts and cultural sponsorships are growing at the fastest rate. Overall, spending on






1





2

sponsorship is increasing faster than spending on advertising or Wes promotion (Andrews 1996).

While the bulk of the growth in sponsorship activities has taken place over the last twenty years, the idea of creating an exploitable association between a commercial and a noncommercial entity is not a new one. In fact, an association between the product Bovril and the Nottingham Forest Football Club dates from 1898 (Marshall and Cook 1992). In 1927, Charles Lindbergh named his plane the "Spirit of St. Louis" because his financial sponsors wanted to promote St. Louis, Missouri, as an aviation city. Thus, the first transatlantic fight was, in fact, a sponsored event (Hildreth and Nalty, 1969; Lindbergh, 1953). When Jesse Owens won four gold medals in the 1936 Olympics, he did so in Adidas shoes (Ryssel and Stamminger, 198 8, p. 111). Finally, in the first known example of a sporting event being named after a brand, The "Whitbread (Ale) Gold Cup," a horse racing event, first took place in 1956 (Marshall and Cook, 1992).

In spite of this long history as a phenomenon and rapidly growing levels of

expenditure, the mechanisms through which sponsorship affects brand equity, preference and choice, are not well understood. The purpose of this dissertation is to begin systematically exploring the ways in which sponsorship works. It begins with a review of extant sponsorship research and is followed by a summary of relevant research from related "commercial association domains, like celebrity endorsement and brand and line extension. Measurable marketing variables upon which sponsorship may have an impact are then discussed.

Exploratory research and managerial intuition as to how sponsorship works are

then developed into seven testable mechanisms through which sponsorship may function in





3

various situations. These mechanisms are simple awareness, affect transfer, image transfer, affiliation, implied size, implied endorsement, and reciprocity. Constructs which are hypothesized to have an impact upon the functioning of these mechanisms are then developed. These constructs include various types of "fit" between sponsor and event, as well as "sacrilege" or consumer outrage at an association which is perceived to be inappropriate. A theoretical discussion of how other communications can be deployed to maximize the impact of the sponsorship association wraps up the background and theory chapters.

Three empirical studies of how the event and associational factors (event size, brand-event fit, duration of association and share of presence) of a sponsorship impact consumer perceptions of unfamiliar brands are presented in chapter eight. Chapter nine reports the results of two studies which investigate the role of audience characteristics (involvement with and knowledge about the event) on consumer response to a pair of familiar brands. In chapter ten, the impact of sponsorship-based inference on brand choice is investigated for several categories of brands. Chapter 11 provides an overview of the findings of chapters eight through ten and suggests some interesting areas for further study in this emerging area.














CHAPTER ONE

EXTANT SPONSORSHIP RESEARCH: WHERE ARE WE NOW?



Since 1984, there have been at least 35 journal articles which deal, either directly or indirectly, with sponsorship. Hundreds more can be found in the trade press. Despite these facts, very little work beyond describing the occurrence frequency of sponsorship activities has taken place. What has gone beyond description of frequency often involves the reporting of levels of awareness, attitude or purchase intention following a sponsorship. A few tangential topics have also received some attention. Finally, there have been a good number of papers which put forth some ideas of "how sponsorship works". For the most part, though, these papers rely entirely upon anecdotal evidence for support of their assertions. These assertions, which come largely from papers written by practitioners and consultants, however, do provide a wealth of insight into potential mechanisms through which sponsorship may work. The systematic development of such mechanisms follows in a subsequent chapter.

One of the questions most frequently addressed in the sponsorship literature is "why do companies engage in sponsorship?" Abratt, Clayton, and Pitt (1987) found that among 60 sponsoring organizations in South Africa, 11% of respondents claimed that their sponsorship was altruistic "to a large extft," 53% said it was altruistic to "some




4





5


extent" and 36% said it was not altruistic (p. 309). Apart from this one study, however, the issue of altruistic motives in sponsorship is unexplored.

Armstrong (1988) presents the results of a case study analysis of the sponsorship activities of 17 large international electronics companies. Interviews were conducted in four countries (United States, Great Britain, Germany, and France) with corporate and regional marketing managers from the companies, representatives of the sports being sponsored and any intermediary agencies involved (80 interviews total). Within the sample, American firms more frequently listed business managers as the primary target for the sponsorship, while European and Japanese firms tended to use sponsorship to reach the consumer audience. Interestingly, the demographics of these two target audiences were nearly identical (male, upper management, age 35-55). The interviews also revealed that ffims which have been involved with a sponsorship for several years tended to have image enhancement (as opposed to awareness) goals. It is unclear from the data presented in the paper, however, whether companies changed their focus over time, or if companies with awareness goals stopped engaging in sponsorship after awareness goals were met.

Marshall and Cook (1992) asked why 200 British companies from the Times 1000 engaged in sponsorship. Reasons given were "reach target audience" (49%), "product linked" (34%), "image of sport" (14%) and "othee'(3%) (p. 319). They also asked for importance ratings for a number of potential sponsorship goals. Mean importance ratings for "building corporate image," "national link with product or service," "potential advertising exposure," "potential TV coverage," "potential press coverage," and "covers target audience" indicated that all were important or very important. Only "ability to





6

provide corporate hospitality" was rated below the midpoint of the five point importance scale.

Wither et al. (1991) found that companies used sponsorship of cultural and art institutions to reach opinion leaders and to achieve community relations goals, whereas sports sponsorship was used primarily to reach the general public directly. Scott and Suchard (1992) factor analyzed the results of a 23 question survey of 108 Australian businesses. The resultant four factors were regressed on the amount spent on sponsorship by each company. Two of the four factors, performance and client relationships, were found to be significantly correlated to expenditure on sponsorship. The significant correlation of "client relationship" and expenditure is said to be somewhat counter to previous studies (i.e. Marshall and Cook 1992) which found hospitality to be only a marginally important goal.

Another issue which has received some attention in the sponsorship literature is "how do companies go about choosing an event, once they have decided to engage in sponsorship?" Gardner and Shuman (1987) surveyed Fortune 500 companies about their choice of sponsorship activities. Eighty-four percent reported using in-house expertise for screening, while 24% percent utilized outside counsel to evaluate and choose particular events.

In a survey of companies sponsoring football (soccer) in the LT, Thwaites (1995) found that sponsors of Premier League Clubs, which have high profile and cost, were more proactive in securing sponsorship opportunities, compared more clubs before the decision and negotiated longer term contracts, than did sponsors of Division One Leagues, which have moderate profile and cost (all significant at p<.05). This seems to suggest that





7

larger, more sophisticated companies spend more time and effort choosing what they will sponsor than do smaller, less sophisticated firms.

The types of evaluation employed by firms to determine sponsorship success has also been investigated. Gardner and Shuman (1987) found that 17% of firms using sponsorship engaged in audience research. Among consumer firms, 27% used sales or share to determine success while only 8% of commercial firms used this measure. Somewhat startling, however, is that My 47% of responding firms involved in sponsorship reported no post-event evaluation.

There is one other paper which reports the results of an empirical investigation of sponsorship evaluation. In a survey of UK soccer sponsors, Thwaites (1995) found a heavy reliance upon amount of generated media exposure as a measure of sponsorship success. He also found that this exposure was more carefully measured by firms which sponsored large and prominent teams.

Research has also reported on amounts of expenditure on corporate sponsorship. For, example, Meerabeau et al. (199 1) reported $5 billion in global expenditure on sports sponsorship during 1989. This study also details the amount spent by different segments of the "drinks industry," as well as percentages of marketing budgets spent on sponsorship by each segment. Armstrong (1988) surveyed 17 large international electronics companies. He found that larger companies, while spending more on sponsorship in absolute numbers, spent a smaller percentage of their marketing budget on sponsorship than did smaller companies.

Studies have also reported the impact of sponsorship on consumers. By far, the most widely reported measure is awareness. Chew (1992), for example, found that





8

viewers of a sponsored public television program (MacNeiVLehrer News Hour) were more aware that AT&T sponsored the News Hour than were nonviewers.

Crimmins and Horn (1996) point out the need to take into account the number of people who mistakenly believe that a competitor is a sponsor, in addition to actual awareness that your brand is the sponsor of an event. They point out that while 3 5% of a consumer mail panel could correctly identify Coke as the official soft drink of the NFL, that 34% thought it was Pepsi, and note that Coke had paid $250 million to become the official soft drink.

Awareness levels of sponsorship have also been studied in an experimental setting. Hoek, Gendall and Stockdale (1993) exposed New Zealand secondary school students to tapes containing a cricket match with a cigarette brand as the title sponsor (the 1990 Rothman's Cricket Series") and control subjects to a match without cigarette sponsorship. The only significant effect found was that treatment subjects subsequently had a higher "awareness for Rothman's" than did a control group. No significant impact on attitude towards smoking was found.

Some papers have focused on awareness levels "that a company is a sponsor". Wright (1988), for example reports awareness levels for sponsors of different football (soccer) teams in Great Britain. Time series data are provided which demonstrates the various awareness levels over time.

Rajaretnam. (1994) reports the findings of a "unique experiment' in which a large Indian tire maker stopped all product message advertising and spent almost its entire marketing budget on sponsorship from 1984-1987. While the author does not consider in detail other potential factors impacting awareness and image variables (e.g., increased





9

distribution and word of mouth), the numbers are impressive. During the period of the experimentn" top of mind awareness went from 4% to 2 1 %, which is a 425% increase. Unaided recall went from 38% to 68%, and aided awareness went from 92% to 99% (p. 65). The author notes that while the impact of sponsorship on awareness was "almost immediate," the impact on brand preference took a longer time.

One paper attempted to address determinants of sponsorship awareness. Hansen and Scotwin (1995) reported that in an experimental setting, those interested in soccer (measured) had higher awareness of the team sponsors than did those who were not interested. The authors, however, did not report whether these differences are statistically significant.

Other impacts of sponsorship have also been reported such as beliefs about the

company. Javalgi et al. (1994) studied perceptions of five local companies and awareness of their sponsorship activities. They hypothesized that persons who are aware of corporate sponsorship have a more favorable view of the sponsoring company than persons who are unaware. This was not supported. Awareness of the sponsorship activity was only related to the statements "is involved in the community" and "only wants to make money" (reversed), and neither was significant overall. "Is involved in the community" was significantly correlated with awareness of sponsorship for only one of the five companies.

In a study which did not directly involve sponsorship, Nebenzahl and Jaffe (199 1) found that evaluation of Korean products by Israeli consumers was higher after the 1988 Olympics in Seoul than it was before the games among heavy Olympic viewers, but not among a low viewing controlr' group. This suggested to them that an association with the Olympic games had raised people's evaluations of Korea and in turn Korean products.





10

Another impact of sponsorship reported in the literature is purchase intention. In a survey of consumers, Gardner and Shuman (1987) asked people what they perceived to be the impact of sponsorship on their purchasing behavior. Fifty-three percent of respondents said that sponsorship makes them more likely to buy. Eight percent of respondents said sponsorship makes them less likely to buy, and 39% said it had no impact.

Another topic which has received some attention in the sponsorship literature is

the concept of "ambush marketing" (Sandier and Shani, 1989; Meenaghan, 1994). Sandler and Shani (1989) identify and define ambush marketing as a practice whereby a nonsponsoring brand tries to associate itself with a sponsored event. This is done to either gain the benefits of being associated with an event without paying for the privilege, or to limit the impact of a competitor's legitimate sponsorship effort.

Their research found that ambushers did achieve greater false awareness of being sponsors than did non-ambushing, non-sponsoring companies.' The ambushers, however, did not approach the awareness levels achieved by the actual sponsors.

Meenaghan (1994) defines various types of ambushsh. He outlines several ways legitimate sponsors might prevent negative effects of such actions. He concludes by presenting the question of whether "ambush" is moral. The answer to this question, he proposes, is a matter of both the form the ambush takes and "one's perspective". The final issue addressed in the extant sponsorship literature relates to how sponsorship actually works. Most of the papers which address this issue rely almost exclusively upon



1 As measured by recall data, the combined recall-recognition measure showed no difference.








anecdotal evidence drawn from colorful examples (Hastings, 1984; Meenaghan, 1991b; Thomas, 1985; Crowley, 1991; Parker, 1991; McDonald, 1991; Otker, 1988; Ryssel and Stamminger, 1988; Schoch, 1994; Walshe and Wilkinson, 1994). They have titles like "Professional Football Sponsorship--Profit or Profligate" (Thwaites 1995) and "Arts sponsorship: Harmony or discord?" (Wolton 1988). While there is little formal conceptual development, construct definition or theory, they do contain a good deal of intuition about the processes involved. These papers are drawn upon heavily for examples of constructs proposed in the following chapters













CHAPTER TWO

SPONSORSHIP WITHIN A BROADER CATEGORY OF COMMERCIAL ASSOCIATIONS

As alluded to in the introduction, sponsorship is not unique as a commercial

endeavor which seeks to associate two things. The core idea behind brand extension, for example, is that a new brand can sometimes benefit from an association with an existing brand. Likewise for celebrity endorsers, an underlying assumption is that the brand can sometimes benefit from being associated with a human being. Other areas of study in marketing which involve commercial associations include co-branding (O'Connor, 1996; Carpenter, 1994), the use of one brand as an ingredient in another (Park, Jun and Shocker, 1996), "brand alliances" (Rao and Ruekert, 1994), and retail brand assortment (Laforet and Saunders, 1994; Bawa, Landwehr and Krishna, 1989; Bultez et al., 1989). There is even anecdotal evidence of the importance of customers as commercial associates in terms of providing prestige (Smith and Park, 1992).

The two areas which have received the most attention in terms of understanding the impact of such associations on consumers are the brand extension and celebrity endorsement literatures. While brand extensions and celebrity endorsers share with sponsorship issues related to association, they also differ in some conceptually important ways. The goal of this chapter is to fit sponsorship into a larger framework of commercial associations. Due to the different assumptions about the underlying mechanisms of brand




12





13


extension and celebrity endorsers, the issues addressed in this chapter are limited to the very basic, "common denominatoe, issues and predictions.

Since the goal is to set up a broad understanding of commonalties among

commercial associations, the level of precision in this chapter is predictably broad. The propositions put forth here, therefore, are also predictably broad. The specific details of how sponsorship is likely to function, as well as detailed predictions about main effects and moderators are left for later chapters.

In order to facilitate discussion of broad issues relating to commercial

associations, it is necessary to introduce a bit of vocabulary. In the discussion below, associate" can refer to either member of an association. For example, a brand can be an associate of a celebrity, and a celebrity can be an associate of a brand. "An associate of simply means that an object is associated with something else. "Acceptor," however, refers to the object which an association is attempting to help. In celebrity endorsement, the acceptor is usually the brand being associated with the celebrity. In a brand extension, the acceptor is the new brand which is being associated with the old one. The "provider," on the other hand, is the object or person providing the help. In celebrity endorsement, the celebrity is usually the provider. In a brand extension, the preexisting brand is the provider.

Perhaps the most basic premise, across domains of commercial associations, is that the more desirable the provider, the more beneficial the association will be to the acceptor. Aaker and Keller (1990), for example, find that in the brand extension domain the higher the quality of the core brand (the provider) the better the consumer acceptance of the extension. Sunde and Brodie (1993) obtain similar results. Likewise, in the celebrity





14


endorser domain, Kahle and Homer (1985) find that the better (more attractive) the provider (the endorser) the more the acceptor (product) is liked. This same effect has been found when the provider is a speaker, and the acceptor is an argument (Horria, Naccari and Fatoullah, 1974; Chaiken, 1979).

As stated above, there are many different mechanisms through which provider quality impacts acceptor quality. The specifics of these mechanisms are discussed in future chapters. In the most general terms, however, it can be proposed:


PI: The more highly regarded the event, the more highly a consumer will
evaluate the sponsor.

P2: The more well liked the event, the more well liked will be the sponsor.


An equally frequent finding in the literature in commercial associations is that the more qualitatively similar the two associates, the more benefit will come to the acceptor from the association. Aaker and Keller (1990) find this to be true in the brand extension domain. Beyond simple qualitative similarity, however, they define fit dimensions along which extended brands (acceptors) may match up with the original branded product (provider). Specifically, they find that if the acceptor is a complement of, or a substitute for, the provider, the acceptor will be better received. Similarity between provider and acceptor, however, can take place at a much higher level of abstraction. Park, Milberg and Lawson (199 1), for example, find that extensions work best when both core brand and extension are either prestige or functional in nature. Benefits to the acceptor of qualitative similarity with the provider have also been frequently demonstrated for





15

celebrity endorsers (Speck, Schumann and Thompson, 1988; Misra and Beatty, 1990; Frieden, 1984).

The details of how different kinds of similarity can impact sponsorship receive detailed coverage in chapter five. By adhering to the goal of fitting sponsorship into the broader category of commercial associations, however, the above findings on the impact of similarity between providers and acceptors lead to the following general proposals about sponsorship:


P3: The more a sponsor is like an event, the more positively consumers
evaluate the sponsor.

PC The more a sponsor is like an event, the more positive affect is
generated about the sponsor through the association.


A third, extremely common finding in the commercial association literature is the interaction between quality of provider and associate similarity. In general, it has been found that the fit between associates interacts with the positive relationship between the quality of the provider and the evaluation of the acceptor. Aaker and Keller (1993) for example, conclude that high quality brands can be stretched farther and extended into more dissimilar categories" (p.56). Broniarczyk and Alba (1994) "agree that, all things being equal, the best scenario for success involves extension of a highly regarded brand to a very similar product class." (pg. 215).

With regard to brand extensions, Keller and Aaker (1992) state that "when similarity is very low and the extension seems to be very unrelated to the core brand, extensions will not be evaluated favorably regardless of the nature of the core brand associations because those associations will not be seen as relevant to judging the





16

extension ... That is, the perceived quality of the core brand influenced brand extension evaluation only when there was some basis of fit between the core brand and proposed extension products" (p. 37).

Sheinin and Schmitt (1994) find complex interactions between brand affect,

category breadth, and attribute congruity between the original and the extended brand. Sunde and Brodie (1993) find that consumer acceptance of a brand extension will be higher if the original brand is of high quality and, there is a perceived fit between core brand and extension.

McCracken (1989), in the context of celebrity endorsers, points out that the ability of celebrity endorsers to act as providers for a brand comes from the cultural meaning with which they are endowed. Iiis argument is that if an endorser does not fit with a brand, there will be no overlap in cultural meaning and, therefore, no positive impact upon the acceptor.

Perhaps the best example of the interaction between quality of provider and

associate similarity in the domain of endorsers is found by Friedman and Friedman (1976). They investigated whether the effectiveness of endorser-type is dependent on the type of product being endorsed. They find a significant product by endorser-type interaction such that ads for jewelry were evaluated highest when done by a celebrity, whereas ads for cookies were better received when done by a homemaker. Thus, the study seems to show the importance of the fit between endorser and product type.

The common finding of an interaction between provider quality and associate similarity leads to the following proposal for sponsorship:





17




P5: The impact of perceived quality of an event on evaluation of, and positive
affect toward, a sponsor depends upon the level of fit between sponsor and
event. High fit facilitates the impact of event quality, whereas poor fit
impedes it.


The commercial association literature also provides evidence of many moderators of the impact a commercial association may have. Examples include previous associations, arousal, symbolic value, type of good, and priming. Each of these topics is considered below.

In the case of previous associations, Walker, Langmeyer and Langmeyer (1992) find that previously unendorsed products (bath towels and VCRs) "pick up" more of the personality of the endorser than does a previously endorsed product (jeans). This suggests the possibility of the following propositions for sponsorship:


P16: Brands in product classes which have not previously engaged in
sponsorship form stronger associations in the minds of consumers than
brands in categories where sponsorship is common.

P?7: Brands which have not previously engaged in sponsorship form stronger
associations with events than brands engaged in multiple sponsorships.


Another moderator of commercial association effectiveness is arousal.

Sanbonmatsu and Kardes (1988) find that high arousal causes an increase in the effectiveness of using a celebrity vs. noncelebrity endorser. Pham (1993) replicates and extends this finding. From these results it is possible to develop the following propositions:





18


P8: Events which are highly arousing cause higher evaluation of more
positive affect for a sponsored brand than do equally evaluated events which
are less arousing.


Given the intrinsically more arousing nature of five events over televised ones,

P9: The differential effect proposed in P8 is greater for live spectators than
for people who watch an event on TV.


Another moderating factor of commercial associations is the symbolic value of the provider. In general, providers with greater symbolic value are more capable of having an impact on the acceptor through the association. This is the foundation for many of the ideas put forth by McCracken (1989) with regard to celebrity endorsers. In the context of brand extensions, Reddy, Holak, and Bhat (1994) propose that brands with high "symbolic value" will be more readily extended than others. Using statistical analysis on time series sales data, they find evidence that highly symbolic brands extended better than less symbolic brands. This is true using both expert ratings of symbolism and consumer ratings of symbolic value. From these findings in the general area of commercial associations, it can be argued that for sponsorship,


P10: The higher the symbolic value of an event for a consumer, the more
impact the association will have upon the evaluation of the sponsoring brand.


The type of good involved as the acceptor has also been shown to impact the outcome of an association. Smith and Park (1992) find that in general, brand strategy (use of extensions vs. not) accounts for only 4% of market share, whereas for goods with mostly experience attributes it accounts for 9% of the variance (p. 309). This increase in the impact of brand extension upon share for experience goods over all





19

goods indicates that experience goods may be more susceptible to the impact of commercial associations than are search goods. In the context of "brand alliances" Rao and Ruekert (1994) "view the search-experience dichotomy as a continuum" (p. 95). They give examples of how brand name ingredients in food products which are likely to affect taste are better "brand alliances" than are alliances involving products with mainly search attributes. These studies seem to indicate that experience goods may be more easily impacted by commercial associations than are search goods. Therefore it can be proposed:

P11: AD else being equal, sponsorship will be more effective for brands with
more experience attributes and fewer search attributes.

P12: Brands which are considered credence goods will be more greatly
impacted by sponsorship than will search or experience goods.


One final issue needs to be touched upon in this review of the commercial

association literature. That is the issue of the acceptor having a negative impact upon the provider. Aaker (1990) notes that risky associations may backfire and gives examples of several notable cases where a failed brand extension actually harmed the image of the core product. Dawar and Anderson (1994) propose and find evidence that 44an extension to a particular product may make it more difficult for the brand to extend in a different direction in the future" (p. 124). Loken and Roedder John (1993) explain how a dilution effect (harming the provider) occurs under a "bookkeeping model." This bookkeeping model dilution effect will occur for extensions which have attributes that are incompatible with the image of the core brand. They note that this can be avoided if





20


the association can be perceived as either an anomaly or a separate category. With regard to sponsorship, therefore, it can be proposed:

P13: Effects of an event/brand association can impact the event as weH as
the brand.



The goal of this chapter was to provide a conceptual framework for commercial associations. The similarities in findings from celebrity endorsement, brand extension and other commercial association literatures were presented. Likely implications of these common findings for sponsorship were given as formal propositions. Overall, this chapter has demonstrated the ways in which sponsorship is like other commercial associations. The unique aspects of sponsorship and their likely impact on brand equity and positioning are provided in the following chapter












CHAPTER TBREE

THE POTENTIAL RAPACT OF SPONSORSHIP ON BRAND EQUITY AND POSITIONING.


As stated in the introduction, with sponsorship, the "association is the message." There is no argument with which to measure agreement, no claims, attributes or benefits on which to measure recall. It seems appropriate, therefore, to examine marketing variables which may be affected by sponsorship and which do not rely entirely upon formal argument. Two such variables are brand equity and brand positioning. The goal of this chapter, therefore, is to investigate what sponsorship can do. It does not deal with how it does it nor when it will work best. These issues are addressed in subsequent chapters.

While the complex associative mechanisms through which sponsorship may impact brand equity have received little attention to date, the act of associating a brand with "a socially intrusive activity possessed of its own personality in the eyes of the receiving audience" (Meenhaghan, 1991b, p. 8) should certainly be able to impact the things one associates with a brand. This is relevant because several authors have conceived of brand equity as consisting largely of brand associations.

For instance, Keller (1993) describes what he calls "customer-based brand equity." He defines this as "the differential effect of brand knowledge on consumer response to the marketing of the brand" (p. 1). The keys to this effect, according to Keller, are familiarity with the brand and holding favorable, strong, and unique brand associations in memory




21





22


He goes on to propose that marketing activity can enhance or maintain the favorability, strength, and uniqueness of a brand. Given an association-based definition of brand equity:


P14: Significant sponsorship activities are capable of impacting a brand's
equity.


In his book on managing brand equity, Aaker (199 1) devotes a good deal of time to explaining the importance of associations to brand equity. With regard to this importance, he suggests that "the association not only exists but has a level of strength. A fink to a brand will be stronger when it is based on many experiences or exposures rather than few. It will also be stronger when it is supported by a network of other links" (p. 109). Clearly, then, by Aaker's definition as well, sponsorship associations should sometimes be able to affect brand equity.

It should be noted, however, that Aaker has not endorsed the ability of

sponsorship to impact brand equity through building favorable associations. In fact, he states that "the primary role of most event sponsorship is to create or maintain awareness" (p. 75). Given, therefore, that sponsorship receives but a single line in the book, it does not seem likely that Aaker had fully considered the ways in which sponsorship might impact upon brand equity in great detail.

Other definitions of brand equity take a slightly different perspective. Simon and Sullivan (1993), for example, have proposed a model of brand equity based upon market valuation of the stock equities of corporations with substantial brand assets. Similarly, Farquhar (1989) defines brand equity as "the 'value added' with which a given brand





23


endows a product"(p. 24). He notes that Coopers and Lybrand calculate brand equity by comparing the premium price paid for a branded product compared to a generic one. It should be noted that while these definitions focus on the economic value which comes from brand equity, they do not preclude an associative component in their formation.

In fact Farquhar et al. (1992) propose ways in which the economic value of brands can be maximized and exploited through management of associations. They develop the concept of "master brands" which are very exploitable. They define these "master brands" as those brands which "own an association." That is, when the associate is mentioned, the brand comes to mind. While the associates used in most of the examples are product categories, they acknowledge that "other types of brand associations include celebrity endorsers, sponsored events, geographic affiliations, user groups, etc." (pp. 32-33, emphasis added).

In addition to specifically including the management of commercial associations as an important component of brand equity management, they provide insight into methods of measurement. They suggest that associations can be determined through free associative tasks, projective methods, and focus groups. They also claim that the process of building brand equity consists of developing favorable associations, making these associations accessible, and developing "points of Merence" which can be easily remembered by consumers.

These ideas have clear implications for sponsorship. In addition to the awareness building contribution of sponsorship to brand equity, it seems feasible that more complex associative mechanisms might be able to play a role as well. It should be possible to develop functional forms for the impact of sponsorship on brand equity. These might





24


measure or manipulate magnitude, accessibility and relevance of associations between brand and event, and use these variables to predict impact on brand equity in a variety of situations. Specific possibilities are explored in later chapters.

Another critical marketing goal upon which sponsorship may have an impact is positioning. One anecdotal account of a company using sponsorship to reposition their brand is Tanqueray gin. Tanqueray decided to sponsor the "California AIDS Ride," a bicycling event which raises money for AIDS (New York Times, July 24,1995). Tanqueray was in the process of trying to reposition the brand away from being a classic martini component for the 50+ crowd, toward being a more versatile beverage for the 2540 crowd. They saw sponsorship of the AIDS Ride as a bold way to associate themselves with an issue of great importance to their new target market.

Several authors have described brand positioning mechanisms which may be

susceptible to the influence of sponsorship associations. Blackston (1995), for example, conceives of repositioning as a change in brand meaning. He asserts that because value depends on the meaning, changing brand meaning is equivalent to changing the value of the brand. He proposes a brand relationship model where the objective brand is distinguished from the subjective brand. He defines the "objective brand" as consisting of a set of associations, images and personality characteristics. Within this framework, changing the set of associations should change the meaning of the brand and, therefore, change its positioning. In cases where associating a brand with an event leads to a change in the mix of brand associates, sponsorship should be able to aid in the repositioning of a brand.





25


Park, Jaworski, and MacInnis (1986) have conceptualized positioning as

management of "brand image." They present a framework for managing brand image at Merent times in a brand's fife cycle. They define brand image as "the understanding consumers derive from the total set of brand-related activities engaged in by the firm" (p. 13 5). They stress the importance of various brand associations on brand image, stating that "each element of the (marketing) mix can affect the inferences consumers develop about the brand" (p. 13 8). Therefore,


P15: Whenever sponsorship of an event causes a change in the associations a
consumer has with the brand, brand image changes, and positioning within
multidimensional space is altered.


Crawford (1985) proposes that the primary function of product positioning is to Merentiate among products. He defines a product positioning typology in which three things, product features, benefits, and surrogates, can be used for positioning. Of relevance to sponsorship is the "surrogates" positioning strategy. He lists "coming from respected source" or "endorsed by a respected personality" as tactical means to employ a surrogates positioning strategy. While sponsorship is not mentioned per se, it seems reasonable that an association with a respected event or organization should also aid one's surrogate positioning strategy. Therefore,



IP16: To the extent that a sponsorship represents a unique association, it is
exploitable as a differentiating positioning strategy.


Finally, DiMingo (1988) points out that product positioning is a process of

distinguishing one's company or product from competitors along real dimensions which





26


are important to customers. By his definition, there is no "idear' position for all brands. Positioning must be based upon advantages of the brand. He contends that these advantages should be communicated through a consistent, integrated communications strategy involving brand name, packagmg, and advertising. While he does not specify a role for sponsorship, the idea that positioning should be used primarily for Merentiation from competition suggests that to the extent to which a sponsorship represents a point of differentiation from competitors, it will be an effective positioning tactic.

This can be formally stated as:


P17: Sponsorship will have a greater impact on perceived brand positioning
when there is less competitor sponsorship of similar events.




This chapter has examined the marketing variables which may be affected by sponsorship. Specific proposals about the impact of sponsorship on brand equity and positioning were presented. The mechanisms through which sponsorship can effect these variables, however, was not addressed. This issue is covered in the following chapter.
















CHAPTERFOUR

SEVEN NIECHANISMS THROUGH WHICH SPONSORSHUP MAY WORK



Almost no conceptual or empirical studies have, thus far, addressed "how

sponsorship works." It is the contention of this dissertation that at least seven distinct mechanisms exist through which sponsorship may function. As proposed in the previous chapter, these mechanisms may impact brand equity and brand positioning. They may also impact other relevant marketing variables of interest. These mechanisms are not hypothesized to be exclusive of one another. Two or more of them may function in any given sponsorship association. Together, however, they are proposed to represent a reasonably complete set of "the ways sponsorship works."

The simplest way to order these mechanisms is by the amount of cognitive elaboration which is required by a consumer for the mechanism to function. In approximate order of necessary cognitive elaboration, they are simple awareness, affect transfer, image transfer, affiliation, implied size, implied endorsement, and reciprocity. The relationship of the mechanisms depicted in Figure I below.








27





28


Increase in required cognitive elaboration



( simple affect image 1~r imlied implied
awarenes tansfer transfer sizeaio enosee reciprocity


FIGURE 1

RELATIONSHIIP BETWEEN THE SEVEN MECHANISMS



The only of these mechanisms to receive significant direct acknowledgment are simple awareness, affect transfer and, to a lesser extent, image transfer. Evidence for the existence of the affiliation, implied size, implied endorsement, and reciprocity mechanisms is largely intuitive and/or anecdotal. The seven mechanisms are described below. Propositions related to the main effects of these mechanisms are also provided. Potential moderators of these mechanisms are provided in later chapters.



Simple Awareness

For who is going to buy a product or service from a company that one has never
heard of? Wright (1988, p. 104)


Awareness, as a mechanism through which sponsorship is likely to function, is widely accepted by practitioners and academics alike. A typical explanation of how this mechanism might function is as follows. A consumer goes to an event or watches it on TV and is "exposed" to the brand name. Since repeated exposure to an object has been found to lead to positive affect toward the object (Zajonc 1969), the consumer feels better about the brand. This good feeling then has all sorts of benefits to the brand, including,





29


but not limited to, greater attention paid to subsequent commercial communications (or less counterarguing, see Pracejus, 1993), and greater chance of inclusion in the consideration set. While there is little doubt that such things take place, to end here with awareness ignores the rich possibilities involved in the association. In short, if simple awareness were the only mechanism in operation, then it would, for example, make no difference whether a brand name was seen on the side of a race car or on the side of a bam.

As mentioned above, there is little doubt that awareness functions as an important mechanism in sponsorship. Examples of simple awareness, however, where there is little or no chance of cognitive elaboration are not easy to cite. One which comes close is the practice of soft drink bottlers paying for convenience store signage, provided it prominently displays the appropriate brand logo. This example meets the criterion for simple awareness because of the low likelihood of elaboration by a store patron as to the origins of the large "Coke" logo on the "ABC Beverage" sign. Even if there were some attribution on the part of a customer, it is unlikely that he would have enough involvement with the store to care that the owners received a new sign courtesy of the local Coke bottler. The most likely impact (if any) is that the customer may have high top of mind awareness of "Coke" as he faces his soft drink selection. There is little or no difference between the logo appearing on the sign or on the side of a nearby building. The only thing accomplished is simple awareness of the brand name.

Despite the relative simplicity of this mechanism, two basic propositions can be offered. Given the reasonably well established mere exposure effect,





30


P18: The more frequently a sponsored brand name is encountered by a
consumer in the context of an even4 the more positive feelings s/he has
toward the brand.

P19: The effect of P18 is greater for less familiar brands than for more
familiar brands.




Affect Transfer

Affect transfer in a sponsorship context is analogous to affect transfer in an

advertising context, whereby one's attitude toward the ad "transfers" to one's attitude toward the brand, with little cognitive mediation (MacKenzie, Lutz and Belch 1986; MacKenzie and Lutz 1989). As a mechanism through which sponsorship may function, affect transfer refers to positive feelings toward an event transferring to the sponsoring brand through the sponsorship association. This mechanism requires no real cognitive elaboration. It does not even require conscious awareness of the association, although such an awareness might be beneficial to the transfer taking place.

Affect transfer is a mechanism which has received some mention in the sponsorship literature, although much of it is rather vague. In a paper which reports significant differences in demographic and lifestyle variables among several Hallmark events (major city-wide destination events), Roslow, Nicholls and Laskey (1992) note that, "Hallmark events are attended by diverse audiences who have come together to enjoy a specific experience or entertainment. This heightened level of expectation may provide a more beneficial exposure content than traditional media (p. 54). They elaborate that, "It is probable that this form of promotion would benefit the advertiser from increased levels of excitement and interest associated with Hallmark events" (p. 58). These authors seem to





31


be proposing that the positive affective responses to the "Hallmark evenf' may transfer to the sponsoring brand.

Crimmins and Horn (1996) also address the idea of affect transfer. They discuss the transfer of positive feelings from the event to the brand in terms of Heider's (1946) "elementary human calculus." They posit that, given the relative weakness of many brand beliefs, the impact of associating a lowly regarded brand with a highly regarded event should usually lead to an upward estimation of the brand. While this seems plausible, the authors provide no data to support the assertion. The idea, however, is that the "positiveness" of the event can transfer to the sponsor.

Stipp and Schiavone (1996) attempt to demonstrate that attitude toward the Olympics can actually "rub off' onto Olympic sponsors. Through multiple regression, they find that attitude toward Olympic sponsorship in general has a significant impact, net of ad recall and ad evaluation, upon the sponsor's image. While this provides some support for the affect transfer mechanism, it should be noted that the survey methodology employed somewhat stacked the deck in favor of their finding.

Respondents were first asked a series of agree-disagree statements like "official Olympic sponsors provide money to: Help athletes train, Make it possible for the American teams to go to the Olympics, . ."(p. 24). Setting up the survey with these questions at the front end certainly had the potential to bias responses toward a strong association between attitude toward sponsorship and attitude toward a specific sponsor (a demand effect).

Indication of an affect transfer mechanism is also found in other commercial

association domains. In the context of celebrity endorsers, for exarnple, Misra and Beatty





32


(1990) found that the transfer of affect from spokesperson to brand was facilitated when the two were matched. In the domain of brand extensions, Rangaswamy, Burke and Oliva (1993) propose "affect transfer through categorization" as a method by which utility of the brand name impacts utility of the product.

Affect-transfer-like mechanisms are also found in the retail assortment domain. Jacoby and Mazursky (1984), in a factorial design, associate high and low quality brands with high and low quality stores. They attempt to determine how differential affect (i.e. positive affect toward one, negative affect for another) in the associate pair interact. Weak support is found for an averaging process.

Like simple awareness, the cognitively simple mechanism of affect transfer allows basic propositions to be made:



P20: The more positive the affective response to the an event, the more
positive the affective response to a sponsoring brand.

P21: The stronger the associative linkage between the brand and the event,
the stronger is the impact posited in P20.






Image transfer

The aim of sponsorship is image transfer: the transfer of the personality, aura
and competence ... to the brand Ryssel and Staniminger (198 8, p. 111)


While affect transfer is the process of assigning event-based affective response to the brand, image transfer refers to the assignment of abstract associations of the event to





33


the brand. Image transfer is the mechanism requiring the least cognitive elaboration in which the association can be formally stated as a message. The message of image transfer is that, along some dimension, "this brand is like this event."

The idea that the image, or personality of an event can impact upon the image or personality of a brand has been alluded to several times in the context of sponsorship. Parker (1991), for example, describes the relationship between sponsorship and brand image. Marshall and Cook (1992) found that 14% of respondents listed "image of sport" as the most important determinant in selecting an event to sponsor. Wolton (198 8) contends that art sponsorship is used to alter or modify the image of a company, lending a human face to cold, technologically-oriented companies, and providing a positive association with companies such as insurers which are usually associated with unpleasant situations.

McDonald (199 1) describes what he refers to as "product relevance" (p. 3 6) which can be either direct or indirect. Indirect relevance is proposed to imply association at an abstract, more image-based level. A strong and traditional bank sponsoring the traditional sport of football (soccer) is given as an example of indirect relevance, where the link is "tradition." Relating back to the discussion above, in this example the message would be that this bank is like the soccer league in that they are both traditional.

Perhaps the best anecdotal evidence of how the image transfer mechanism might be used for a specific marketing objective, like repositioning a brand, is given by Meenaghan (1991a). He states that sponsorships can achieve specific objectives through what he calls the "image by association effect"(p. 42). He gives as an example of how a "very American7' company, Gillette, was made to seem "more British" through its





34


sponsorship of cricket, a traditionally British sport. He also cites the case of IVECO trucks, which were viewed as weak European vehicles, compared to more macho American trucks. Through a sponsorship of the "macho" sport of heavyweight boxing, IVECO was able to shift this image in a highly successful way.

The image transfer mechanism is also found in other domains of commercial

association. In the context of celebrity endorsers McCracken (1989), points out that the cultural meaning with which celebrities are endowed is the root of their success in product endorsement. Clearly from his arguments, the image of the celebrity (and the meaning of this image within the culture) are key to an effective "meaning transfer" process. This process of transferring meaning from the celebrity to the product is more complex than explanations which consider attributes of a celebrity, such as credibility and attractiveness, as unidimensional, linear constructs.

In an empirical study of a transfer mechanism, Walker, Langmeyer and Langmeyer (1992) have subjects rate brands and endorsers along several abstract dimensions. They compare before association scores with after association scores along these dimensions to see how products "pick up" some of the personality of the endorser. They find that previously unendorsed products "pick up" more of the personality of the endorser than did a previously endorsed product This leads to the following testable proposition with regard to image transfer in sponsorship:



P22: Brands involved in few sponsorship activities acquire more of the image
and abstract associations of an event than brands which sponsor many
events.





35


As mentioned previously within the brand extension domain, Reddy, Holak, and Bhat (1994) find that brands with high "symbolic value" extend better than brands with low symbolic value. Given that in the case of sponsorship, the brand is the acceptor, one might expect that brands with high symbolic value already would be less likely to have the image of the event transferred onto them. Therefore it can be proposed:



P23: Transfer of image from event to brand takes place more easily, and
with greater magnitude, for low symbolic brands compared to high
symbolic brands.


Within the domain of combined brands Rao and Ruekert (1994) discuss how an averaging model of attribute levels would lead to a zero sum game in which brand alliances would be almost impossible to arrange. They argue in favor of a "maximum rule" whereby the "combined brand" may receive the higher of the two brand perceptions on each dimension. An example of this is the association between Lays potato chips and KC Masterpiece barbecue sauce. While the combined brand gains the desired associations of "tangy" and "hickory smoke flavor" from KC Masterpiece, an average "crispness" between the chips and the sauce is not perceived. Crispness remains at the higher of the two levels, the level of regular Lays. This leads to the following proposition for the domain of sponsorship:


P24: Due to a "maximum rule," brand images not fully congruent with an
event will not cause appreciable harm to the image of the event.


In fact, anecdotal evidence of reverse image transfer without negative impact upon the brand has been cited. Jones and Dearsley (1989 cf. McDonald, 1991) find that the





36

image of football (soccer) was actually improved by a sponsorship by Barclays Bank during a time when football was receiving a lot of bad press. Apparently this association transferred "stability" from Barclays to the soccer league without transferring "rowdiness" to Barclays. Perhaps positive associations are more easily transferred than negative ones. This is certainly an interesting empirical question for future study.



Afriliation

Affiliation, like image transfer, can be stated as a message. The message of the affiliation component of an association is that this brand "is for people like me." The primary requirement for the functioning of this mechanism is that the event be perceived as "for people like me." The more this statement is perceived to be true, the more the message about the brand being "for me" will be facilitated.

While there is no formal conceptual development of the way affiliation might function as a mechanism of sponsorship effectiveness, some authors have hinted at the idea. Meenaghan (199 lb), for example, points out that "the (sponsorship) message is delivered by association with a socially intrusive activity" (p. 8). Parker (199 1) posits that today's consumers, in addition to demanding products with desired attributes at a reasonable price, also seek products which are "for people like me." Sponsorship, he proposes, is uniquely suited to helping a brand achieve the perception of being "for people like me." No further development of these ideas, however, is presented by either of these authors.

There have been several studies of the impact of strong, social affiliations such as political affiliations (Ellen, Wiener and Cobb-Walgren, 199 1) and ethnic affiliations





37


(Deshpande, Hoyer and Donthu, 1986) on consumer behavior. At a broader level, Westbrook and Black (1985) find that "desires for affiliations with others" is an important determinants of some people's shopping behavior. In addition to these general impacts of affiliation with a group on consumer behavior, the possibility that an affiliation with a brand or organization can be beneficially developed through the use of marketing variables has been explored.

Macchiette and Abhijit (1992) describe "affinity marketing" as the use of group affiliation to produce a strong promotional program. The resultant "affinity groups" are said to have high levels of social bonding and cohesiveness. Members of affinity groups are also said to be are committed to the objectives of the group. The authors propose "marketing generated affinity" as one source of group centeredness.

An example of how marketing can take advantage of people's existing affiliations is the "affinity credit card." Such cards bear the name and symbol of an organization to which the card holder belongs. High levels of card holder loyalty have been claimed for these cards (Worthington and Horn, 1992).

Bhattacharya, Rao and Glynn (1995) investigate the correlates of social

identification with an organization. By defining a museum as the product and members as customers, the study investigates the determinants of affiliation with the museum. They find that perceived prestige is positively correlated with one's affiliation with the organization. Visiting frequency is also found to predict affiliation. Finally, participation in similar organizations is negatively correlated with affiliation. Based upon these findings, it can be proposed that





38


P25: Affiliation with an event is positively correlated with perceived
prestige of the event.

P26: Affiliation with an event is positively correlated with frequency of
attendance.

P27: Affiliation with an event is negatively correlated with one's
participation in other activities.


In a related area, the manipulation of "affiliation motivation" has been cited as a frequently employed advertising tactic (Zinkhan, Hong and Lawson, 1990). These authors content analyzed a sample of magazine ads from selected years between 1935 and 1985. They found that. over the time period studied, affiliation motivational appeals increased in frequency. The authors contrast this increase with a decline in the use of "achievement appeals."



Implied size

The implied size mechanism also has a message which can be stated in words. The message of implied size is that "this company is big." The cognitive elaboration required for this mechanism to occur is considerably higher that those previously listed. A consumer has to, at some level, think to himself "this must be a big company to have this event named after it."

The idea that sponsoring an event can influence the perceived size of the

sponsor has been examined and has received some empirical support. McDonald (199 1) explored this possibility, but did not find evidence of an effect. He states, however, that the reason sponsorship did not appear to affect perceptions of size was that most of the companies studied were already clearly perceived to be very big and important (e.g.,





39


Phillips and Cadbury). He posits that had the sample included more smaller companies, an impact of sponsorship on perceived size might have been found.

Evidence that sponsorship can influence perceived size was found by Rajaretnam (1994). In the evaluation of the "unique experiment" where a tire maker spent almost its entire marketing budget from 1984-1987 on sponsorship, perceived size of company increased by 19.7% during the period. Perceived financial health of the company, a related measure, increased by 24.6% (p. 68). While this field experiment with no control group does not prove that sponsorship can impact the perceived size of the firm, it certainly allows for the following proposals:



P28: Event sponsorship influences the perceived size of a sponsoring firm.

P29: The larger the perceived size of the even4 the greater is the impact of
the sponsorship on the perceived size of the firm.


Anecdotal evidence of the size of the sponsor affecting the perceived size of the event have also been reported. Parker (199 1) proposes that Mars' sponsorship of the London Marathon was the key to getting people to believe that it was a "rear' event. The following can therefore be proposed:



P30: Sponsorship of an event by a large sponsor increases the perceived size
of the event.


The marketing literature provides evidence of the benefits of perceived size.

Kirmani and Wright (1989) propose three possible mechanisms through which perceived expenditure on advertising might lead to inferred quality. One is the notion that





40

confidence in quality leads companies to spend more on promoting a product. Another is a perceived relationship between promotional spending and quality in some markets. Finally, they cite the idea that "perceived advertising expense is a signal of a firm's financial strength, probable social acceptance or some other factor that defines quality in some market' (p. 344). This last mechanism is closest to the idea of perceived size. The idea is that a company couldn't afford to sponsor a prominent event unless it is a large, stable company. The authors do not investigate the perceived financial strength possibility, however, stating that the "default attribution" is the "most interesting possibility and the one we pursue most directly in this research" (p. 346).

Following up this work, Kirmani (1990) proposes that perceived costs will be used as cues to quality when they are more diagnostic than other cues. She hypothesized that this should take place under low involvement with the product class. The hypothesis, however, was not supported. No effect of involvement was found.

i and Wright (1989) also list several "undermines" to the positive

relationship between perceived advertising expenditure and perceived quality. They describe the "desperation undermine' as occurring when expenditures are so high that the consumer assumes that the company is desperate to move a bad product. The "immunity undermine" occurs when the consumer perceives that the company has nothing to lose, as is the case with a fly-by-night operation. The "no pain undermine' is said to occur when a high expenditure is perceived to be a small part of a company's budget. Finally, they present the "basic premise undermine' which occurs when a consumer is confronted with information which argues against the "default attribution."





41

While both Kirmani (1990) and Kirmani and Wright (1989) report evidence of an inverted "U' relationship between perceived advertising cost and perceived quality, whereby beyond a certain point, quality perceptions actually go down. The data reported in both cases could, however, simply reflect diminishing returns. Asymptotic functions could have produced the significant quadratic terms reported. Also, no means or least squared means were reported which might have shown the true shape of the expenditure/quality relationship.

While it is not difficult to imagine situations under which perception of firm size would have an impact on perceived quality, very little research has directly investigated this effect. Within the context of the banldng industry there is evidence that perceived size of a bank can affect consumer beliefs about the likelihood of receiving a loan (Leonard and Spencer, 199 1). In the arena of legal services, however, there is evidence to suggest that service providers believe perceived firm size to be more important than do clients (Gaedeke and Tootelian, 1988).

Given the results of this research, it can be proposed that


P31: The sponsorship impact on perceived size has a positive impact on
perceived quality of the brand.





Implied Endorsement

The implied endorsement mechanism requires the most cognitive elaboration on the part of the consumer, save reciprocity. The message of the implied endorsement mechanism is "this brand must be 01C, or people wouldn't participate in an event named





42


after it." The implied endorsement mechanism requires that, from a sponsorship association, the consumer infers that the event is somehow endorsing the quality of the brand. Several examples of this can be cited. Seiko's sponsorship of track meets at which it is the official timer is provided by McDonald (199 1) as an example of what he calls "direct relevance"(p. 36). By direct relevance is meant that one might logically conclude that the official status of a sponsor implies some measure of quality.

Crimmins and Horn (1996) cite Visa and Seiko as brands which actually affected consumer perceptions of quality due to their association with the Olympics (as measured by DDB Needham's "SponsorWatch" service). Since Seiko was the official timekeeper and Visa was the only credit card the Olympics would accept for tickets, it seems that an implied endorsement may have led to the Olympic sponsorship success of these two brands.

For the implied endorsement mechanism to function, however, does not require

some official sanctioning of the sponsoring brand. As long as a consumer infers that some minimum quality level is required before an event will let a brand sponsor it, the implied endorsement mechanism will be engaged. Usage of the brand at an event, however, should increase the likelihood of this type of inference. Seeing top pool players compete for large prizes on Brunswick tables, at the "Brunswick Shoot Out" for example, might cause some consumers to infer a level of quality for Brunswick tables.

It can therefore be proposed:

P32: Use of the brand in the sponsored event facilitates the inferences
necessary for the implied endorsement to function.

P33: Having the brand name in the event name facilitates the inferences
necessary for the implied endorsement to function.





43


The ways in which implied endorsement can impact perceptions, attitudes, and choice, have been well delineated in the source effects literature. Factors affecting the ability of a source to persuade include trustworthiness (Brehm and Lipsher, 1959), expertise (Petty, Cacioppo and Goldman, 1981), attractiveness (Landy and Sigall, 1974), and credibility (Birnbaum, Wong and Wong, 1976). While the literature on source effects is rich with predictive constructs, proposing the impact of these constructs in the context of sponsorship first requires some empirical investigation. For example, it is not known whether a credible or trustworthy event exists in the minds of consumers. If such things do exist, it is not obvious what the determinants of these source characteristics would be. Further elaboration on the role of source characteristics in sponsorship, therefore, requires prior empirical investigation.



Reciprocity

Reciprocity is the mechanism which requires the highest level of cognitive

elaboration. The message of reciprocity is "the sponsor supports events you care about, so you should patronize the sponsor." The reciprocity mechanism is engaged when a consumer makes a conscious decision to go out of her way to support the brands that support the events which she cares about. While this may not be the most common response to sponsorship, it is certainly a powerful one.

Surprisingly, there is evidence to support that fans of certain events demonstrate significant reciprocity (or so they claim). Crimmins and Horn (1996) present what they call "gratitude" whereby the fan of an event goes out of his way to buy the products of an event sponsor. They cite a survey of NASCAR fans of whom 48% say they would





44

"almost always" purchase a sponsor's product over a closely priced competitor. Reported reciprocity for the Chicago Gospel Festival was even more impressive. As measured by "it makes me more likely to buy," fan reciprocity here was 82%. The authors go on to point out that it is unlikely to get these effects from casual observers or uninvolved spectators. Simply stated, they claim that 'Tans are grateful for sponsorship. Others are nof' (p. 17). From these observations, it is possible to propose:



P34: The more involved consumers are with an event, the more likely the
sponsorship is to engage reciprocity.




Conclusion

Any attempt to study sponsorship must consider the multiple ways that an

association of a brand with an event can impact consumers. Seven mechanisms and their psychological underpinnings were proposed and discussed in this chapter By breaking sponsorship into these component mechanisms, it becomes possible to posit and test how various aspects of a sponsorship association will impact success. Since there are multiple ways in which sponsorship may work, considering which mechanisms are likely to be engaged under various types of sponsorship associations allows many specific hypotheses to be developed and tested. Empirical testing of numerous hypotheses relating to how audience, event, and associational factors impact the functioning of these mechanisms is presented in chapters eight, nine and ten. A thorough exploration of one important associational factor, fit, is presented in the following chapter.















CHAPTER FIVE

DIMENSIONS OF FIT.


The better the fit of all elements, the better the chance that the sponsorship is
effective, because the relations between the sponsor and sponsored activity are
logical, natural, and easy to understand Otker (1988, p. 83)


The notion of "fit" is seemingly ubiquitous in the extant sponsorship

literature. (Roslow, and Laskey, 1992; Thomas, 198 5; Crinmmins and Horn, 1996; Otker, 1988; Waishe and Wilkinson, 1994). This chapter attempts to define fit between an event and a brand. "Generic" or respondent-reported perceived fit requires no development or investigation of elaborate antecedents and component structures. It is simply a measure of whatever a respondent thinks it is. Operationally, generic fit is measured by a subject's responses to a question like "how good is the fit between event X and brand Y." While this measure may be a bit crude, the importance of quickly getting a handle on the fit construct makes it useful.

A brief review of how the fit construct has been presented in sponsorship and other commercial association domains is reviewed. Other, more technically defined types of fit, such as "attribute fit," "audience fit" and "positioning fit" are then discussed and defined. The chapter concludes with a discussion of measurement issues.

Most sponsorship research refers to fit as the degree to which an association "makes sense." Walshe and Wilkinson (1994), for example, report that IIBM's



45





46

sponsorship of a Pompeii museum exhibit was received much better than corporate sponsorship of the arts had typically been received up to that point. Because IBM was able to use its computers to bring a computerized archive of the dig into the museum, the attendees not only saw a fit between the exhibit and IBK they actually were able to perceive a value added to their experience from the association. Negative reactions to the association (measured by polling attendees) were, as a result of the association making sense, much lower than negatives toward other similar art sponsorships.

Within the context of evaluating sponsorships, McDonald (199 1) argues that "synergy" between the brand and the event is essential. He proposes that assessing synergy would be useffil, as opposed to doing simple favorability tracking. He does not, however, specifically define synergy nor does he propose how to, measure it.

The idea of "fif' is also found throughout the rest of the commercial associations literature. In the domain of brand extensions, Aaker and Keller (1993) define various types of "fit." They state that "it is clear that the interactive relationship between "quality" and the fit variables is sensitive to the levels of perceived quality and fit (in addition, possibly, to different types of fity' (p. 56). These authors operationally defined fit to be the degree to which the brand extension is a substitute or a complement for the core brand. Such an operational deftition, however, cannot be applied directly to the domain of sponsorship.

Broniarczyk and Alba (1994) reassert the importance of fit in brand extensions,

stating "an assumption common to virtually all research, however, is that brand affect and product category similarity play important roles." (p. 214, emphasis added). They continue, "we agree that, all things being equal,:the best scenario for success involves extension of a highly regarded brand to a very similar product class" (pg. 215, emphasis





47

added). They then qualify these statements by explaining that this is not news to brand managers, and that it misses the obvious differences in the "brand specific associations" which are not common to all brands within a category. While they go into some detail on why brand specific associations are an important consideration, they do not develop a method for determining the brand specific associations for each brand, and assessing their strength or importance.

Rather, they use pretests. to identify brands with at least one strong, unique

association. They then test to see whether these brands are more successful at entering product classes for which their unique associations are relevant than are brands which do not possess these unique associations. While the authors make a significant point about the unique associations possessed by brands (as opposed to categories), their method of seeking out brands with strong, unique associates still does not answer the question "what is fit?." It does, however, suggest that if both the provider and the acceptor posses the same strong and unique association, fit will be relatively high.

While specific antecedents of different kinds of fit are discussed below, the

implications for sponsorship of the Broniarczyk and Alba (1994) study can be formally stated as:

P35: Perception of generic "it" by consumers will be high when the event
and brand share unique, salient attributes.


Within the context of "composite branding alliances," Park Jun and Shocker (1996) propose structural determinants of fit between two brands. Specifically, they propose that when the "header brand" (first brand in composite name) and the "modifier brand" (the brand which follows "by" in the name) have complementary attributes, and are





48

not seen as logically incompatible, the resultant composite brand will have a good likelihood of success. Key to their definition of "good fit" are complementary attributes which are not logically inconsistent. In other words, sharing attributes is not necessary. Here they argue that so long as the attributes are logically consistent, non-overlapping attributes help the combination. Implications of this idea are discussed in detail in the "attribute fit" section.

Fit has also received attention in the celebrity endorsement literature. While several empirical outcomes of fit have been documented, little attention has been devoted within this domain to defining or measuring things which lead to fit. Kamins; and Gupta (1994), for example, found that increased congruence between the spokesperson and the product resulted in a more favorable product attitude. Speck, Schumann and Thompson (1988) found that when the endorser's relationship to the product is more obvious, information can be processed in a shorter time.

Ryssel and Stamminger (1988) report that Adidas has engaged in some interesting measurements of the perceived personalities of various tennis stars, and how these measures fit with the Adidas personality. They do not report in this paper whether or not they measure brand personality along similar scales to those used to evaluate the celebrity, but the article gives the impression that the company uses some sort of mechanism for comparing celebrity personality with brand personality for fit. This points out the interesting possibility that fit can include more abstract attributes, like personality characteristics.

Despite its frequency of mention in the above literatures, there has been almost no serious attempt to systematically define or measure the fit construct. There is little





49

detailed consideration of its potential antecedents, properties and consequences. There are few details on how one might measure it. The rest of this chapter represents a first step toward remedying this situation.

There are at least three kinds of (or approaches to) fit. While the three are likely to be highly correlated in many situations, there are reasons to believe that they have different impacts on the seven mechanisms discussed in the previous chapter. The three types are "attribute fit," "audience fit," and "positioning fit." This chapter first describes attribute fit in terms of various psychological constructs, such as similarity (Tversky, 1977; Tversky Gati, 1982) and categorization (Mervis and Rosh, 1982; Rosch, 1976) which give insight into possible antecedents, properties and consequences of attribute fit, as well as suggestions for, its measurement. Audience fit and positioning fit are then presented. Issues relevant to their special impact on several of the seven mechanisms are discussed, as are issues related to their measurement. Specific predictions on how these types of fit impact the seven mechanisms, however, are left for a future chapter.




Attribute Fit

Although similarity is recognized as fundamental to marketing and marketing
strategy, little attention has been paid to the way similarity judgments are
produced Johnson (1986, p. 59)


Attribute fit is the most general type of fit. It is not specific to sponsorship in any way. All of the antecedents, properties and consequences are, in theory, as applicable to any other form of commercial association as they are to sponsorship. Attribute fit is defined as the degree to which two associates share common attributes, and do not





50


possess incongruous attributes. By attributes are meant associations -with the brand or event which are commonly elicited in the minds of consumers when thinking about the event or brand. These associations can vary along a concrete-abstract continuum. An example of a concrete attribute shared by sponsor and event is "automotive." "Speed" is a more abstract association, and "macho" is more abstract still.

This definition closely resembles Tversky's (1977) definition of similarity, which increases with the addition of common features or the deletion of distinctive features. More will be said of the relationship between "similarity" and "attribute fit." First, however, possible antecedents of attribute fit are discussed.

One thing which should influence attribute fit is the frequency and recency with which two things are experienced together. Barsalou (1982), for example, investigated the ways in which context-dependent properties of an object might be made to be more context-independent. In other words, he explored how situation-dependent associates of an object can be generated more spontaneously. He found that repeated simultaneous exposure to the context-dependent property of an object can make the property more context-independent.

An example of a context dependent property which can be made to be more

context independent is the ability of a basketball to float (Barsalou 1982). Most people, when asked to fists things associated with basketball would not fist its ability to float. People who had frequently and or recently seen a basketball floating in a pool, however, would be much more likely to come up with this association. This leads to the following proposition:





51


P36: When attributes which overlap between event and brand are context
dependent, exposing consumers to the relevant context increases attribute rit.



Sponsorship associations are likely to be a context-dependent "property" initially, but may, over time become context-independent. Initially, for example, Miner beer may be a context-dependent property of NASCAR. At this initial stage, one would expect Miller to cue NASCAR only when the sponsor is also cued, perhaps while one is watching a race. As the association becomes stronger (as might be the case for a serious fan who is spontaneously aware that Miller sponsors his favorite driver), Miller might become a context independent property of NASCAR. In this latter case, one of the spontaneous associations that comes to mind when this person free-associates about Miller beer would be NASCAR.

It therefore seems possible that



P37: High involvement with an event leads to greater spontaneous awareness
for the brand-event association.

P38: Frequent attendance at an event leads to greater spontaneous awareness
for the brand-event association.

P39: Recent attendance at an event leads to greater spontaneous awareness
for the brand-event association.


In another study, Barsalou (1985) found that central tendency did not predict

graded structure in goal-derived categories. Goal-derived categories are categories which are formed for a specific purpose, like assembling things to sell at a garage sale. From this example, it is clear that goal derived categories are often categories which are formed in





52

order to make sense out of a group of objects which could not easily be separated by preexisting, taxonomic categories. Barsalou found that frequency of instantiation predicted graded structure in both goal derived and taxonomic categories to a large and approximately equal extent. This provides evidence that, to the extent to which seeing a brand paired with an event leads one to create an "ad hoc category" (Barsalou 1983), in order to make sense out of what is being seen, frequently and recently seeing this brandevent pairing should enhance attribute fit. Noting the continuous nature of this relationship, Barsalou explains that the "strength of association is free to vary continuously as a function of how frequently and recently an association has been active in working memory" (p. 212).

P40: Frequent attendance at an event leads to greater salience of common
attributes.

P41: Recent attendance at an event leads to greater salience of common
attributes.




Markman and Gentner (1993) argue that mental representations (which could include degree of fit between two objects) are formed through "structural alignment" which is derived from analogical reasoning. They find support for the proposition that "the likelihood of a structural alignment between two stimuli is greater after a similarity comparison than before" (p. 435). It seems, then, that consciously considering the similarity between two objects can shift attention away from the obvious, surface dimensions of fit, to the more relational dimensions of fit. Since conscious consideration of the similarity between the sponsor and the event is something which may be fiLcilitated





53

by communications, the details of this finding will be discussed in the "associationenhancing communications" chapter. If, however, one supposes that frequency of instantiation of the brand/event pairing might spontaneously lead people to "consider the similarity," then it may be possible that over several encounters with the pairing, not only is fit facilitated, but the more abstract, relational dimensions of fit might become more prominent. These ideas are formally stated:

P42: Frequent attendance at an event leads to greater increase in the salience
of abstract relational common attributes than in the salience of more
concrete common attributes.

P43: Recent attendance at an event leads to greater increase in the salience of
abstract relational common attributes than in the salience of more concrete
common attributes.


Within the broader research stream involving how people fit things into categories, Cohen and Basu (1987) suggest that "there is a real possibility... that the consumer environment favors category definition in terms of specific exemplars rather than category defining features..." (p. 470). This is in reference to a perceived uncertainty as to when categories will be defined by exemplars as opposed to being defined in a "probabilistic" fashion (see also Smith and Medin, 198 1). The critical issue here is that because brands are not incredibly important to most people, the mechanism associated with categorizing brands is not likely to involve capacity-depleting probabilistic rules. This seems to suggest that when it comes to the fit between a brand and an event, the large bulk of the fit will be determined by a perceived match between exemplary attributes, with less of a role for attribute matches which do not immediately or spontaneously come to mind.





54


P44: Given people's low motivation to form associated brands and events
into categories, attribute fit is determined primarily by overlap in chronically
available attributes of the two associates.



Having explored several potential antecedents of attribute fit, possible properties of attribute fit are now considered. As mentioned above, the attribute fit construct is quite close in definition to Tversky's "similarity." It therefore seems appropriate to consider other properties of similarity which make sense in the context of attribute fit. One seemingly important fact is that Tversky (1977) generally argues that similarity should not be treated as symmetrical. He proposes and demonstrates, for example, that North Korea is perceived to be more similar to China than China is to North Korea. This is an example of a more general idea, that "the variant is more similar to the prototype than the prototype is to the variant, because the prototype is generally more salient than the variant" (Tversky, 1977, p.333). This has several implications for the perceived fit between sponsor and event.



P45: When the brand is larger (more salient) than the event, rit is perceived
to be greater when the association is stated in the form of "the event fits with
the brand" than "the brand rits with the event"

P46: When the event is larger (more salient) than the brand, fit is perceived
to be greater when the association is stated in the form of "the brand fits
with the event" than "the event rits with the brand."



It should be noted that similar results have been obtained in a marketing context. Johnson (198 1) finds Shasta was judged to be more similar to Coke than Coke was to Shasta. Here again, asymmetries are demonstrated such that an object with many





55


distinctive features is judged as less similar to an object with fewer distinct features than is an object with few distinct features to an object with many. Put another way, a small (less salient) brand is perceived to be more similar to a big (more salient) brand than is the big brand to the small one.

Tversky (1977) also claims and demonstrates that at similar levels of salience,

similarity asymmetry is reduced, or nonexistent. An example of this case is a comparison of the similarity between Hungary and Czechoslovakia. Since neither country represented a "prototype," the perceived similarity of Hungary to Czechoslovakia was about the same as was the perceived similarity of Czechoslovakia to Hungary. The implications for the perceived fit between sponsor and event is:



P47: For sponsors and events of equivalent perceived size (salience) there is
little or no "fit asymmetry."


Another property Tversky (1977) finds for similarity is that it is not transitive. For example, while Cuba is similar to Jamaica (both Caribbean islands), and Cuba is similar to Russia (both Communist countries at the time) Jamaica is not perceived to be similar to Russia at all. While this seems incredibly intuitive in the context of countries, it may be less so in the context of the fit between sponsor and event.

For example, it seems entirely possible that a marketing manager might develop rough heuristics by which to evaluate potential sponsorship opportunities. One possible strategy might be to consider looking for other sponsors of events which his company currently sponsors, then examine which other events those companies are sponsoring. In fact, databases for We to companies involved in sponsoring are set up to, among other





56


things, do this very thing (e.g., "IEG Sponsordex7 software). While such a database might prove useful as a decision aid, it seems quite reasonable (and not unimportant). to propose that



P48: Attribute fit between sponsors and events is not transitive.


A final property of similarity proposed and demonstrated by Tversky is its relationship with "dissimilarity." Similarity has generally been found to be strongly negatively correlated with dissimilarity, but "people attend more to the common features in judgment of similarity than in judgment of difference" (Tversky, 1977, p. 340). As a result, prominent pairs (e.g., East Germany and West Germany) are found to be more similar than non-prominent pairs (e.g., Ceylon and Nepal) in a similarity condition and more dissimilar than non-prominent pairs in a "dissimilarity" condition. This effect has also been demonstrated in a marketing context. Johnson (198 1) finds that Coke and Pepsi are very similar in the context of similarity, and very dissimilar in a dissimilarity context. This suggests that

P49: When an event has many salient associations which are similar to associations of the brand as well as many salient associations which are
incongruent, the magnitude of similarity depends on the frame (fit vs. lackof-fit) such that the sponsorship association is perceived as having either high
fit or high lack-of-fit.


Before considering the consequences of attribute fit, it should be noted that

similarity has received some criticism for representing at best an intermediate measure for predicting category membership, and at worst a midpoint in a circular argument. Murphy and Medin (1985) point out that "any two entities can be arbitrarily similar or dissimilar by





57

changing the criterion of what counts as a relevant attribute." They go on to present the use of similarity to explain categorization as analogous to citing "scoring more points than the opposing team!' as a reason for victory in sports. The authors posit that rather than being a true determinant of conceptual coherence, similarity may be a mere "by-product" (p. 291). While the merits of this conclusion in the context of understanding category structure is debatable, the criticism seems less relevant to the issue at hand. As stated above, the process of evaluating fit between a sponsor and an event is likely to be only tangentially related to taxonomic category structure. For this reason, the usefulness of the similarity construct in the exploration of the properties of attribute fit seems to be reasonably whole.

In order to understand how attribute fit might moderate the seven mechanisms, it is first important to consider the direct consequences of attribute fit. As mentioned above, to the extent that sponsorship associations elicit an attempt on the part of a consumer to make sense of the association through a process of categorization, the categories formed are likely to be what Barsalou (1983) refers to as "ad hod" categories. He goes on to describe and demonstrate several outcomes of such categories. Specifically, it is demonstrated that ad hoc categories tend to be less established in memory, and therefore less available in memory. This means that category concepts, concept-to-category associations and category-to-concept associations will generally be less established in memory compared to those of common categories (Barsalou, 1983).

This seems to indicate that the weaker the attribute fit, the less the sponsorship associations will be available in memory. Ther6foire, it is possible to predict that





58


P50: The greater the attribute fit, the greater the recall for the association.

P51: The greater the attribute fit, the greater the recognition for the
association.

P52: The greater the attribute fit, the greater the recall for the event when
the brand is cued.

P53: The greater the attribute fit, the greater the recall for the brand when
the event is cued.

P54: The greater the attribute fit, the greater the recognition for the event
when the brand is cued.

P55: The greater the attribute fit, the greater the recognition for the brand
when the event is cued. '

Murphy and Medin (1985) propose another interesting consequence of attribute fit (or conceptual coherence). They propose that such fit can facilitate inference-making about the relationship. Specifically, they present the possibility that "coherent or useful categories are the ones that allow the most inferences to be made." They continue by proposing that "a vague category ... enables few if any inferences to be made"( p. 293). While they do not present conclusive evidence to support these assertions, it is certainly possible to imagine how strong attribute fit might facilitate the kind of inferences necessary for the functioning of mechanisms such as "implied size" and "implied endorsement." Therefore, it is tentatively proposed that



P56: Strength of attribute fit is positively correlated with the number of
inferences made about the association.



'It should be noted that the linear form assumed is only proposed, and that other research indicates other possible functional forms. The functional form of the relationship between attribute fit and memory, therefore, is an empirical question.





59

P57: Strength of association is positively correlated with the confidence in
one's inferences about an association.


Cohen and Basu (1987) reach some similar conclusions about categorization and inferences. They elaborate on the potential of affect to serve as an intermediate variable in the inference-making process. They propose that "since categorization produces a reduction in uncertainty, positive affect may result from a successful fit and negative affect from an inability to categorize an item particularly if the resulting judgments or inferences are important" (p. 470). They go on to say that "by relating the assignment of affect to more fundamental categorization processes we are also more likely to anticipate the type of inferences that consumers use to fill in 'information gaps' when direct product knowledge is absent or when motivation to seek such information is low" (Cohen and Basu 1987, p. 471).

These observations are relevant to sponsorship because they point out that the affect generated by successfully fitting things into a category may moderate the number and type of inferences generated. It can therefore be proposed that



P58: The affect generated by the association of a brand with an event moderates the relationship between attribute fit and inference making
proposed in IP56 and P57.


Ortony (1979) also provides some insight into the consequences of attribute fit. This work attempts to go "beyond literal similarity" to include what he refers to as metaphorical similarity. He defines many properties of metaphors, and relates these properties to Tversky's (1977) concept of similarity. Key to this discussion is the idea that





60


the more metaphorical is a comparison, the more asymmetry there will be (e.g., cigarettes are like buffets, but bullets are not like cigarettes). This is the idea which led to the quote .".the degree of symmetry is inversely related to the degree of metaphoricity"( p. 172).

Ortony suggests that when no attribute of the second term in a comparison seems to apply to the first, that "a speaker (or listener, reader or writer) may reorder the salience of the attributes of (especially) the second term in the comparison." (Ortony, 1979, p. 173). This may have relevance to sponsorship. If one equates the metaphoricity construct with the levels of abstraction of the association, then it is possible to make some interesting propositions regarding levels of abstraction and attribute salience.

For example, when Porsche cars are compared to race cars, the comparison is

quite literal, and the salient attribute of a race car (speed) can be easily linked to the brand Porsche. When McDonalds sponsors a race, however, the literal similarity is quite low, and the desirable "fast" attribute is less likely to transfer. When a full-size pickup truck sponsors a race car, however, the most salient attribute of the race car, "speed," is not relevant, and a reordering may take place until a match, such as "can take abuse" is found. If this takes place, not only should the toughness of the truck be made salient, but the toughness of the race cars should be made salient as well. Based upon Ortony's (1979) results, the following propositions can be made:

P59: When attribute rit between event and brand is the result of concrete attribute overlap, metaphorical elaboration does not enhance attribute fit.

P60: When attribute fit between event and brand is the result of abstract
attribute overlap, metaphorical elaboration enhances attribute fit.





61


Another consequence of attribute similarity may be a tendency to infer that one associate has the attributes of the other associate. In other words, attribute fit between sponsor and event along several attributes may lead to inferences that a brand possesses an additional attribute of the event, which the brand is not known to possess. There is evidence to support the idea that two things which are strongly associated can lead to inferences about shared attributes which are not known to be shared. Sedikides, Olsen and Reis (1993) investigate whether providing information about the relationship between two individuals (that they are married) would lead to greater spontaneous transference of attributes (e.g., "is vegetarian") during recall than for pairs of non-married individuals. They do, in fact, find that for married couples, more attributes of one spouse are mistakenly attributed to another spouse than is the case for non-married pairs of people (p<.05). The implications of these findings for sponsorship might be



P61: The tendency to "transfer attributes" from event to sponsor occurs
more frequently under high attribute fit than under low attribute fit.


In addition to investigating possible consequence of fit, it is necessary to consider possible consequence of lack-of-fit. Within the domain of investigating complex social conjunction categories, Hastie, Schroeder and Weber (199 1) cite evidence which suggests likely characteristics of attributions made about compound categories which are seemingly incongruent. In this paradigm, the authors compare conjunctive categories formed by pairing either congruent or incongruent attributes of a person. They find that attributions about the incongruent compounds are more extreme e along trait adjective scales than are either the original components of the pairing or the congruent compounds. They also find





62


that freely elicited attributes of the incongruent compound include many attributes which are not elicited by either simple category. They suggest that evaluation of incongruent combinations may be based upon a more "complex judgment process" than evaluation of simple categories.

Put another way, their results seem to indicate that when incongruent things are paired, evaluation of the pair is extreme, and attribute elicitation about the pair includes attributes which are not elicited by either of the paired items alone. Assuming the pairing of event with brand engages similar mechanisms,

P62: Lack of attribute fit between sponsor and event leads to more extreme
evaluation of the association along trait adjective scales.

P63: Lack of attribute fit between sponsor and event leads to the elicitation
of attributes which are not elicited for either the event or sponsor alone.

P64: Evaluation of high lack-of-attribute-fit sponsorships requires consumers
to engage in more complex judgment processes than do sponsorships which
have better attribute fit.



Finally, it is important to qualify all of the assumptions of a linear relationship

between attribute fit and consequences in light of findings which suggest more complex relationships. Meyers-Levy and Tybout (1989), for example, demonstrate that products that are somewhat congruent with their associated category schemes will stimulate processing that leads to a more favorable evaluation relative to products that are either congruent or extremely incongruent. In addition to the preexisting notion that very incongruent pairings are unnatural, weird, and/or poorly liked, this research suggests the possibility that extremely similar pairings are seen as mundane or borin& and therefore,





63

poorly evaluated. While the result obtained by these authors is not entirely intuitive, its existence suggests some possibilities for the impact of attribute fit in sponsorship.



P65: In some situations, extremely high attribute fit results in lower
evaluation of the association than does moderate fit.'


No discussion of a construct like attribute fit would be complete without an examination of possible methods of operational measurement. Several interesting possibilities are presented in the literature. Others are developed below.

Johnson (1986) performs a direct test of the contrast model (Tversky 1977) in a consumer product domain. Specifically, the study finds that similarity judgments are predicted by "a linear combination or contrast of the average number of common and distinctive features of the products being judged" (p. 56). This suggests that attribute fit might be measured by free elicitation of attributes (abstract and concrete) from independent groups of subjects for an event and a brand. The attribute fit measure could then be calculated as a function of the number of common attributes and unique attributes of each associate. Given the possibility that events may generally elicit more abstract, and fewer concrete attributes than brands, it may be necessary to normalize elicited attributes along an abstract-concrete continuum before calculating the attribute fit measure.

Similar techniques have been suggested within the domain of brand equity

evaluation. Farquhar and Herr (1992) point out the distinction between strong category to brand associations (category dominance) and brand to category associations (instance



'When this is the case remains an empirical question.





64


dominance). The authors also point out the implications of different levels and types of dominance in the domain of brand extension. They point out that many researchers either ignore directionality or look at the wrong associative direction. The authors also hint at other brand associates, such as usage situations, benefits and attributes, but do not discuss the implications of these associations in any detail.

They then discuss three methods for assessing category dominance. They first

describe naming methods, which involve a subject naming as many brands in a category as possible in some short time, or naming some pre-specified number of brands. Brands named, and especially named early-on are said to have high category dominance. Such methods could be applied to sponsorship by asking subjects to fist "events sponsored by X' or "brands sponsoring Y," using unaided recall as a surrogate for fit.

Farquhar and Heff (1992) also describe latency methods, which use reaction time measures to determine category dominance. According to the authors, brands which are quickly and correctly identified as members of the category have high category dominance. Here a slight modification of the measurement method might be appropriate for sponsorship. Cueing a consumer with an event, and measuring his/her reaction time and accuracy to whether or not a brand is a sponsor would be a fairly good method of measuring the "spontaneous availability" component of attribute fit. Reaction time measures of this type have been demonstrated to discriminate between spontaneously available properties and context dependent properties (Barsalou, 1982, experiment one).

Finally, the authors discuss facilitationn' methods, which involve providing

subjects with a category, then presenting them with a brand covered up with dots. The authors claim that brands which are recognized earlier are category dominant. In the





65


context of sponsorship, an event rather than a category could be cued. Brands for which the event cue resulted in "facilitation" of recognition could be said to have better attribute fit.

Although all these measures could be used as surrogates for attribute fit, it seems that all three techniques would be poor direct operationalizations of the attribute fit construct. Such measures would really describe how successfid a previous association had been. This might be strongly influenced by things like duration of the sponsorship, and amount of money spent on promoting awareness of the association. While these things have been proposed to impact attribute fit, they are not truly components of attribute fit.

As mentioned at the beginning of the chapter, one measure against which others could be compared would be some sort of respondent-defined, generic fit measure. Direct measures of generic fit could be measured along continuous scales. Correlations between these respondent-defined measures of generic fit and other measures mentioned above would certainly help to illuminate the properties of attribute fit.



Audience Fit

One type of fit which is more relevant to sponsorship than other commercial

associations is "audience fif '. Audience fit refers to the nature of the overlap between the audience for the event and target market for the brand. As mentioned earlier, audience fit, in many circumstances, will be highly correlated with other types of fit. It does, however, have some interesting properties with unique implications for several of the seven mechanisms. Before describing these properties in detail, a brief review of the notion of audience fit in the sponsorship literature is presented.





66


Many authors have espoused the commonsense idea that having overlapping

audiences is a good idea. Varadarajan and Menon (1988) suggest that any "cause-related marketing" effort should share the demographics of the target market for the product. Nicholls, Laskey and Roslow (1992) suggest that understanding the differences in audience from one "Hallmark Event" to another may allow marketers to develop optimal messages. Otker (1988) states that "the chance of an effective sponsorship is maximized when there is a perfect marriage of target groups.... of sponsor and the sponsored body" (p. 85). Marshall and Cook (1992) found that 49% of respondents fisted reaching a specific target audience as the most important determinant in selecting an event to sponsor. Thomas (1985) states that "the first concern of a company with sponsorship in mind should be to select the sport, the art or the kind of music which is ... most appealing to its consumer/audience" (p. 322).

Other authors have written about the ability of sponsorship to reach different and/or prized audiences. Crowley (199 1) points out that sponsorship reaches many audiences, including current and potential customers, suppliers, employees, the general public and local community, the business community, and shareholders. He presents the results of a survey which shows the relative importance of each of these audiences by firm type, and orientation. Wither et al. (199 1) found that companies used sponsorship of cultural and art institutions to reach opinion leaders and to achieve community relations goals, whereas sports sponsorship was used primarily to reach the general public directly. Wolton (1988) states that arts sponsorship is often employed in order to reach "decision makers, government leaders, and opinion leaders" (p. 89).





67


Finally, there has been some elaboration on the notion that there are several

distinct "event audiences" which may respond Merently to sponsorship. Hastings (1984) makes a distinction between audience members who are "directly involved in the sponsored events as participants," those who "are directly involved as personal spectators," and those who "follow the event through the media!'. The author notes that event sponsorship differs from advertising in the clear delineation between various audiences for an event. He points out that these differences must be taken into account when creating potential objectives for a sponsorship. These different event audiences could probably serve as a surrogate for an "event-involvement" construct which may have an impact on the functioning of the seven mechanisms. Specifically, it can be proposed that

P66: Event-involvement is highest for participants, lower for those who
attend the event live, and lower still for those who view the event through the
mass media.



Apart from these quotes and ideas, there is no extant elaboration on the audience fit construct, There has been no discussion of its nature, properties, or impacts on sponsorship success. This shortcoming is addressed below.

Audience fit involves two distinct dimensions. The first, which is widely

recognized in other contexts, is "intersection," which simply refers to the size of the overlap in the audience of the event and the target market for the sponsor. Intersection is analogous to the media planning concept of "reach!' (see Sissors and Surmanek, 1986). As far as the intersection component of audience fit is concerned, an overlap of the target market for brand X with event Y of 100,000 is equivalent in value to an overlap of the





68

target market for a brand X with event Z of 100,000, regardless of the absolute size of Y and Z, or the relative size of event audience and brand market (see Figure 2 below). For several of the seven mechanisms, intersection is probably the only component of audience fit that matters.

The second component of audience fit is "subsumption," or the degree to which one audience is subsumed by the other. Here the relative audience size of the two events does make a difference, as does the percentage of each audience which is represented in the intersection. A few examples of this construct are given in Figure 2 for clarification.

Equivalent E B 10% Brand

Subsumption Subsumed


GMC Bog" Event 50% Brand
TR Subsumption Subsumed
.,@Brand 100% Brand
Subsumption Subs






FIGURE 2

TYPES ANDI AMOUNTS OF SUB SUMPTION



As can be seen in the above figures, the type and degree of subsumption depends upon the area of intersection between the brand and the event, the size of the event, and the size of the brand. "Equivalent subsumption" is defined by the intersection representing the same proportion of the event audience as the brand audience. "Event subsumption" is





69


defined by the intersection of brand and event audiences representing a larger percentage of the event audience than the brand audience. "Brand subsumption" is defined by the intersection representing a higher percentage of the brand audience than the event audience. Percent subsumption is simply the percentage of the subsumed associate represented by the intersection.

Looking briefly at the right side of the above figure, it should be clear that there are several differences in the nature of the 100,000 person intersections presented. In the case of equivalent, low subsumption, the intersection of "Indy Car People" and "GMC People" is not particularly predictive of the attributes of either group. In the case of Event subsumption, the relatively high percentage of people watching the "Mud Bog" race who are also in the pickup truck target market may cause people in the intersection to increasingly believe that "GMC Trucks are for people like me". Likewise, in the case of brand subsumption, the high percentage of people in the "high performance carburetor market," who also watch the Indy 500, may lead people in the intersection to increasingly believe that "the Indy 500 is for people like me". In other words, when there is sufficient percent of subsumption, affiliation with the smaller, more individuating associate should flow to the larger, less individuating associate. From these ideas, the following propositions can be made:

P67: Sponsorship under event subsumption leads to greater affiliation with
the brand than sponsorship under equivalent or brand subsumption

P68: Sponsorship under brand subsumption leads to greater affiliation with
the event than sponsorship under equivalent or event subsumption





70




Positioning Fit

Positioning fit refers to the fit between the message of the commercial association and the message of the rest of the integrated marketing communications campaign. To the extent that this is explained by the similarity between the audience for an event and the audience for traditional media vehicles employed by the brand, positioning fit will be highly correlated with audience fit. But positioning fit also has a temporal component, and a sub-segmentation component that make it worth discussing as a unique type of fit. The temporal component refers to the fact that the marketing campaign for some products may reflect certain seasonal characteristics. This makes the timing of events a determinant of positioning fit. The sub-segment component refers to the fitct that marketers often give slightly different messages about their product to different sub-segments. This later component is related to the unique ability of sponsorship to engage in "partial positioning," which is defined below. First, however, a brief summary of the literature relevant to the positioning fit concept is presented.

Very little has been said about positioning fit as a determinant of success in

commercial associations. One exception to this is an article by Crimmins and Horn (1996) who discuss the role of sponsorship within the overall marketing campaign. They stress how important it is to "check the fit between the sponsored property and your marketing objectives" (p.20). This is perhaps the closest thing to a direct reference to positioning fit in the existing literature.

Within the related domain of "corporate image management," Gray and Smeltzer (1985) discuss the importance of having all of a company's communications have a








coherent message. Specifically, they state that "the various sources that communicate corporate image must be coordinated so that a congruent and effective image is projected" ( p. 77). While sponsorship associations are not directly cited, their argument clearly could include sponsorship as an image-communicating device.

Finally, within the domain of brand extensions, Park, Milberg and Lawson (199 1) point out the value of having all of a company's brands share a coherent position. They state that "the degree of perceived fit is a function of' among other things, "brandconcept-consistency perceptions" (p. 186). They also find that extensions work best when both core brand and extension are either prestige or functional in nature. Therefore, at a very high level of abstraction, two associates having a similar positioning seem to be an indicator of success in some circumstances.

As mentioned in the introduction to this section, one reason why positioning fit is particularly relevant to sponsorship is the temporal nature of events. An example of a brand which takes on a somewhat seasonal positioning strategy is Coors Light In the Summer, Coors Light is positioned as fun time beer to take with you to the beach so you won't be too full to play volleyball. In the winter, the brand position seems to move toward being a cozy, by the fire, after-ski beverage. In short, the changing seasonal consumption situations for the product are reflected in the advertising of the product. By sponsoring beach volleyball in the summer and downhill ski racing in the winter, the seasonal positioning is reflected in the sponsorships chosen.

A related, and perhaps more interesting facet of the positioning fit between event and brand is the sub-segment component. Much like some brands have seasonal positioning strategies, others try to position their brands as different things to different





72


sub-segments. Clearly, it is not possible for a brand to be all things to all people, but the degree to which brands modify their messages to fit different segments is somewhat remarkable.

Budweiser beer is a case in point. While Bud positions itself as the "working

man's beer" on weekend sports shows, the message it presents on David Letterman and MTV is quite different. This takes place for two reasons. One reason for this is that the MTV audience may not be interested in a beer that is just for blue collar types. Secondly, and more importantly is the fact that the "working mew' do not stay up on Sunday night to watch 120 Nfinutes" on MTV. Hence, a partial positioning strategy is possible because the more avant-garde message is never seen by the core market who would find it confusing at best, and disturbing at worst.

Sponsorship, where the association is the message, represents a nearly nodownside way for a firm to try partial positioning. An example of this is AT&T's major sponsorship of the Rock and Roll Hall of Fame. While AT&T is not generally considered to be a "rock and roll kind of company," anyone who would find this association to be a negative for AT&T never sees it. The key to the ability of sponsorship to engage in partial positioning in this case is the fact that the message is limited to those who go to the museum.


Conclusion

This chapter has examined how the various types of "fif' relate to sponsorship. A brief review of how the fit construct has been presented in sponsorship and other commercial association domains was reviewed. Other, more technically defied types of





73


fit, such as "attribute fit," "audience fit" and "positioning fit" were then discussed and defined. The chapter concluded with a discussion of measurement issues. WAhe this chapter did address the idea of "lack-of-fit," it did not explore the potential outcomes of associations which are viewed by consumers as inappropriate. This issue is explored in the next chapter.















CHAPTER SIX

SACRILEGE.


Tottenham Hotspur, Liverpool, Manchester United: these names have magical
associations for those who worship the great football god Singing, swaying
masses on Wembley terraces. Muddy mayhem in the twilight gloom. The glory
and the gory. Red vs. Blue. Catholic vs. Protestant. North vs. South. And now...
AVCO vs. Crown Paints. JVC vs. Sharp. Holsten vs. Home Ales. Wright, (1988,
p. 104)


No discussion of sponsorship would be complete without acknowledging the fact that some consumers may think that it is going too far. In an age of increasingly bold associations, the problem should only continue to grow. One acknowledgment that people may be getting fed up is evidenced by Taco Bell's "April Fools Day prank" in which the company placed an ad in the April 1, 1996 New York Times proclaiming:


TACO BELL BUYS THE LIBERTY BELL
In an effort to help the national debt, Taco Bell is pleased to announce that we
have agreed to purchase the Liberty Bell, one of our country's most historic
treasures. It will now be called the "Taco Libert Bell" and will still be accessible
to the American public for viewing ....


Clearly, at least among the advertising creatives at Taco Bell, there may be a problem.

The idea that maybe there are some things which are "sacred," and therefore

should not be involved in commercial association is an intriguing one. The idea that some products may be so "profane" that any association with a non commercial entity would



74





75


result in poor evaluation and even outrage is equally attractive. When combined, the two notions form a complex and largely unexplored area in marketing. This area is referred to here as "sacrilege".

The notion that consumption objects can take on dimensions of being "sacred" or "profane" has received some attention in the consumer behavior literature. Relc, Wallendorf and Sherry (1990) describe how the processes of the secularization of religion and the sacralization of the secular are at work in contemporary society. They find that consumers maintain the sacred in consumption through separation of sacred and profane, and through the performance of rituals. Relic and Wallendorf (1990) explore the sacred and profane dimensions of money itself. They find that the sacred and profane aspects of money depend upon its sources and its uses.

Hirschman (1991 a) explores the way advertisers utilize sacred and profane

consumption imagery to persuade. In an exploration of process of the secularization of the sacred, Hirschman (1991ib), describes how the commercial transactions involved with surrogate motherhood and in-vitro fertilization can be made less profane. She finds that this is achieved by stressing the sacred aspects of the acts of childbirth and family creation.

While the sacred and profane nature of consumption objects has received attention, there has been no examination of these issues in the context of commercial associations. Some might argue that words like "sacred," profane" and "sacrilege" are too strong to be used in the context of commercial associations. There are several recent examples, however, which, seem to meet the more traditional notion of sacrilege. DeCoursey (1995) reports on the widespread feeling of over-commercialization associated with the Pope's 1995 visit to the United States. Brewer (1993) describes attempts to associate the Pilot





76


Precise Pen with Middle East peace'. Brewer (1993) also reports on an attempt by KSwiss shoes to capitalize on the fact that the Pope bought a pair of their sneakers while in Denver. Finally, there is the case of the McDonalds World Cup Soccer promotion in which sacred Koran scripture was featured on two million throw-away hamburger bags. This promotion greatly offended Muslims in Great Britain (Alonzo, 1994).

The above examples seem to provide evidence of "sacred" events or objects which are difficult to successfully exploit through commercial association. Similar positive examples of associations which failed because the product was too "profane," are not as readily available. While Tambrands was among the corporate sponsors of the Women's National Basketball Team (Rosner, 1995) and Angel Soft bathroom tissue is involved with a "cares for kids progranf (Weisz, 1995), there are no known cases of either feminine hygiene products or bathroom tissues having an event named after them. The same is true for other product classes found to be "causes of irritatiorf' (Aaker and Bruzzone, 1985), such as laxatives and breath fresheners.

An interesting and complicated possibility is that sacrilege may be perceived for brand-event associations in which the brand is not strictly profane and the event is not strictly sacred. Some combinations of associates, for reasons which have not yet been examined, may lead consumers to consciously elaborate on the "inappropriateness" of the association. This may sometimes go beyond mere elaboration in one's own mind, leading to actual verbalization. Discussion among consumers about the crass and commercial nature of the association can then ensue. The impact of such discussion among consumers about associations is also currently unexplored. The pen was apparently used in signing some of the documents.





77


An example of this right be the case of the "Tostitos Fiesta Bowl," which was

played to determine the national college football champion of the 1995 season. While the Fiesta Bowl may not have been strictly "sacred" and corn chips may not have been strictly "profane," discussions which were seemingly the result of perceived sacrilege could be heard among Florida Gator fans around the time of the game. Comments like "That's not right ... the taco chip national championship" and "Good grief, the Tostitos logo on the sweatshirts is bigger than the words 'Florida Gators' were common.

Using a small, convenience sampling of comments, it also seemed that the people who were bigger fans were more likely to express disdain for the association. While there is absolutely no empirical work in this area, it is possible to propose the following:



P69: The more mundane the brand, the more likely the association to result
in perceived sacrilege.

P70: The more sacred the event, the more likely the association to result in
perceived sacrilege.

P71: For associations which run the risk of sacrilege, greater consumer
involvement with the event results in a greater likelihood of perceived
sacrilege.

P72: For associations which run the risk of sacrilege, lower levels of attribute
fit lead to a greater likelihood of perceived sacrilege.

P73: The more blatant the perceived commercial intent of the association, the
greater the likelihood of perceived sacrilege.


The most frequent response to perceived sacrilege may well be for the consumer to elaborate that "those two things don't belong together". This elaboration may provide the necessary condition for a "dissimilarity frame'. Given that "people attend more to the





78


common features in judgment of similarity than in judgment of difference" (Tversky, 1977, p.340), it is possible to propose:

P74: Under perceived sacrilege, attributes common to event and brand
become less salient.

IP75: Under perceived sacrilege, attributes not shared by event and brand
become more salient.

Therefore,

P76: Under perceived sacrilege, attribute fit is reduced.


The elaborations generated by perceived sacrilege may well be the commercial

association equivalent of "counter-arguing". This is serious, because one of the benefits of commercial associations in general is the inability of consumers to counter-argue in the classic sense, when "the association is the message'. It is therefore possible to propose that:



P77: Sacrilege-inducing associations will lead to more counterarguing than
will non-sacrilege-inducing associations.



Conclusion

While consumer outrage at associations considered to be sacrilegious is understudied, the growing number and forms of commercial association make it an especially salient avenue for further research. As the boundaries of what can be sponsored get pushed farther and farther, the importance of this area of investigation should only increase. One way to minimize the chance of negative consumer response to an association is through additional





79


promotional effort which explains the association to the public. The potential benefit of such additional communication about the association is presented in the following chapter.














CHAPTER SEVEN

ASSOCIATION-ENHANCING COMMUNICATIONS (AEC)



As alluded to earlier, sponsorship associations do not exist in a vacuum. Firms can engage in a variety of communication activities in order to maximize the impact of the association for which they have paid. These communications are referred to here as "association-enhancing communications," or AEC. The expenditures above and beyond the fee paid to "be the sponsor' are generically referred to by practitioners as "leverage"(Andrews 1996). Leverage can attempt to enhance awareness of the sponsorship among consumers through advertising. It can also, however, take the form of tie-in promotions, sales promotions, case lot discounts, and other perks to the channel members in conjunction with the event.

Leverage has received some attention in the sponsorship literature. Meenaghan

(1991 a) defines "leveraging" as "the additional effort, largely promotional, which must be invested by the sponsor in order to properly exploit the opportunity provided as a result of securing particular sponsorship rights" (p. 43). He goes on to say that it is generally agreed that an amount at least equal to the sponsorship fee should be allocated to leveraging the sponsorship. Otker (1988 ) presents a somewhat higher estimate, stating that "for each dollar spent on buying, another three dollars is spent on exploitation" (p. 77).


so





81


Gardner and Shuman (1987) empirically investigate the amount spent on leverage. They survey 300 Fortune 500 companies about their sponsorship activities. Eighty-nine percent of responding companies engaged in sponsorship report some form of project support, such as advertising tie-ins. The authors find that within their sample, an average of forty percent of sponsorship cost was spent on leveraging the sponsorship (including advertising, promotion, and employee tie-ins). They also find that less than ten percent of companies spend more money leveraging the sponsorship than the sponsorship actually costs.

Although the optional level of spending on leverage seems to be a matter of some dispute, the facts seem to support the idea that most sponsorships involve some additional spending on communications about the association. The interesting question then becomes "what should be said and to whom?" Crimmins and Horn (1996) propose that when the link between an event and a brand is clear, as with basketball and athletic shoes, consumers can make the link themselves. When the link is harder to forge, however, the link must be made for them, as with Visa driving home the message that the Olympics don't take American Express. While this advice seems to be face valid, it does not answer the biggest question. When does the meaning of the association "need to be driven home"? When does this help? When might it hurt? When might it be merely a waste of money?

In the related area of brand extensions, Aaker and Keller (1990) find that

elaborating on brand extension attributes improves evaluations of the extension. They propose that this occurs by alleviating negative inference making. Explaining which





82


attributes the extension shared with the core brand is found to be superior to elaborating upon the quality of the original brand in generating positive evaluations of the extension.

An example of how this is relevant to sponsorship may make the point clear. Suppose a dry cleaner is sponsoring a stock car. While attributes of a race car like "speed" are positive for a dry cleaner, attributes like "dusty" or "oily" are not. Given the Aaker and Keller (1990) findings, and the implication of the above, it can be stated:


P78: Communications stressing the positive attributes of an event for a
brand have greater impact on brand evaluation when the event possesses
both brand-positive and brand-negative attributes than when the event
contains only brand-positive attributes.


Previously the findings of Markman and Gentner (1993) were discussed. These

authors found that the act of merely considering an association had an impact on subject's structural alignment. Subjects who were told to make a similarity comparison tended to associate two scenes along more abstract, relational dimensions. Assuming that association enhancing communications (AEC) can successfully induce people to make a similarity comparison, then it can be proposed:

P79: AEC efforts which explicitly tell a consumer to "see the fit" between
event and sponsor will lead to greater perceived fit.

P80: The effects predicted in P79 are greater when the event and brand share
attributes which are abstract and relational than when the event and brand
share attributes which are concrete.


Quite a bit of research has been done which focuses on whether explicitly stating conclusions is helpful. Kardes (1988) found that more positive brand perceptions are derived from allowing the consumer to draw his/her own conclusions in advertising.





83


Sawyer and Howard (1991) demonstrated that for high involvement individuals, openended advertisements were superior to closed-ended ads in terms of attitudes, purchase intention, and choice. The authors found no significant difference, however, for the low involvement groups.

These studies seem to contradict early work which found that explicitly drawing the conclusion for the audience was superior to letting them draw the conclusion themselves. Hovland and Mandel (1952), for example, found that over twice as many subjects agreed with a speaker's argument when the conclusion was explicitly drawn by the speaker as opposed to being left up to the audience (neither intelligence nor personality type had a measurable impact on this effect). The theoretical explanation for this seeming inconsistency is that the impact of omitting the conclusion depends upon how difficult the conclusion is to draw. If it is very difficult, then few people will draw the appropriate inference, and effectiveness is reduced. If people are capable of "getting it" on their own, however, omitting the conclusion should be more effective.

Several authors have hypothesized as to why this might be so. Linder and Worchel (1970) demonstrate how mechanisms like cognitive dissonance can lead to difficult self drawn conclusions being believed more than conclusions which were easier to draw. One explanation for this finding is that people want (or need) to find value in accomplishing a more difficult task, and that discovering something with which one agrees is more valuable than discovering something with which one does not agree.

Similar ideas have been put forth in the domain of categorization, which is more closely related to the pairing of commercial associates. Cohen and Basu (1987), for example, propose a direct impact on affect of discerning what two things have to do with





84


one another. Specifically they state that "since categorization produces a reduction in uncertainty, positive affect may result from a successful fit and negative affect from an inability to categorize an item, particularly if the resulting judgments or inferences are important"(p. 470). Combining this idea with the findings of Linder and Worchel (1970), it can be proposed that



P81: For those who determine how a sponsor is like in event on their own, the more difficult the association is to understand, the more positive affect
will be generated for the association.

P82: The effect predicted in P81 is greater for consumers who are highly
involved with the event than for those who are less involved with the event


The above propositions must be qualified, however, by the generally accepted idea that for most people, figuring out what a sponsor has to do with an event is not an activity likely to receive much cognitive effort. Therefore, any association that requires more than the simplest consideration to understand will likely be ignored. This fact does not, however, make the above discussion irrelevant. On the contrary, it allows some specific predictions about the effectiveness of certain strategies for association-enhancing communications. Some specific AEC strategic proposals are as follows:



P93: Communications which explain obvious associations harm the
association.

P84: Communications which "all but" explain the associations that are not
obvious are superior to communications which actually do explain the
association.





85


Conclusion

Given the industry rule of thumb of spending three times the sponsorship fees on "leverage," the study of how and when these communications about the association will be most effective is an interesting area of investigation. While the extant literature has done a thorough job of looking at the impact of omitted conclusions in the context of traditional advertising, the unique mechanisms involved in sponsorship warrant finther research into how to best spend these billions.

This chapter has considered how firms can maximize the impact of engaging in sponsorship through "leverage," or additional spending above the sponsorship fee. Such additional spending on "association enhancing communications" was posited to have a measurable impact on several sponsorship variables. Consumer difficulty in determining what a brand and event have in common was proposed as a key variable in deciding how much should be said about an association. Formal propositions about how AEC will impact these variables were also presented.

The last several chapters have explored broad aspects of the relationship between sponsorship variables at a somewhat abstract level. The following three chapters focus more narrowly on a subset of the issues raised thus far. These chapters develop and test a two-stage model of building brand equity through sponsorship. After an explanation of the model's conceptual underpinnings, eight empirical studies are presented which test various components of this model. Chapter I I then summarizes the findings from the eight studies, and discusses their theoretical and managerial implications. The dissertation concludes with suggestions for future research.














CHAPTEREIGHT

TBE "ACT OF EVENT AND ASSOCIATIONAL FACTORS ON INFERENCES ABOUT TBE BRAND


The preceding chapters have provided an overview of what is currently known about sponsorship. They then went on to map out the somewhat vast "unexplored territory" of how sponsorship may work. While the wide array of unexplored issues makes the general area of sponsorship an exciting domain in which to work, it also makes it necessary to limit to a subset the number of issues which can reasonably be addressed within the scope of a dissertation. This first empirical chapter, therefore, begins with a brief discussion of which issues were selected for empirical investigation and why they were chosen.

The core unexplored issues in sponsorship are how and when does associating a brand with an event build equity for the brand, and, ultimately, increase choice probability for a segment of consumers. Given that the mechanisms at the lower end of the "necessary elaboration continuum' (i.e. awareness, affect transfer) are not expected to differ functionally from their counterparts in the domain of advertising, it seems most fruitful to pursue those mechanisms at the upper end of the continuum. Very little is known about mechanisms like implied size, implied endorsement and reciprocity. These mechanisms are also likely to be more greatly impacted by sponsorship than by other forms of promotional activity.


86





87


These mechanisms are so far up the "necessary cognitive elaboration continuum' that they all strongly rely upon consumer inference making in order to function. For a sponsorship association to impact the perceived size of a company or brand, a consumer must make an inference that "this company must be big in order to be able to sponsor such an event." In order for the implied endorsement mechanism to be engaged, a consumer must make the inference that "this brand must be OIK, or else it wouldn't be allowed to be the sponsor of this event. In order for reciprocity to be engaged, a consumer must make the inference that "this brand is doing something to make possible and/or better an event I care about."

The realization that inference making is key to so many important, uniquely

sponsorship-related mechanisms has led to the development and testing of "the inference based model of building brand equity through sponsorship." This model yields many specific hypotheses about how audience, event, and associational factors impact the inferences consumers draw about sponsoring brands, and how these inferences impact choice probabilities for various classes of brands. The inference-based model is depicted graphically in Figure 3.

As the figure makes clear, the inference-based model consists of two stages. In stage one the consumer who is exposed to the sponsorship association makes inferences about the brand. The number and kind of inferences made are determined by event and associational factors, as well as by characteristics of the audience members' relationships with the event (audience factors). Stage two consists of converting these inferences into changes in choice probability for a particular brand on the basis of the sponsorship-based inferences. Here, categories which differ in experience ambiguity, and brands which differ


I





88


in their goodwill position, are expected to be differentially impacted by each type of

sponsorship-based inference.




-Big br
-Legitimate
Audience Factors: -Makes event E v ent Factor. Intrinsic personal E Frelevance
Size Domain knowledge Q
Felt involvement :







Asogat o odilnpsto
(msaesn Made
Asoiational Factors! sEvent/Brand Fit
Duration of Assoc.l















FIGURE 3

THE INFERENCE-BASED MODEL OF BUILDING BRAN]) EQUITY THROUGH SPONSORSHIP



The remainder of this chapter details a series of three studies which investigate the

impact of event size, brand/event fit, duration of association, and share of presence

(defined below) on consumer inference-making about a sponsoring brand. The following





89


chapter describes two experiments which explore the role of audience factors on consumer inference making. Chapter ten explores how sponsorship-based inferences differentially impact different brands and categories.



Conceptual overview and construct definition for studies la, lb and 1c

Studies I ab and c investigate how event and associational. factors impact inference making. Specifically, they explore how event size, brand/event fit, share of presence, and duration of association impact inferences about brand size, brand legitimacy and event facilitation. These constructs are defined below.



Event Factors

"Event Factors," or attributes of the event being sponsored, are expected to

directly impact consumer inference making about the sponsoring brand. Many aspects of an event, such as perceived size and prestige, might influence inferences about sponsoring brands. Pretesting was undertaken to determine the relatedness of these constructs. One hundred eleven actual sponsored events were evaluated by independent groups of subjects. Perceived prestige was evaluated along a seven point scale anchored by "very prestigious" and "not prestigious at all" (n7-21). Perceived size was evaluated along a seven point scale anchored by "very large" and "very smalr' (n718). Mean scores for each construct were calculated. The means for both constructs, for all I I I events, were then submitted to simple correlational analysis. The correlation coefficient between the two constructs was found to be .785.





90

While this correlation was quite high, further investigation of the data attempted to determine if the two constructs could be orthogonally manipulated experimentally. While it was possible to find events which differed in prestige and size, these events fell into very specific event categories. That is, nearly all of the high prestige, low size events were arts related, while nearly all the high size, low prestige events were dirt-track motor sports or rodeos. Rather than limit the initial test of the inference-based model to these rather unique categories of event, where size and prestige were poorly correlated, a decision was made to include only one of the two constructs. This was done with the knowledge that the chosen construct is most likely correlated with the other. Given the interesting theoretical predictions about the impact of event size on the perceived size mechanism, the perceived event size construct was chosen for exploration.



Perceived Size

Perceived size of the event is expected to impact inferences about the size of the sponsoring firm. Here, consumers are expected to infer that only a large brand could afford to sponsor a large event. Perceived size of the event is also expected to affect inferences about the commitment of the brand (or firm) to supporting the class of activity. For example, an athletic shoe sponsoring a large tennis match might be perceived as more committed to tennis than an athletic shoe which sponsors a smaller, less important tennis match. Perceived size of the event is also expected to impact consumer inference making about the necessity of the sponsor to the event taking place ("event facilitation"). Here, sponsoring a small event is more likely to lead consumers to infer that without the support of the brand, the event would not take place.








91




TABLE 1
MEAN PERCEIVED PRESTIGE AND SIZE OF 111 EVENTS AND

ORGANIZATIONS




EvnIantta mi P i 11yardfmnaht flat hm Event~rgnatU Sm. ren
AMA PuRsiig 4.00 3.33 Pra Swig FW 230 Z2 Plftewh Three~. W e 269 4.00
AMC Ccer Reseac Cmle 5.30 6.00 NiW d fti Owlse Rgal 3.00 3.43 Polow Aript6Wi 2.67 3.33
Arr~iara Conlwrsod*y 6.11 6.76 He" Ford Muszn 3.44 4.110 Patcs Rs&g Teern 200 1.6
Anwic. Dl~ius Asoddan 5.94 6.67 Houla Uveisodc Show 212 2.3 Pw Twn 267 2.67
Arrierice HeW OWNf 5.17 6.00 IMAX Capwai 4.60 3.05 Reverdi Hwlw /ALSAC 4.86 5.14
Arrierma Ham Shms Asem 2.46 3.42 hdIy 5.00 4.14 ftk~ihg Smu. Ck=w 6.56 4.10)
Arnercim Ljz,*qrJad 1.39 2.3 6 IrdV 1 Ope LA (Tawwf) 4.06 5.14 Rodde Cwlsr 5.61 6.14
Anerka Podplayurs 2M 2-61 htimakn Raftvd U"t 4M3 4.19 Rodcy Mainm Blk Fourdda 3.11 3.52
Anwilm Clueter Hams 239 3.43 Jaifte Balle 2.67 4.71 Roal Acaduny a 3.44 4.48
Anwic.oo Recng Swiss 3.61 3.33 Jurgar Adwni..nk ftc 3.76 3.81 Sm DWR o"1a Soe 283 2176
Artvt Paaxi 5.30 5.52 Kuvuedy 0~ or P. Ala 4.56 5.461 Son DlageZoo 3.81 4.81
Belftar Zoo 4.11 4.43 Kuimdy Kaleg WO~M 3.33 4.43 Son From*= Juz Fes" 3.46 4.43
Big Apple Circs. 11.. 3.39 3.33 KtO Drugs SakE 3.39) 3.14 Sdwmc Musurn d kmd 275 4.05
Brvodleld ZO*vCftW 38 4.0 Lakdard Pweluu Hok& Town 2M6 243 VOWSi Pwtun* As Carde 3.17 4.33
Builders S""mAlarm Sam! 42B 3.57 Waty Sciece Caerw 23 3M8 &We A~,ukm 2A4 3.67
Calgary Slompes 2683 252 Lhndet Pak1 Zoo CWNgo. IL 3M3 4.52 SL. Maya Food Bark 4238 4.71
CaigaZac 3.50 4.24 Macho D~lirs 5.61 6.00 Tesd Gaiamudo 239 281
Cwidmi FturuSloskg Ass=c 3.3 4.02 Mvry~dpkugkuCa& qsl~ 4.11 5.3 Tiuo 00y 3.17 4.10
Cerodimn CO~C Ajeadellm 4.33 5.67 M*d Mum=n of FINe Ats 23l 4.16 Town Dewrout DENMARK 244 3.81
CW City Rkrufu 3.00 3.52 Muiple Slerois Sociey 4.78 5M8 The Oulduwu Tmesl 3.44 4.36
Conegle Hlld 5.11 6.481 Wasm or Sduica & bdhay 4.00 4M5 Thle NskCowAM~c 4.00 4.16
Chlsli Maiu Seefty 4.44 3.0 MYeGC Seqwt M-ss 3.11 3.71 Ta*f Tervil Classc 3.72 5.06
Ctvry Oui hAs F W 2.94 3.62 lng. 3.17 3.66 U.S. 10K Cmlai 256 3160
O'duMi- Ha1s Sociey 3.50 4.57 Not MAsc. of fter. At"=dc 4.16 4.81 U.S. Lug@Asbaw 294 3.33
Chk smHoqWW d Alsbana 4.11 4.71 N@L AnxO.S@=Wwdy Sdvd 3.3 32D U.S. Tvalwado hilai 27 3.38
Claiddi HWalsM dPhtsbw 4.67 5.06 Neadi Ce~iiln kw%^~s Ic. 256 3.43 Ikte Sim" SakW 3.17 4.10
C~lrsna Miadle Netallcr 6.17 6.00 NCAA 6.06 5.71 Urda Stas Twuts Asaocdk 4.61 5.3
Ci'dicu* Musm" d kldwuipdal s 36 4.14 Ndini OeagrqNc Soddly 5.44 5.38 Urdied Way 6.39 5.06
COlda Play TOAVn Theses 3.33 4.24 Ndad Wilmsf Federao 5.17 5.24 Lkivwully Muics!d Sod*y 3.61 4.52
Circus du Sold 3.3 4.57 Now Jersay BW=c Fee"h 218 257 USA iii, 94 3.24
City d Hape 3.3 4.02 NMMA Bad Tour 222 3.06 USA WMOslt 4.00 3.10
cadmi Va. Cwcv 5.00 5.38 Ntngc 4.00 4.14 VmcAaw Art Gtery 216 4.30
Owiaduac kwflab d Sdwms 3.00 4.81 OKO)Im 267 320 Vaidme Fae" 23 4.00
Dowintmw DNOWc Housim. TX 3.60 3.43 Oro" a" 6.3 &5 Wal DlmeW* Wald ailm 6.3 4.57
ShYm Preo Grooded 3.44 3.52 Oegim She~lae Fe94 267 3.71 Wad*cgbl DC P-~c 3.0 4.3
Epiepy Facimd ft~ji des 4.72 5.81 CudanAmerica 239 3.14 Wasum n air 272 2.24
Feud Via Chikuc 5.17 5.17P0~ Ndbara Ec4*IAm 2633 3.24 Whkle CmeervdM SWW~Y 4.00 4.5







The idea that sponsoring an event can influence the perceived size of the sponsor



has been examined and has received same empirical support. McDonald (199 1) did not



find evidence of such an effect. This may have been due to the fact that perceived size of



companies studied was already very high (e.g., Phillips and Cadbury). Evidence that



sponsorship can influence perceived size was found by Raijaretnam (1994) who reports



that a tire maker that spent almost its entire marketing budget on sponsorship



dramatically increased its perceived size.





92


The marketing literature provides other evidence of the benefits of perceived size. Kirmani and Wright (1989) propose three possible mechanisms through which perceived expenditure on advertising might lead to inferred quality. One is the notion that confidence in quality leads companies to spend more on promoting a product. Another is a perceived relationship between promotional spending and quality in some markets. Finally, they cite the idea that "perceived advertising expense is a signal of a firm's financial strength, probable social acceptance or some other factor that defines quality in some markets" (p. 344). Kirmani (1990) proposes that perceived expenditure will be used as cues to quality when they are more diagnostic than other cues.



Associational Factors

Associational factors, or attributes of the association between the event and the brand, are also expected to impact inference making about the brand. Associational. factors include fit between the event and the brand, the duration of association between the event and the brand, and the share of presence a particular brand has relative to all the brands at the event. Brand/event fit was discussed at length in chapter five. Duration of association refers to the event-brand associational "history" (number of years the brand has sponsored the event). Share of presence is a measure of consumer perception of how large a presence a brand has at an event relative to the other sponsors.

The fit between a brand and an event is expected to be a strong determinant of the number and kinds of inferences a consumer makes about a sponsoring brand. As previously discussed, fit can be directly measured as a unidimensional construct which is perceived by consumers. Operationally, it has been defined in several ways by researchers




Full Text
72
sub-segments. Clearly, it is not possible for a brand to be all things to all people, but the
degree to which brands modify their messages to fit different segments is somewhat
remarkable.
Budweiser beer is a case in point. While Bud positions itself as the working
mans beer on weekend sports shows, the message it presents on David Letterman and
MTV is quite different. This takes place for two reasons. One reason for this is that the
MTV audience may not be interested in a beer that is just for blue collar types. Secondly,
and more importantly is the fact that the working men do not stay up on Sunday night to
watch 120 Minutes on MTV. Hence, a partial positioning strategy is possible because
the more avant-garde message is never seen by the core market who would find it
confusing at best, and disturbing at worst.
Sponsorship, where the association is the message, represents a nearly no
downside way for a firm to try partial positioning. An example of this is AT&Ts major
sponsorship of the Rock and Roll Hall of Fame. While AT&T is not generally considered
to be a rock and roll kind of company, anyone who would find this association to be a
negative for AT&T never sees it. The key to the ability of sponsorship to engage in partial
positioning in this case is the fact that the message is limited to those who go to the
museum.
Conclusion
This chapter has examined how the various types of fit relate to sponsorship. A
brief review of how the fit construct has been presented in sponsorship and other
commercial association domains was reviewed. Other, more technically defined types of


139
implications of this finding, however, it is necessary to make sure that event facilitation is
being driven by fit, and not some other aspect of Evian and Texaco. To this end, study 2b
checks to see if the main effect result replicates when Texaco is high fit and Evian is low
fit.
Study 2b
Overview
This study replicated the method in 2a by using an event with high perceived fit
with Texaco and low perceived fit with Evian Open and closed ended measures of event
facilitation and commercial motive were taken. Since the event which best flipped the
fit was exclusively a spectator sport, it did not make sense to measure the audience
factors used in 2a. This study simply checks for the confound of fit with brand in fits
impact in inference making.
Subjects
Forty-one undergraduates in an introductory marketing class at a major
southeastern university took part. Each received course extra credit for participation.
Design
The design is identical to 2a, except the event was changed so as to make Evian
the high fit event and Texaco the low fit event.


CHAPTER FOUR
SEVEN MECHANISMS THROUGH WHICH SPONSORSHIP MAY WORK
Almost no conceptual or empirical studies have, thus far, addressed how
sponsorship works. It is the contention of this dissertation that at least seven distinct
mechanisms exist through which sponsorship may function. As proposed in the previous
chapter, these mechanisms may impact brand equity and brand positioning. They may also
impact other relevant marketing variables of interest. These mechanisms are not
hypothesized to be exclusive of one another. Two or more of them may function in any
given sponsorship association. Together, however, they are proposed to represent a
reasonably complete set of the ways sponsorship works.
The simplest way to order these mechanisms is by the amount of cognitive
elaboration which is required by a consumer for the mechanism to function. In
approximate order of necessary cognitive elaboration, they are simple awareness, affect
transfer, image transfer, affiliation, implied size, implied endorsement, and reciprocity.
The rela tionship of the mechanisms depicted in Figure 1 below.
27


103
TABLE 3
PERCIEVED LEVEL OF FIT BETWEEN A RUNNING EVENT AND SELECTED
PRODUCT CATEGORIES
Mean
St Dev
Liquor Store
1.65
0.99
Oil Company
2.00
0.79
Mexican Food
2.20
0.89
Realtor
2.20
0.95
Fast Food Restaurant
2.25
1.12
Wine
2.25
1.02
Ice Cream Bar
2.60
1.05
Candy Bar
2.70
1.98
Software Company
2.70
0.98
Coffee
275
1.45
Brewery
2.95
2.24
Snack Food
3.00
1.78
Soft Drink
3.00
1.92
Steak house
3.00
1.08
Insurance Company
3.30
1.03
Music Store
3.45
1.43
Automobile Manufacturer
3.55
1.47
Drug Store
3.85
1.90
Airline
4.10
1.52
Newspaper
4.10
1.62
Cellular Phone Company
4.20
1.40
Hotel
4.40
1.57
Bank
4.45
1.15
Grocery store
4.55
1.54
Radio Station
4.70
1.42
Film
4.75
1.65
Hospital
4.95
1.70
Low Fat Frozen Entre
4.95
1.67
Pasta
5.00
2.00
Herbal Tea
5.25
1.33
Fruit juice
6.30
0.98
Health Care System
6.35
0.88
Mineral Water
6.55
0.76
Sporting Goods Store
6.70
0.47
Health Club
6.75
0.91
Performance Drink Concentrate
6.90
0.31
Sports Drink
6.90
0.31
Athletic Shoe
7.00
0.00
Energy Bar
7.00
0.00
* l=very low fit; 7=very high fit


109
regressed on fit, size, duration and share, as well as the interactions of fit with size, fit with
duration and fit with share. Unless otherwise specified, the statistics reported for these
predictors on the dependent variables are based on the parameter estimates from a seven
degree of freedom, dummy coded regression model.
Impact on Perceived Size of the Brand
Two measures were taken of the perceived size of the brand. Absolute size was
measured by a single seven point scale anchored by one of the smallest sports drinks
(coffees) in Belgium and one of the largest sports drinks (coffees) in Belgium.
Perceived market share was measured by a single seven point scale which asked how large
they thought the brands market share was compared to other category brands in the
country. It was anchored by very small share and very large share. Analysis revealed
no significant impact of the independent variables on the absolute size measure.
Regressing on the perceived market share variable revealed two significant main
effects. First, the size of the event sponsored significantly predicted perceived market
share (F=5.29, p<03). As predicted, brands which sponsored large events were perceived
to have larger market shares. Thus, HI is supported.
The duration of association also predicted perceived market share (F=4.87, p<05).
As predicted, brands associated with an event for five years were perceived as having
larger market shares than were new sponsors. Therefore, H2 is supported as well. No
support, however, was found for H3. Share of presence at an event had no measurable
impact on perceived market share of the brand (F<1).


130
perceived size was also obtained from the close-ended data in study la. This lends
additional support for the robustness of the effect.
In the comparison of Study lc to la, title sponsorship was found to impact event
facilitation inferences. The same analysis revealed, however, that being a title sponsor can
also lead to a greater number of negative, commercial motive inferences. As mentioned
previously, an important issue for future research will be to explore when consumers make
each type of inference.
This series of studies has examined how event and associational factors impact
consumer inference making about brands. The first stage of the inference based model
(see Figure 3) also includes the potential for audience factors to impact the inference
making process. This issue is addressed in the next chapter.


82
attributes the extension shared with the core brand is found to be superior to elaborating
upon the quality of the original brand in generating positive evaluations of the extension.
An example of how this is relevant to sponsorship may make the point clear.
Suppose a dry cleaner is sponsoring a stock car. While attributes of a race car like
speed are positive for a dry cleaner, attributes like dusty or oily are not. Given the
Aaker and Keller (1990) findings, and the implication of the above, it can be stated:
P78: Communications stressing the positive attributes of an event for a
brand have greater impact on brand evaluation when the event possesses
both brand-positive and brand-negative attributes than when the event
contains only brand-positive attributes.
Previously the findings of Markman and Gentner (1993) were discussed. These
authors found that the act of merely considering an association had an impact on subjects
structural alignment. Subjects who were told to make a similarity comparison tended to
associate two scenes along more abstract, relational dimensions. Assuming that
association enhancing communications (AEC) can successfully induce people to make a
similarity comparison, then it can be proposed:
P79: AEC efforts which explicitly tell a consumer to see the fit between
event and sponsor will lead to greater perceived fit
P80: The effects predicted in P79 are greater when the event and brand share
attributes which are abstract and relational than when the event and brand
share attributes which are concrete.
Quite a bit of research has been done which focuses on whether explicitly stating
conclusions is helpful. Kardes (1988) found that more positive brand perceptions are
derived from allowing the consumer to draw his/her own conclusions in advertising.


161
underpinnings of these seven mechanisms were proposed and discussed. This allowed
many formal propositions to be developed.
Constructs hypothesized to have an impact upon the functioning of these
mechanisms, such as fit and sacrilege were then developed. The impact of various types
of brand/event fit were discussed at length. A review of how the fit construct has been
presented in the study of sponsorship and other commercial association domains was
presented. Other, more technically defined types of fit, such as attribute fit, audience
fit and positioning fit were then defined. The discussion of fit concluded with
measurement issues.
Sacrilege arising from associations perceived to be inappropriate by consumers
was then discussed. Formal propositions about when it may arise and its likely impact on
sponsorship success were also presented. A theoretical discussion of how other
communications can be deployed to maximize the impact of the sponsorship association
concluded the background and theory chapters.
A two stage inference-based model of building brand equity through sponsorship
was then developed to explain the functioning of several of the seven mechanisms. The
first stage in the model considered how audience, event and associational factors of a
sponsorship impact consumer perceptions about brands. Five studies were presented
which explored these issues. The second stage of the model considered how these
inferences impact choice for different brands and product categories. Three studies
explored these issues using stated-preference choice experiments.


85
Conclusion
Given the industry rule of thumb of spending three times the sponsorship fees on
leverage, the study of how and when these communications about the association will be
most effective is an interesting area of investigation. While the extant literature has done a
thorough job of looking at the impact of omitted conclusions in the context of traditional
advertising, the unique mechanisms involved in sponsorship warrant further research into
how to best spend these billions.
This chapter has considered how firms can maximize the impact of engaging in
sponsorship through leverage, or additional spending above the sponsorship fee. Such
additional spending on association enhancing communications was posited to have a
measurable impact on several sponsorship variables. Consumer difficulty in determining
what a brand and event have in common was proposed as a key variable in deciding how
much should be said about an association. Formal propositions about how AEC will
impact these variables were also presented.
The last several chapters have explored broad aspects of the relationship between
sponsorship variables at a somewhat abstract level. The following three chapters focus
more narrowly on a subset of the issues raised thus far. These chapters develop and test a
two-stage model of building brand equity through sponsorship. After an explanation of
the models conceptual underpinnings, eight empirical studies are presented which test
various components of this model. Chapter 11 then summarizes the findings from the
eight studies, and discusses their theoretical and managerial implications. The dissertation
concludes with suggestions for future research.


46
sponsorship of a Pompeii museum exhibit was received much better than corporate
sponsorship of the arts had typically been received up to that point. Because IBM was
able to use its computers to bring a computerized archive of the dig into the museum, the
attendees not only saw a fit between the exhibit and IBM, they actually were able to
perceive a value added to their experience from the association. Negative reactions to the
association (measured by polling attendees) were, as a result of the association making
sense, much lower than negatives toward other similar art sponsorships.
Within the context of evaluating sponsorships, McDonald (1991) argues that
synergy between the brand and the event is essential. He proposes that assessing
synergy would be useful, as opposed to doing simple favorability tracking. He does not,
however, specifically define synergy nor does he propose how to measure it.
The idea of fit is also found throughout the rest of the commercial associations
literature. In the domain of brand extensions, Aaker and Keller (1993) define various types
of fit. They state that it is clear that the interactive relationship between quality and
the fit variables is sensitive to the levels of perceived quality and fit (in addition, possibly,
to different types of fit) (p. 56). These authors operationally defined fit to be the degree
to which the brand extension is a substitute or a complement for the core brand. Such an
operational definition, however, cannot be applied directly to the domain of sponsorship.
Broniarczyk and Alba (1994) reassert the importance of fit in brand extensions,
stating an assumption common to virtually all research, however, is that brand affect and
product category similarity play important roles. (p. 214, emphasis added). They
continue, we agree that, all things being equal, the best scenario for success involves
extension of a highly regarded brand to a very similar product class (pg. 215, emphasis


17
P5: The impact of perceived quality of an event on evaluation of, and positive
affect toward, a sponsor depends upon the level of fit between sponsor and
event. High fit facilitates the impact of event quality, whereas poor fit
impedes it.
The commercial association literature also provides evidence of many moderators
of the impact a commercial association may have. Examples include previous
associations, arousal, symbolic value, type of good, and priming. Each of these topics is
considered below.
In the case of previous associations, Walker, Langmeyer and Langmeyer (1992)
find that previously unendorsed products (bath towels and VCRs) pick up more of the
personality of the endorser than does a previously endorsed product (jeans). This
suggests the possibility of the following propositions for sponsorship:
1P6: Brands in product classes which have not previously engaged in
sponsorship form stronger associations in the minds of consumers than
brands in categories where sponsorship is common.
F7: Brands which have not previously engaged in sponsorship form stronger
associations with events than brands engaged in multiple sponsorships.
.Another moderator of commercial association effectiveness is arousal.
Sanbonmatsu and Kardes (1988) find that high arousal causes an increase in the
effectiveness of using a celebrity vs. noncelebrity endorser. Pham (1993) replicates and
extends this finding. From these results it is possible to develop the following
propositions:


202
Simon, Carol J. and Mary W. Sullivan (1993) The measurement and determinants of
brand equity: a financial approach. Marketing Science. 2, 1, 28-52
Sissors, Jack Z. and Jim Surmanek (1986) Advertising media planning Lincolnwood, IL:
NTC Business Books
Shavitt, Sharon (1990) The role of attitude objects in attitude functions. Journal of
Experimental Social Psychology, 26, 124-148
Sheinin, Daniel A, and Bemd H. Schmitt (1994) Extending brands with new product
concepts: the role of category attribute congruity, brand affect, and brand
breadth. Journal of Business Research. 31, 1-10
Smith, Daniel C., and C. Whan Park (1992) The effect of brand extensions on market
share and advertising efficiency Journal of Marketing Research. 29, 3, 296-313
Smith, Edward E. and Douglas L. Medin (1981) Categories and concepts. Cambridge,
MA: Harvard University Press.
Speck, Paul S., David W. Schumann, and Craig Thompson (1988) Celebrity
endorsements scripts, schema, and roles: theoretical framework and preliminary
tests Advances in Consumer Research. Vol. 15 Michael J. Houston, ed. Provo,
UT: Association for Consumer Research, 69-76
Stipp, Horst, and Nicholas P. Schiavone (1996) Modeling the impact of Olympic
sponsorship on corporate image. Journal of Advertising Research. 36,4, 22-28
Sunde, Lorraine, and Roderick J. Brodie Consumer evaluation of brand extensions:
further empirical results. International Journal of Research in Marketing. 10,
47-53
Thomas, Howard (1985) Sponsorship, an advertisers guide. International Journal of
Advertising. 4, 319-326
Thwaites, Des (1995) Professional football sponsorship-profit or profligate.
International Journal of Advertising. 14, 149-164
Tversky, Amos (1977) Features of similarity. Psychological Review. 84, 4, 327-352
and Itamar Gati (1982) Similarity, separability, and the triangle inequality.
Psychological Review. 89, 2, 123-154
Varadarajan, Rajan P., and Anil Menon (1988) Cause related marketing: a coalignment of
marketing strategy and corporate philanthropy. Journal of Marketing. 52, 3, 58-
74


107
Immediately afterward subjects were instructed to turn to the following page
which contained seven thought listing boxes, identical to those on the first thought listing
page. Subjects were given 90 seconds to list their thoughts, this time, specifically about
the target brand. At the end of 90 seconds, they were instructed to finish working on
your current thought, but do not go onto an additional box. When all subjects had
finished their last thought, they were instructed to turn to the following pages which
contained the closed ended inference measures.
At their own pace, subjects evaluated the target brand as to their perception of its
size, legitimacy, quality and its facilitation of the event. Manipulation checks for event size
and brand/event fit were then taken. Finally, subjects were asked if they had ever: been to
Europe, been to Belgium, heard of Rombouts coffee (Extran Sports Drink), run in a
marathon, or run in a 1 OK race.
After completing these closed ended questions, subjects were instructed to turn
back to the thought listing pages, and to self code the thoughts they had listed in the
boxes. They were instructed to place a mark on the lines to the left of each thought box.
They were asked to place a + if the thought was positive, a if the thought was
negative, and am n if the thought was neutral. These self coded valence marks were to
used to help the coders understand what the subjects might have meant in ambiguous
situations. After completion of the self coding task, the subjects filled out paperwork
associated with the extra credit and were allowed to go.


70
Positioning Fit
Positioning fit refers to the fit between the message of the commercial association
and the message of the rest of the integrated marketing communications campaign. To the
extent that this is explained by the similarity between the audience for an event and the
audience for traditional media vehicles employed by the brand, positioning fit will be
highly correlated with audience fit. But positioning fit also has a temporal component, and
a sub-segmentation component that make it worth discussing as a unique type of fit. The
temporal component refers to the fact that the marketing campaign for some products may
reflect certain seasonal characteristics. This makes the timing of events a determinant of
positioning fit. The sub-segment component refers to the fact that marketers often give
slightly different messages about their product to different sub-segments. This later
component is related to the unique ability of sponsorship to engage in partial
positioning, which is defined below. First, however, a brief summary of the literature
relevant to the positioning fit concept is presented.
Very little has been said about positioning fit as a determinant of success in
commercial associations. One exception to this is an article by Crimmins and Horn (1996)
who discuss the role of sponsorship within the overall marketing campaign. They stress
how important it is to check the fit between the sponsored property and your marketing
objectives (p.20). This is perhaps the closest thing to a direct reference to positioning fit
in the existing literature.
Within the related domain of corporate image management, Gray and Smeltzer
(1985) discuss the importance of having all of a companys communications have a


168
Impact of Audience Involvement and Knowledge on Inference Making
One area which has been left somewhat unresolved is the issue of how involvement
and knowledge impact inference making. Given that absolute levels of inference making
need to be calculated, the role of differential knowledge and involvement cannot be
overlooked. Studies 2a and 2b measured both constructs at an individual level. Field
research, however, will probably have to rely upon average audience involvement and
knowledge across an event. This more aggregate measure, while less valuable for
investigating specific mechanisms, should prove quite adequate as an input to predicting
event-wide rates of inference making. This level of abstraction will be appropriate to the
task of calculating absolute rates of inference making across each event.
One interesting finding from the individual level data must be considered in the
aggregate. Study 2a revealed that within the sample, low knowledge was highly correlated
with lower perceived event size. The possibility that people low in knowledge about an
event generally perceive it to be smaller than high knowledge people is intriguing. This
finding provides a new insight into the relationship between knowledge and event
facilitation inferences. Future research should investigate when event knowledge is and is
not correlated with size. It is quite possible that people low in knowledge about events
generally perceive them to be smaller than high knowledge people. Determining the
frequency with which this is true will allow a deeper understanding of the role of domain
knowledge in inference making.


166
impact varies from brand to brand. Some general guidelines for generating brand size
vs. event facilitation inferences were presented above. Such general guidelines are only
a first step towards more rigorous analysis which should eventually make more specific
recommendations possible. The additional information which is needed before such
analysis will be possible is presented below.
Absolute Measures of Inference Making
Studies la, lb, lc,2a, and 2b provided some absolute measures of inference making.
These studies support the notion that sponsorship associations can lead to inferences
about the brand. It is not known, however, how the level of inference making exhibited in
this particular experimental setting compares with inference making from more intrusive
exposures to associations such as event promotional messages, mass media coverage and
actual event attendance.
The absolute amount of inference making in the five studies is easy to calculate.
Across the five studies which measured inference making (la,lb,lc,2a,and 2b) 99 out of
310 subjects (31.93%) made at least one inference about the size, legitimacy, event
facilitation or commercial motive of the sponsoring brand. This level of inference making
is fairly uniform across the five experiments, from a low of 25% in lb, to a high of 38% in
lc. It could be argued, however, that this number is inflated due to the task demand
involved with asking subjects to list thoughts. That is, consumers might not make as many
inferences when they are not asked to list their thoughts.
It is quite possible, however, that the amount of spontaneous inference making in
these experiments is much lower than would occur with actual sponsorship. It should be


INTRODUCTION
As a promotional tool, sponsorship is unique. It is, perhaps, the only means of
communicating with customers without the benefit of a verbal message. Any message
received must be derived from the association in the mind of a consumer. While
traditional advertising often employs simple visual associations between objects and
products, the headline, body copy and/or audio usually help to explain what the
association means. With sponsorship, however, the association is the entire message. For
this reason, the study of sponsorship involves the development and testing of constructs
which ar e quite different from those employed in the study of traditional advertising.
Before going any further, it is necessary to define what is meant by sponsorship.
Sponsorship is a commercial association of a brand with a sports teams or sporting event,
entertainment tour or attraction, festival, fair or annual event, charitable cause, the arts, or
cultural institutions (Andrews 1996). While purchasing advertising time on TV and radio
has sometimes been referred to as sponsoring, buying space in traditional media (i.e.
TV, radio, print) is not what is meant by sponsorship here.
Spending on sponsorship (as defined above) has grown more than fivefold in the
last ten years, and is expected to exceed $5 billion this year in North America alone
(Andrews, 1996). Currently, the bulk of this sum (66%) is spent on sports sponsorship,
but arts and cultural sponsorships are growing at the fastest rate. Overall, spending on
1


65
context of sponsorship, an event rather than a category could be cued. Brands for which
the event cue resulted in facilitation of recognition could be said to have better attribute
fit.
Although all these measures could be used as surrogates for attribute fit, it seems
that all three techniques would be poor direct operationalizations of the attribute fit
construct. Such measures would really describe how successful a previous association had
been. This might be strongly influenced by things like duration of the sponsorship, and
amount of money spent on promoting awareness of the association. While these things
have been proposed to impact attribute fit, they are not truly components of attribute fit.
As mentioned at the beginning of the chapter, one measure against which others
could be compared would be some sort of respondent-defined, generic fit measure. Direct
measures of generic fit could be measured along continuous scales. Correlations between
these respondent-defined measures of generic fit and other measures mentioned above
would certainly help to illuminate the properties of attribute fit.
Audience Fit
One type of fit which is more relevant to sponsorship than other commercial
associations is audience fit. Audience fit refers to the nature of the overlap between the
audience for the event and target market for the brand. As mentioned earlier, audience fit,
in many circumstances, will be highly correlated with other types of fit. It does, however,
have some interesting properties with unique implications for several of the seven
mechanisms. Before describing these properties in detail, a brief review of the notion of
audience fit in the sponsorship literature is presented.


157
impact on one brand vs. another. Given that sponsorship has been shown to influence
perceived brand size and perceived event facilitation, the findings of this study
demonstrate a potential difference in the value of a sponsorship between two brands in the
same category.
Overall Discussion of Studies 3a-3c.
The goal of this chapter was to determine whether attribute perceptions, which had
been shown to be impacted by sponsorship, would have different impacts on choice for
different brands. The three studies were set up so as to detect differential impact in
different categories, as well as on different brands within a single category. A third, and
unexpected differentiating factor, market structure, may also have an impact on the
importance of various attribute perceptions.
A comparison of studies 3 a and 3b demonstrates that an attribute which can have
its perceptions shaped by sponsorship (i.e. event facilitation) can have an impact on one
product category (yogurt), but not on another (auto insurance). This finding supports the
notion that managers need to consider the overall needs of their category when setting
sponsorship strategy.
Study 3b demonstrates that an attribute which can have its perceptions shaped by
sponsorship (i.e. company size) can have an impact on one brand (Kobayama), but not on
another (Wilson). This finding supports the notion that managers must also consider the
specific situation of their brand when developing a sponsorship strategy.
A comparison of studies 3 a and 3 c demonstrates that an attribute which can have
its perceptions shaped by sponsorship (i.e. company size) can have an impact under one


187
EXPERIMENT IB ALL LOW
Leuven City Library Literature
Festival and Poetry Competition
Leuven Belgium
October 241h-26,h 1997
Speakers
Include:
Wilfried
Vanhonacker
Head libraran
Luk Warlop
Poetry Club
President
and
Christoph Van
Den Bulte
Literature Professor.
Events Include:
Student Poetry
Readings
Amateur
Performance Art
and
Presentations of
the Leuven
Poetry Club
Award Ribbons
Sponsors:
Proud to be a New
Sponsor


117
Brandy event fit
Pretesting was conducted to find a category of event which had significantly higher
fit with coffee than with sports drinks. Brainstorming led to a list of potential events.
Fourteen subjects rated perceived fit between each product category and both coffee and
sports drink. Literature festivals met this criterion (F=71.7, P<001). Mean perceived fit
with coffee was 5.71, whereas mean perceived fit with sports drinks was 1.64.
Event size
Manipulation of event size was accomplished as similarly to that used in study la
as possible. Posters for two fictitious events were created. They were pretested for
perceived size by separate groups of subjects, drawn from the same population as the main
experiment. Eighteen subjects rated their perception of the size of The European Union
International Literature Festival and Poetry Competition, along a seven point scale
anchored by a very small event and a very large event (mean=5.1) Eighteen others
rated the Leuven City Library Literature Festival and Poetry Competition along the
same scale (mean=4.6). The difference between these two was not statistically significant.
In order to increase the strength of the manipulation, additional information about
the events was included in the posters. The poster for the small event listed the speakers as
Wilfried Vanhonacker(Head Librarian), Luk Warlop (Poetry Club President) and
Christophe Van Den Bulte (Literature Professor). The poster for the large event listed
Her Royal Highness Queen Fabiola, Kurt Vonnegut, Hugo Claus (Poet laureate of
Belgium) and Seamus Heaney (Winner of the 1995 Nobel Prize in Literature).


104
TABLE 4
MEAN FAMILIARITY WITH VARIOUS EUROPEAN BRANDS
Mean*
St Dev
Max Ice Cream
1.05
0.22
Feast Ice Cream
1.05
0.22
Vescovi Caffe'
1.05
0.22
Rombouts Coffee
1.05
0.22
Jacobs Kaffee
1.05
0.22
Rader Candy Bar
1.05
0.22
Verkade Candy Bar
1.05
0.22
Dextro Energy Bar
1.10
0.31
Brao Caffe'
1.10
0.31
Douwe Egberts Koffie
1.10
0.31
Kanis & Gunnink Koffie
1.10
0.31
Simba Snacks
1.15
0.37
Coffee Rio Candy
1.15
0.37
Cup O Gold Candy
1.20
0.52
Callebaut Candy Bar
1.20
0.52
Droste Candy Bar
1.20
0.70
Extran Energy Bar
1.25
0.72
Torrefazione Italia (coffee)
1.25
0.72
Max Havelanr Koffie
1.25
0.72
Milka Candy Bar
1.25
0.55
Extran Sports Drink
1.35
0.81
Glycomax Sports Drink
1.35
0.75
Magnum Ice Cream
1.35
0.75
Leo Candy Bar
1.35
0.93
Dextro Energy Drink
1.40
1.35
Cometto Ice Cream
1.40
1.35
P-Nuttles Candy
1.50
1.19
agio Ice Cream
1.55
1.28
Whirty pop Candy
1.55
1.10
Zip Candy Bar
1.65
1.18
B Molino Coffee Roasters
1.70
1.38
Fairtime Taffy
1.70
1.17
Bella Caf
1.80
1.58
Solero Ice Cream
1.85
1.76
Baker Street Snacks
2.00
1.62
Calippo Ice Cream
2.05
2.52
Amanti Coffee
2.10
1.77
Lindt Candy Bar
2.15
2.18
B Presidente Coffee
2.50
1.88
Bounty Candy Bar
2.60
2.16
*l=never heard of it; 7=very familiar


69
defined by the intersection of brand and event audiences representing a larger percentage
of the event audience than the brand audience. Brand subsumption is defined by the
intersection representing a higher percentage of the brand audience than the event
audience. Percent subsumption is simply the percentage of the subsumed associate
represented by the intersection.
Looking briefly at the right side of the above figure, it should be clear that there
are several differences in the nature of the 100,000 person intersections presented. In the
case of equivalent, low subsumption, the intersection of Indy Car People and GMC
People is not particularly predictive of the attributes of either group. In the case of Event
subsumption, the relatively high percentage of people watching the Mud Bog race who
are also in the pickup truck target market may cause people in the intersection to
increasingly believe that GMC Trucks are for people like me. Likewise, in the case of
brand subsumption, the high percentage of people in the high performance carburetor
market, who also watch the Indy 500, may lead people in the intersection to increasingly
believe that the Indy 500 is for people like me. In other words, when there is sufficient
percent of subsumption, affiliation with the smaller, more individuating associate should
flow to the larger, less individuating associate. From these ideas, the following
propositions can be made:
P67: Sponsorship under event subsumption leads to greater affiliation with
the brand than sponsorship under equivalent or brand subsumption
P(58: Sponsorship under brand subsumption leads to greater affiliation with
the event than sponsorship under equivalent or event subsumption


121
Low Fit
High Fit
Study la
Running
Rombouts <
£ Extran
Event
Coffee
Sports Drink
Study lb
Extran
> Rombouts
Literature
Sports Drink
Coffee
FIGURE 6
DIFFERENCES IN TOTAL INFERENCE MAKING BETWEEN HIGH AND LOW FIT
GROUPS FOR STUDIES 1A AND IB
Study la seems to support a positive impact of fit on inference making. Study lb
seems to support a negative impact of fit on inference making. The parsimonious
explanation for this pattern of data is that some aspect of Extran Sports drink, and not
brand/event fit, is driving the differences in inference making. Without the second study,
researchers in the domain of sponsorship might have begun to look for boundary
conditions (i.e. when fit facilitates inferences and when it does not). Fortunately, the
intentionally conservative design employed in these first two studies has eliminated any
chance of following such a false path.
There are any number of reasons why one might expect more inferences to be
made about a sports drink than a coffee. Perhaps there are more attributes to think about.
Perhaps encountering an unfamiliar sports drink is more unusual than encountering an
unfamiliar coffee. Whatever the reason, it seems clear that the particular brand, not fit
with the event, is not driving the number of inferences made.


194
Bawa, Kapil, Jane T. Landwehr, and Aradhna Krishna (1989) Consumer response to
retailers' marketing environments: an analysis of coffee purchase data. Journal of
Retailing. 65, 4, 471-495
Belk, Russell W., and Melanie Wallendorf (1990) The sacred meanings of money.
Journal of Economic Psychology. 11, 1, 35-67
, Melanie Wallendorf, and John F. Sherry Jr.(1989) The sacred and the profane
in consumer behavior: Theodicy on the Odyssey. Journal of Consumer Research.
16, 1, 1-38
Berger, Seymour (1965) Persons as mediators: a four-stage paradigm. Journal of
Personality & Social Psychology, 2, 3, 467-469
Bhattacharya, C B., Hayagreeva Rao, and Mary Ann Glynn (1995) Understanding the
bond of identification: an investigation of its correlates among art museum
members. Journal of Marketing, 59, 4, 46-57
Bimbaum, Michael H., Rebecca Wong, and Leighton K. Wong (1976) Combining
information from sources that vary in credibility. Memory & Cognition, 4, 3,
330-336
Blackston, Max. (1995) The qualitative dimension of brand equity. Journal of
Advertising Research. 35, 4, RC-2-7
Bloch, Peter H. (1986) The product enthusiast: implications for marketing strategy
Journal of Consumer Marketing. 3,3,51-62
Boulding, William, and Amna Kirmani (1993) A consumer-side experimental examination
of signaling theory : do consumers perceive warranties as signals of quality?
Journal of Consumer Research. 20, 1, 111-123
Brehm, Jack W., and David Lipsher (1959) Communicator-communicatee discrepancy
and perceived communicator trustworthiness. Journal of Personality. 27, 352-361
Brewer, Geoffrey (1993) Here's good news for advertisers. Sales & Marketing
Management. 145, 15, 20
Broniarczyk, Susan M., and Joseph W. Alba, (1994) The importance of the brand in
brand extension. Journal of Marketing Research. 31 (May), pg. 214-228
Bultez, Alain., Philippe Naert, Els Gijsbrechts and Piet Vanden Abeele (1989)
Asymmetric cannibalism in retail assortments. Journal of Retailing. 65, 2, 153-
192


78
common features in judgment of similarity than in judgment of difference (Tversky, 1977,
p.340), it is possible to propose:
P74: Under perceived sacrilege, attributes common to event and brand
become less salient.
P7S: Under perceived sacrilege, attributes not shared by event and brand
become more salient
Therefore,
P75: Under perceived sacrilege, attribute fit is reduced.
The elaborations generated by perceived sacrilege may well be the commercial
association equivalent of counter-arguing. This is serious, because one of the benefits of
commercial associations in general is the inability of consumers to counter-argue in the
classic sense, when the association is the message. It is therefore possible to propose
that:
P77: Sacrilege-inducing associations will lead to more counterarguing than
will non-sacrilege-inducing associations.
Conclusion
While consumer outrage at associations considered to be sacrilegious is understudied, the
growing number and forms of commercial association make it an especially salient avenue
for further research. As the boundaries of what can be sponsored get pushed farther and
farther, the importance of this area of investigation should only increase. One way to
minimize the chance of negative consumer response to an association is through additional


67
Finally, there has been some elaboration on the notion that there are several
distinct event audiences which may respond differently to sponsorship. Hastings (1984)
makes a distinction between audience members who are directly involved in the
sponsored events as participants, those who are directly involved as personal
spectators, and those who follow the event through the media. The author notes that
event sponsorship differs from advertising in the clear delineation between various
audiences for an event. He points out that these differences must be taken into account
when creating potential objectives for a sponsorship. These different event audiences
could probably serve as a surrogate for an event-involvement construct which may have
an impact on the functioning of the seven mechanisms. Specifically, it can be proposed
that
P66: Event-involvement is highest for participants, lower for those who
attend the event live, and lower still for those who view the event through the
amass media.
Apart from these quotes and ideas, there is no extant elaboration on the audience
fit construct. There has been no discussion of its nature, properties, or impacts on
sponsorship success. This shortcoming is addressed below.
Audience fit involves two distinct dimensions. The first, which is widely
recognized in other contexts, is intersection, which simply refers to the size of the
overlap in the audience of the event and the target market for the sponsor. Intersection is
analogous to the media planning concept of reach (see Sissors and Surmanek, 1986).
As far as the intersection component of audience fit is concerned, an overlap of the target
market for brand X with event Y of 100,000 is equivalent in value to an overlap of the


195
Carpenter, Phil M. (1994) Some cobranding caveats to obey. Marketing News. (Nov
7), 28, 23, 4
Celsi, Richard L., Simeon Chow, Jerry C. Olson and Beth A. Walker (1992) The
construct validity of intrinsic sources of personal relevance: an intra-individual
source of felt involvement Journal of Business Research, 25, 2, 165-185.
Chaiken, Shelly (1979) Communicator physical attractiveness and persuasion Journal of
Personality and Social Psychology, 37, 8, 1387-1397
Chew, Fiona (1992) The advertising value of making possible a public television
program. Journal of Advertising Research. 36, 6, 47-52
Cohen, Jacob, (1977) Statistical power analysis for the behavioral sciences.
Orlando: Academic Press
Cohen, Joel B., and Kunal Basu (1987) Alternative models of categorization: toward a
contingent processing framework. Journal of Consumer Research. 13,4, 455-472
Crawford, C. Merle (1985) A new positioning typology. Journal of Product Innovation
Management. 2, 4, 243-253
Crimmins, James, and Martin Horn (1996) Sponsorship: from management ego trip to
marketing success. Journal of Advertising Research. 36, 4, 11-21
Crowley, Martin G. (1991) Prioritizing the sponsorship audience. European Journal of
Marketing. 25,11, 11-21
Dawar, Niraj, and Paul F. Anderson (1994) The effects of order and direction on
multiple brand extensions. Journal of Business research. 30, 119-129
DeCoursey, Jennifer (1995) Papal paraphernalia proffered. Advertising Age, 66, 41, 2
Deshpande. Rohit, Wayne D. Hoyer, and Naveen Donthu (1986) The intensity of ethnic
affiliation: a study of the sociology of hispanic consumption. Journal of Consumer
Research. 13, 2, 214-220
DiMingo, Edward P. (1988) The fine art of positioning. Journal of Business Strategy, 9,
2, 34-38
Ellen, Pam Scholder, Joshua Lyle Wiener, and Cathy Cobb-Walgren (1991) The role of
perceived consumer effectiveness in motivating environmentally conscious
behaviors. Journal of Public Policy & Marketing. 10,2,102-117


171
What do you think the quality level of Extran Sports Drink (Rombouts Coffee) is likely
to be?
Very poor quality | ]_ | | j | | | Very high quality
How important do you think the support of Extran Sports Drink (Rombouts Coffee) is to
the sponsored event taking place?
Not important at all | | | | | | | | Very important
How important do you think the support of Extran Sports Drink (Rombouts Coffee) is to
the success of the sponsored event?
Not important at all | | | | | | | | Very important
How many events do you think Extran Sports Drink (Rombouts Coffee) sponsors each
year?
Very few events | | | | | | | | Lots of events
Compared with other brands of sports drink (coffee) how many events do you think
Extran (Rombouts) sponsors
Far fewer than others | | ¡ |_ | | | | Many more than others
Overall, how large do you perceive the running (literature) event advertised to be?
A very small event | _| J | | | | | A very large event
Overall, how would you rate the fit between the running (literature) event advertised in
the poster and Extran Sports Drink (Rombouts Coffee)
Very poor fit | | | | | | | | Very good fit


186
EXPERIMENT IB ALL HIGH
International Literature Festival
and Poetry Competition
Brussels, Belgium
October 24*-26T 1997
The European Union
Attended by:
Her Royal
Highness Queen
Fabiola
Speakers include:
Kurt Vonnegut
Hugo Claus
Poet Laureate of
Belgium
and
Seamus Heaney
Winner of the
Events Include:
Live Poetry
Readings
Performance Art
Happenings
and
Presentations of
European Union
Poetry
Scholarships
Sponsors:
Proud to be a sponsor
for 5 Years
Coolkat Books
Generate Bank


APPENDIX F SAMPLE CHOICE PROFILE USED IN
EXPERIMENT 3B 181
APPENDIX G SAMPLE CHOICE PROFILE USED IN
EXPERIMENT 3C 182
APPENDIX H SAMPLE EXPOSURE STIMULI FOR
EXPERIMENTS 1A, IB, 1C, 2A, AND 2B 183
APPENDIX I ONE POSSIBLE FORM OF A SPONSORSHIP
VALUATION FUNCTION FOR A SPECIFIC BRAND 192
REFERENCES 193
BIOGRAPHICAL SKETCH 204
vi


CHAPTER ELEVEN
CONCLUSION, IMPLICATIONS, AND FUTURE RESEARCH
The way in which sponsorship impacts consumers is not at all obvious. If you
were to ask a consumer who attended an event, what impact did sponsorship have on
you the likely answer would be none. Should the conclusion be drawn, therefore, that
that the six billion dollars spent per year on sponsorship is wasted money? The results of
the eight studies presented in this dissertation would argue definitely not.
It is easy for a consumer to say that an advertising campaign, a sponsorship, or
even a persuasive salesperson had no influence on his/her decision. Because everyone
has experience as a consumer, it is easy to extrapolate a feeling of no influence on me to
a theory of no influence on anyone. For both personal selling and advertising, there is,
however, a long history of a measurable impact on sales. This body of evidence makes it
easy to prevent the no impact on me feeling from guiding beliefs about aggregate
effects. Such a long history does not exist for sponsorship. It is, therefore, harder to
refute the assumption that the impact of sponsorship on behavior is limited.
The companies pouring the billions into sponsorship must believe that they are
getting something in return. Reading the trade press and talking to practitioners has made
it clear, however, that there is no consensus about the mechanisms through which
159


118
Thirty-six additional subjects were exposed to one of the two revised posters and
rated their perception of event size as described above. With the strengthened
manipulation The European Union International Literature Festival and Poetry
Competition had a mean rating of 5.7 whereas the Leuven City Library Literature
Festival and Poetry Competition had a mean rating of 3.2. This difference is significant
(F=12.78, p<01). These two posters were then used for the main experiment.
Duration of association
Duration of association was manipulated as described previously.
Brand logo size and placement
Brand logos were the same size as in the low share-of-presence conditions in study
la. Due to the nature of the event, one of the filler sponsors(Adidas) was replaced by a
more logical associate (Coolkat Books).
Results
Manipulation Checks
As before, manipulation checks on perceived size and perceived brand/event fit
were taken after all other measures, except for demographic information. Using the same
scales as before, the difference between large size (mean=5.35) and small size
(mean=3.42) was significant (F=29.5, p<001). The difference between high fit (5.25) and
low fit (mean=2.55) was also significant (F=59, pc.001).


Ill
real of a brand do you perceive Extran (Rombouts) to be?. The scale for this question
was anchored by not very real and very real. None of the independent variables had
any significant impact on either of these two variables.
The simple logic that levels of gatekeeping by the event could translate into
legitimacy for a sponsoring brand made exploring legitimacy inferences, as opposed to
overall quality inferences, appealing Because it was anticipated, however, that the
concept of legitimacy would be difficult to convey to subjects in survey form, a direct
measure of perceived quality was also taken. This question asked what do you think the
quality level of Extran (Rombouts) is likely to be? The scale was anchored by very poor
quality and very high quality.
Regression of the independent variables on perceived quality yielded one
significant main effect. Event size significantly predicted perceived quality (F=5.5, p<03).
This was in the predicted direction. Brands associated with large events were perceived to
be of higher quality than were brands associated with small events. This supports H5.
Support for H8 was not found, however, in that fit did not moderate this effect. The
Interaction of event size with fit had no impact on perceived quality (F=l. 1, P>.25). No
support was found for H6 or H7. Neither duration nor share had any measurable influence
on perceived quality (F<1 for both).
Event Facilitation Inferences
Two measures of event facilitation inferences were taken. The first, called event
enabling, attempted to measure inferred importance of the brand to the event taking place


71
coherent message. Specifically, they state that the various sources that communicate
corporate image must be coordinated so that a congruent and effective image is projected
(p. 77). While sponsorship associations are not directly cited, their argument clearly
could include sponsorship as an image-communicating device.
Finally, within the domain of brand extensions, Park, Milberg and Lawson (1991)
point out the value of having all of a companys brands share a coherent position. They
state that the degree of perceived fit is a function of among other things, brand-
concept-consistency perceptions (p. 186). They also find that extensions work best when
both core brand and extension are either prestige or functional in nature. Therefore, at a
very high level of abstraction, two associates having a similar positioning seem to be an
indicator of success in some circumstances.
As mentioned in the introduction to this section, one reason why positioning fit is
particularly relevant to sponsorship is the temporal nature of events. An example of a
brand which takes on a somewhat seasonal positioning strategy is Coors Light In the
Summer, Coors Light is positioned as fun time beer to take with you to the beach so you
wont be too full to play volleyball. In the winter, the brand position seems to move
toward being a cozy, by the fire, after-ski beverage. In short, the changing seasonal
consumption situations for the product are reflected in the advertising of the product. By
sponsoring beach volleyball in the summer and downhill ski racing in the winter, the
seasonal positioning is reflected in the sponsorships chosen.
A related, and perhaps more interesting facet of the positioning fit between event
and brand is the sub-segment component. Much like some brands have seasonal
positioning strategies, others try to position their brands as different things to different


60
the more metaphorical is a comparison, the more asymmetry there will be (e.g., cigarettes
are like bullets, but bullets are not like cigarettes). This is the idea which led to the quote
..the degree of symmetry is inversely related to the degree of metaphoricity( p. 172).
Ortony suggests that when no attribute of the second term in a comparison seems
to apply to the first, that a speaker (or listener, reader or writer) may reorder the salience
of the attributes of (especially) the second term in the comparison. (Ortony, 1979,
p. 173). This may have relevance to sponsorship. If one equates the metaphoricity
construct with the levels of abstraction of the association, then it is possible to make some
interesting propositions regarding levels of abstraction and attribute salience.
For example, when Porsche cars are compared to race cars, the comparison is
quite literal, and the salient attribute of a race car (speed) can be easily linked to the brand
Porsche. When McDonalds sponsors a race, however, the literal similarity is quite low,
and the desirable fast attribute is less likely to transfer. When a full-size pickup truck
sponsors a race car, however, the most salient attribute of the race car, speed, is not
relevant, and a reordering may take place until a match, such as can take abuse is found.
If this takes place, not only should the toughness of the truck be made salient, but the
toughness of the race cars should be made salient as well. Based upon Ortonys (1979)
results, the following propositions can be made:
P59: When attribute fit between event and brand is the result of concrete
attribute overlap, metaphorical elaboration does not enhance attribute fit.
P60: When attribute fit between event and brand is the result of abstract
attribute overlap, metaphorical elaboration enhances attribute fit.


133
Low Fit
High Fit
Study 2a
U.S Open
(Tennis)
Texaco
Evian
Study 2b
Indy 500
Evian
Texaco
FIGURE 8
DESIGN FOR STUDIES 2A AND 2B
Study 2a
Overview
Subjects are exposed to an actual high or low fit sponsorship association by
viewing a poster for the US Open Tennis Tournament. The poster featured either a high
fit or a low fit sponsoring brand. They then responded to open and closed ended measures
of inference making about the sponsoring brand (similar to studies la-lc). They then rate
their involvement with, and subjective knowledge of tennis. Finally, they are quizzed on
their actual knowledge of the rules of tennis.
Subjects
Thirty-eight undergraduates in an introductory marketing class at a major
southeastern university took part. Each received course extra credit for participation.


83
Sawyer arid Howard (1991) demonstrated that for high involvement individuals, open-
ended advertisements were superior to closed-ended ads in terms of attitudes, purchase
intention, and choice. The authors found no significant difference, however, for the low
involvement groups.
These studies seem to contradict early work which found that explicitly drawing
the conclusion for the audience was superior to letting them draw the conclusion
themselves. Hovland and Mandel (1952), for example, found that over twice as many
subjects agreed with a speakers argument when the conclusion was explicitly drawn by
the speaker as opposed to being left up to the audience (neither intelligence nor
personality type had a measurable impact on this effect). The theoretical explanation for
this seeming inconsistency is that the impact of omitting the conclusion depends upon how
difficult the conclusion is to draw. If it is very difficult, then few people will draw the
appropriate inference, and effectiveness is reduced. If people are capable of getting it
on their own, however, omitting the conclusion should be more effective.
Several authors have hypothesized as to why this might be so. Linder and Worchel
(1970) demonstrate how mechanisms like cognitive dissonance can lead to difficult self
drawn conclusions being believed more than conclusions which were easier to draw. One
explanation for this finding is that people want (or need) to find value in accomplishing a
more difficult task, and that discovering something with which one agrees is more valuable
than discovering something with which one does not agree.
Similar ideas have been put forth in the domain of categorization, which is more
closely related to the pairing of commercial associates. Cohen and Basu (1987), for
example, propose a direct impact on affect of discerning what two things have to do with


A series of five studies investigates the role played by event, associational and
audience factors on inference making about a sponsoring brand. Evidence is found that
brand size inferences are best created through sponsoring large events for a long time,
whereas event facilitation inferences are best created through being the title sponsor of
smaller events. Three additional studies demonstrate that these two inferences (brand
size and event facilitation) have a different impact on the choice probabilities of different
categories and brands.
viii


APPENDIX E
SAMPLE CHOICE PROFILE USED IN EXPERIMENT 3 A
Davis Auto
insurance
Wilson Auto
Insurance
Customer service
Hours
8AM-6PM
Mon-Sat
24 Hours 7
days/week
Company Size
Large
Company
Small
Company
Company involvement
with favorite local events
Not involved
Helps make
them possible
Deductible
$1000
$500
Price for 6 months
coverage
$570.00
$615.00
U
u
Please indicate your
choice (check only one)


180


203
Walker, Mary, Lynn Langmeyer, and Daniel Langmeyer (1992) Celebrity endorsers: do
you get what you pay for? Journal of Consumer Marketing. 9, 2, 69-76
Walshe, Peter and Peter Wilkinson (1994) Pompeii revisited: IBM digging for success.
Marketing and Research Today. 22, 1, 89-96
Weisz, Pam (1995) G-P's Angel Soft gets kid-conscious. Brandweek. 6, 39, 4
Westbrook, Robert A., and William C. Black (1985) A motivation-based shopper
typology. Journal of Retailing. 61, 1, 78-103
Williams, Michael (1996) Gen X files: how sony uses sponsorship to reach previously
unattainable goals. Paper presented at the 1995 IEG Event Marketing
Conference, Chicago Illinois.
Withcer, B., G. Craigen, D. Culligan, and A. Harvey (1991) The link between objectives
and functions in organization sponsorship International Journal of Advertising.
10, 1, 10-33
Wolton, Claude (1988) Arts sponsorship: harmony or discord? European Research.
May, 87-96
Worthington, Steve, and Suzanne Home (1992) Affinity credit cards: card issuer
strategies and affinity group aspirations. International Journal of Bank
Marketing. 10, 7, 3-10
Wright, Peter (1985) Schemer schema: consumers intuitive theories about marketers
influence tactics in Advances in Consumer Research. Vol. 13, ed. Richard J.
Lutz, ed. Provo, UT: Association for Consumer research, 1-3
Wright, Rodger (1988) Measuring awareness of British football sponsorship. European
Research. May, p.104-108
Zaichkowsky, Judith Lynne (1985) Measuring the involvement construct. Journal of
Consumer Research. 12, 3, 341-352.
Zajonc, Robert B. (1968) Attitudinal effects of mere exposure. Journal of Personality
and Social Psychology 9, 2, 1-27.
Zinkhan, George M., Jae W. Hong and Robert. Lawson (1990) Achievement and
affiliation motivation: changing patterns in social values as represented in american
advertising. Journal of Business Research. 20, 2, 135-143


87
These mechanisms are so far up the necessary cognitive elaboration continuum
that they all strongly rely upon consumer inference making in order to function. For a
sponsorship association to impact the perceived size of a company or brand, a consumer
must make an inference that this company must be big in order to be able to sponsor such
an event. In order for the implied endorsement mechanism to be engaged, a consumer
must make the inference that this brand must be OK, or else it wouldnt be allowed to be
the sponsor of this event. In order for reciprocity to be engaged, a consumer must make
the inference that this brand is doing something to make possible and/or better an event I
care about.
The realization that inference making is key to so many important, uniquely
sponsorship-related mechanisms has led to the development and testing of the inference
based model of building brand equity through sponsorship. This model yields many
specific hypotheses about how audience, event, and associational factors impact the
inferences consumers draw about sponsoring brands, and how these inferences impact
choice probabilities for various classes of brands. The inference-based model is depicted
graphically in Figure 3.
As the figure makes clear, the inference-based model consists of two stages. In
stage one the consumer who is exposed to the sponsorship association makes inferences
about the brand. The number and kind of inferences made are determined by event and
associational factors, as well as by characteristics of the audience members relationships
with the event (audience factors). Stage two consists of converting these inferences into
changes in choice probability for a particular brand on the basis of the sponsorship-based
inferences. Here, categories which differ in experience ambiguity, and brands which differ


23
endows a product(p. 24). He notes that Coopers and Lybrand calculate brand equity by
comparing the premium price paid for a branded product compared to a generic one. It
should be noted that while these definitions focus on the economic value which comes
from brand equity, they do not preclude an associative component in their formation.
In fact Farquhar et al. (1992) propose ways in which the economic value of brands
can be maximized and exploited through management of associations. They develop the
concept of master brands which are very exploitable. They define these master brands
as those brands which own an association. That is, when the associate is mentioned, the
brand comes to mind. While the associates used in most of the examples are product
categories, they acknowledge that other types of brand associations include celebrity
endorsers, sponsored events, geographic affiliations, user groups, etc. (pp. 32-33,
emphasis added).
In addition to specifically including the management of commercial associations as
an important component of brand equity management, they provide insight into methods
of measurement. They suggest that associations can be determined through free
associative tasks, projective methods, and focus groups. They also claim that the process
of building brand equity consists of developing favorable associations, making these
associations accessible, and developing points of difference which can be easily
remembered by consumers.
These ideas have clear implications for sponsorship. In addition to the awareness
building contribution of sponsorship to brand equity, it seems feasible that more complex
associative mechanisms might be able to play a role as well. It should be possible to
develop functional forms for the impact of sponsorship on brand equity. These might


162
Implications
Early on, it was proposed that sponsorship was the only means of communicating
with customers without the benefit of a verbal message, and that any message received
must be derived from the association in the mind of a consumer. Some authors, however,
have suggested that the primary impact of sponsorship on consumers is simple top-of-
mind awareness (e.g. Aaker, 1991). If this were true, the only pertinent issues in
sponsorship research would be the size and demographics of each event audience.
Brands such as Sony, however, claim to use sponsorship to reach the previously
unattainable goal of making their product a bigger part of the Gen-X lifestyle (Williams
1996). LA Gear claims to have successfully used sponsorship associations to reposition
its product line (Proctor 1996). Tanqueray gin says it used sponsorship to portray a less
stodgy image (Elliot 1995).
It appears, therefore, that the practitioners making sponsorship decisions believe
that associating a brand with an event can do more than simply remind people about the
brand name. This view has been shared by a variety of authors who have used anecdotal
evidence to posit mechanisms through which sponsorship might influence brand
perceptions (Hastings, 1984; Meenaghan, 1991b ; Thomas, 1985; Crowley, 1991; Parker,
1991; McDonald, 1991; Otker, 1988; Ryssel and Stamminger,1988; Schoch, 1994;
Walshe and Wilkinson, 1994). The research presented in this dissertation, however, is the
first to find empirical support for such mechanisms. These findings, and their implication
for sponsorship strategy, are revisited below.


176
undesirable I I I I 1 [ I I desirable
wanted 1111 111! unwanted
not needed 1 I 1 j ¡ ¡ | | needed
Please rate your level of agreement with the following statements:
know a great deal about the game of tennis
strongly agree | j | | | | | |strongly disagree
I know more about the game of tennis than most of my friends
strongly agree | | | | | | | |strongly disagree
I watch most of the major tennis events that are on TV
strongly agree | | | | | | | |strongly disagree
I watch more tennis on TV than most of my friends
strongly agree | | | | | | j ¡strongly disagree
I play tennis quite frequently
strongly agree | | | | | ¡ | ¡strongly disagree
I play tennis more often than most of my friends
strongly agree | | | | ¡ | | ¡strongly disagree
I am a good tennis player
strongly agree | | | ¡ | ¡ | ¡strongly disagree
I am a better tennis player than most of my friends
strongly agree | | ¡ | | | ] ¡strongly disagree


42
after it The implied endorsement mechanism requires that, from a sponsorship
association, the consumer infers that the event is somehow endorsing the quality of the
brand. Several examples of this can be cited. Seikos sponsorship of track meets at which
it is the official timer is provided by McDonald (1991) as an example of what he calls
direct relevance(p. 36). By direct relevance is meant that one might logically conclude
that the official status of a sponsor implies some measure of quality.
Crimmins and Horn (1996) cite Visa and Seiko as brands which actually affected
consumer perceptions of quality due to their association with the Olympics (as measured
by DDB Needhams SponsorWatch service). Since Seiko was the official timekeeper
and Visa was the only credit card the Olympics would accept for tickets, it seems that an
implied endorsement may have led to the Olympic sponsorship success of these two
brands.
For the implied endorsement mechanism to function, however, does not require
some official sanctioning of the sponsoring brand. As long as a consumer infers that some
minimum quality level is required before an event will let a brand sponsor it, the implied
endorsement mechanism will be engaged. Usage of the brand at an event, however,
should increase the likelihood of this type of inference. Seeing top pool players compete
for large prizes on Brunswick tables, at the Brunswick Shoot Out for example, might
cause some consumers to infer a level of quality for Brunswick tables.
It can therefore be proposed:
P32: Use of the brand in the sponsored event facilitates the inferences
necessary for the implied endorsement to function.
P33: Having the brand name in the event name facilitates the inferences
necessary for the implied endorsement to function.


142
replicated when fit was flipped provides strong evidence that fit, and not some other
attribute of the brand is driving the effect.
This effect was not detected in study 1. It is possible that the effect was somehow
enhanced through the use of real brands and events. Perhaps the fact that subjects already
had beliefs about the size and quality of Evian and Texaco allowed more opportunity for
event facilitation inferences to be created. Perhaps the lack of other sponsors on the
poster increased event facilitation inferences in a way which allowed the impact of fit to be
detected. A reasonably parsimonious explanation for not detecting this effect in previous
studies, however, is that there are some anomalous attributes of Rombouts and Extran
which made them unique.
While it appears that fit between brand and event can lead to event facilitation
inferences, it is not clear when this would be expected. Future research should attempt to
find boundary conditions on this effect. Possible studies could examine the role of
familiarity of the brand, and the familiarity of the event, as well as the number of other
sponsors as possible moderators of this effect.
The second major finding of this set of studies is that for the US Open, knowledge
about tennis diminished the impact of fit on event facilitation inferences. That is, groups
of subjects high in tennis knowledge showed no significant difference in event facilitation
inferences for Evian vs. Texaco. Subject groups low in tennis knowledge, however, made
much greater event facilitation inferences about the high fit sponsor (Evian) than the low
fit sponsor (Texaco).
One interesting possibility is that low knowledge subjects are unaware of the large
size of the event. If low knowledge subjects think the event is not as large, they may be


153
Independent Variables
Five independent attribute variables were manipulated orthogonally in exactly the
same manner as in 3 a. Again, these attributes were price (four levels), deductible (two
levels), company size (two levels), event facilitation (two levels), and hours of customer
service (two levels). These attribute levels were identical to those in 3a.
The additional manipulated variable in this study is country-of-origin. In this
study, subjects were choosing between the Wilson auto insurance company of Atlanta,
GA, and the Kobayama auto insurance company of Japan. Pretesting revealed that auto
insurance is a product class where domestic brands were strongly preferred. The same 43
categories as in Table 7 were tested for domestic vs. foreign preference. The same 33
subjects rated these 43 categories along a 9-point scale anchored by Always prefer an
American brand and Always prefer a foreign brand. Results of this pretest are shown
below.


13
extension and celebrity endorsers, the issues addressed in this chapter are limited to the
very basic, common denominator issues and predictions.
Since the goal is to set up a broad understanding of commonalties among
commercial associations, the level of precision in this chapter is predictably broad. The
propositions put forth here, therefore, are also predictably broad. The specific details of
how sponsorship is likely to function, as well as detailed predictions about main effects
and moderators are left for later chapters.
In order to facilitate discussion of broad issues relating to commercial
associations, it is necessary to introduce a bit of vocabulary. In the discussion below,
associate can refer to either member of an association. For example, a brand can be an
associate of a celebrity, and a celebrity can be an associate of a brand. An associate of
simply means that an object is associated with something else. Acceptor, however,
refers to the object which an association is attempting to help. In celebrity endorsement,
the acceptor is usually the brand being associated with the celebrity. In a brand extension,
the acceptor is the new brand which is being associated with the old one. The provider,
on the other hand, is the object or person providing the help. In celebrity endorsement,
the celebrity is usually the provider. In a brand extension, the preexisting brand is the
provider.
P erhaps the most basic premise, across domains of commercial associations, is that
the more desirable the provider, the more beneficial the association will be to the acceptor.
Aaker and Keller (1990), for example, find that in the brand extension domain the higher
the quality of the core brand (the provider) the better the consumer acceptance of the
extension. Sunde and Brodie (1993) obtain similar results. Likewise, in the celebrity


35
As mentioned previously within the brand extension domain, Reddy, Holak, and
Bhat (1994) find that brands with high symbolic value extend better than brands with
low symbolic value. Given that in the case of sponsorship, the brand is the acceptor, one
might expect that brands with high symbolic value already would be less likely to have the
image of the event transferred onto them. Therefore it can be proposed:
P23: Transfer of image from event to brand takes place more easily, and
with greater magnitude, for low symbolic brands compared to high
symbolic brands.
Within the domain of combined brands Rao and Ruekert (1994) discuss how an
averaging model of attribute levels would lead to a zero sum game in which brand
alliances would be almost impossible to arrange. They argue in favor of a maximum
rule whereby the combined brand may receive the higher of the two brand
perceptions on each dimension. An example of this is the association between Lays
potato chips and KC Masterpiece barbecue sauce. While the combined brand gains the
desired associations of tangy and hickory smoke flavor from KC Masterpiece, an
average crispness between the chips and the sauce is not perceived. Crispness remains
at the higher of the two levels, the level of regular Lays. This leads to the following
proposition for the domain of sponsorship:
P24: Due to a maximum rule, brand images not fully congruent with an
event will not cause appreciable harm to the image of the event.
In fact, anecdotal evidence of reverse image transfer without negative impact upon
the brand has been cited. Jones and Dearsley (1989 cf. McDonald, 1991) find that the


36
image of football (soccer) was actually improved by a sponsorship by Barclays Bank
during a time when football was receiving a lot of bad press. Apparently this association
transferred stability from Barclays to the soccer league without transferring rowdiness
to Barclays. Perhaps positive associations are more easily transferred than negative ones.
This is certainly an interesting empirical question for future study.
Affiliation
Affiliation, like image transfer, can be stated as a message. The message of the
affiliation component of an association is that this brand is for people like me. The
primary ¡requirement for the functioning of this mechanism is that the event be perceived
as for people like me. The more this statement is perceived to be true, the more the
message about the brand being for me will be facilitated.
While there is no formal conceptual development of the way affiliation might
function as a mechanism of sponsorship effectiveness, some authors have hinted at the
idea. Meenaghan (1991b), for example, points out that the (sponsorship) message is
delivered by association with a socially intrusive activity (p. 8). Parker (1991) posits
that todays consumers, in addition to demanding products with desired attributes at a
reasonable price, also seek products which are for people like me. Sponsorship, he
proposes, is uniquely suited to helping a brand achieve the perception of being for
people like me. No further development of these ideas, however, is presented by either
of these authors.
There have been several studies of the impact of strong, social affiliations such
as political affiliations (Ellen, Wiener and Cobb-Walgren, 1991) and ethnic affiliations


19
goods indicates that experience goods may be more susceptible to the impact of
commercial associations than are search goods. In the context of brand alliances Rao
and Ruekert (1994) view the search-experience dichotomy as a continuum (p. 95).
They give examples of how brand name ingredients in food products which are likely to
affect taste are better brand alliances than are alliances involving products with mainly
search attributes. These studies seem to indicate that experience goods may be more
easily impacted by commercial associations than are search goods. Therefore it can be
proposed:
PI 1: All else being equal, sponsorship will be more effective for brands with
more experience attributes and fewer search attributes.
P12: Brands which are considered credence goods will be more greatly
impacted by sponsorship than will search or experience goods.
One final issue needs to be touched upon in this review of the commercial
association literature. That is the issue of the acceptor having a negative impact upon
the provider. Aaker (1990) notes that risky associations may backfire and gives
examples of several notable cases where a failed brand extension actually harmed the
image of the core product. Dawar and Anderson (1994) propose and find evidence that
an extension to a particular product may make it more difficult for the brand to extend
in a different direction in the future (p. 124). Loken and Roedder John (1993) explain
how a dilution effect (harming the provider) occurs under a bookkeeping model. This
bookkeeping model dilution effect will occur for extensions which have attributes that
are incompatible with the image of the core brand. They note that this can be avoided if


164
Event facilitation inference making was found to be enhanced by three factors.
Analysis of the pooled data between la and lc showed a significant impact of title
sponsorship on event facilitation inferences. Study la showed a positive impact of
duration of association on event facilitation inferences. Follow-up analysis in 2a, which
found event knowledge to be correlated with the perceived size of the event, supports the
notion that sponsoring smaller events leads to greater event facilitation inferences.
Therefore, if a brand is trying to improve its perception as a good corporate citizen, these
results suggest a sponsorship strategy of being the title sponsor of many, smaller events
for a long time. With this type of strategy as well, it appears that an additional positive
impact of sponsorship accrues over time.
If one were to assume, however, that each type of inference was equally valuable
to all brands, the results of the first five studies could be said to provide limited assistance
to a manager trying to assess the value of different sponsorship opportunities. Assuming
such an equal value for all brands would only support a strategy of being the title sponsor
of high fit events for a long time. Assuming different effects on different brands, however,
allows more specific, and perhaps less obvious recommendations to be made.
The results of studies 3 a, 3 b, and 3 c, in fact, support differential impact of
inferences on different brands. These studies show that the importance of individual
inferences can vary between product categories, between brands within a category, and
even between different competitive situations for a single brand. Finding that each
inference need not impact all brands uniformly makes the findings of the first five studies
much more useful. An example of how this information can be used to select between two
sponsorship opportunities is presented below in Table 12 .


57
changing the criterion of what counts as a relevant attribute. They go on to present the
use of similarity to explain categorization as analogous to citing scoring more points than
the opposing team as a reason for victory in sports. The authors posit that rather than
being a true determinant of conceptual coherence, similarity may be a mere by-product
(p. 291). While the merits of this conclusion in the context of understanding category
structure is debatable, the criticism seems less relevant to the issue at hand. As stated
above, the process of evaluating fit between a sponsor and an event is likely to be only
tangentially related to taxonomic category structure. For this reason, the usefulness of the
similarity construct in the exploration of the properties of attribute fit seems to be
reasonably whole.
In order to understand how attribute fit might moderate the seven mechanisms, it
is first important to consider the direct consequences of attribute fit. As mentioned above,
to the extent that sponsorship associations elicit an attempt on the part of a consumer to
make sense of the association through a process of categorization, the categories formed
are likely to be what Barsalou (1983) refers to as ad hoc categories. He goes on to
describe and demonstrate several outcomes of such categories. Specifically, it is
demonstrated that ad hoc categories tend to be less established in memory, and therefore
less available in memory. This means that category concepts, concept-to-category
associations and category-to-concept associations will generally be less established in
memory compared to those of common categories (Barsalou, 1983).
This seems to indicate that the weaker the attribute fit, the less the sponsorship
associations will be available in memory. Therefore, it is possible to predict that


22
He goes on to propose that marketing activity can enhance or maintain the favorability,
strength, and uniqueness of a brand. Given an association-based definition of brand
equity:
F14: Significant sponsorship activities are capable of impacting a brands
equity.
In his book on managing brand equity, Aaker (1991) devotes a good deal of time
to explaining the importance of associations to brand equity. With regard to this
importance, he suggests that the association not only exists but has a level of strength.
A link to a brand will be stronger when it is based on many experiences or exposures
rather than few. It will also be stronger when it is supported by a network of other links
(p. 109). Clearly, then, by Aakers definition as well, sponsorship associations should
sometimes be able to affect brand equity.
It should be noted, however, that Aaker has not endorsed the ability of
sponsorship to impact brand equity through building favorable associations. In fact, he
states that the primary role of most event sponsorship is to create or maintain awareness
(p. 75). Given, therefore, that sponsorship receives but a single line in the book, it does
not seem likely that Aaker had fully considered the ways in which sponsorship might
impact upon brand equity in great detail.
Other definitions of brand equity take a slightly different perspective. Simon and
Sullivan (1993), for example, have proposed a model of brand equity based upon market
valuation of the stock equities of corporations with substantial brand assets. Similarly,
Farquhar (1989) defines brand equity as the value added with which a given brand


108
Results
Manipulation Checks
Manipulation checks on perceived size and perceived brand/event fit were taken
after all other measures, except for demographic information. The check on the event size
manipulation was measured by asking subjects Overall, how large do you perceive the
running event advertised in the poster to be Responses were given on a single seven
point scale, anchored by a very small event and a very large event Mean response to
this scale were 5.13 for subjects in the large event condition and 4.32 for those in the small
event condition. This difference is significant (F=3.62, P<05), and in the appropriate
direction.
Perceived brand/event fit was measured by asking subjects Overall, how would
you rate the fit between the running event advertised in the poster and EXTRAN SPORTS
DRINK (ROMBOUTS COFFEE)? Responses were given along a seven point scale
anchored by very poor fit and very good fit. Mean response for the high fit groups
was 5.39, and for the low fit groups it was 3.32. This difference is statistically
significant(F=19.55, p<001), and in the expected direction.
Closed Ended Inference Measures
The three types of inferences which were predicted, a priori, to be influenced by
the event and associational factors, are brand size, brand legitimacy and event facilitation.
Because each of these dependent variables is associated with three main effects predictions
(impact of event size, duration of association and share of presence) as well as three
interaction prediction (fit will or will not moderate each), each dependent variable was


122
Thus far, a good deal has been learned about fit, duration and event size. Study
la, however, found no measurable impact of share of presence. Since the manipulation of
share in study la represented a somewhat limited range of the share of presence construct,
study lc explores the impact of a much more intrusive (and managerially pertinent) level
of share of presence: title sponsorship, or putting the brand name into the name of the
event.
Study lc
Overview
This study attempts to extend study la by putting the brand name into the name of
the event during exposure. A completely between subjects 2x2 factorial design exposed
subjects to an the same unfamiliar brands as in study la. Once again, unaided and cued
open ended inference measures were taken, followed by closed ended inference scale
measures.
Subjects
Sixty-two undergraduates in an introductory marketing class at a major
southeastern university took part. Each received course extra credit for participation.
Design
The study consisted of a 2(big/small event size) by 2(high/low brand-event fit)
between subjects design. This design is depicted graphically below:


76
Precise Pen with Middle East peace4. Brewer (1993) also reports on an attempt by K-
Swiss shoes to capitalize on the fact that the Pope bought a pair of their sneakers while in
Denver. Finally, there is the case of the McDonalds World Cup Soccer promotion in
which sacred Koran scripture was featured on two million throw-away hamburger bags.
This promotion greatly offended Muslims in Great Britain (Alonzo, 1994).
The above examples seem to provide evidence of sacred events or objects which
are difficult to successfully exploit through commercial association. Similar positive
examples of associations which failed because the product was too profane, are not as
readily available. While Tambrands was among the corporate sponsors of the Womens
National Basketball Team (Rosner, 1995) and Angel Soft bathroom tissue is involved with
a cares for kids program (Weisz, 1995), there are no known cases of either feminine
hygiene products or bathroom tissues having an event named after them. The same is true
for other product classes found to be causes of irritation (Aaker and Bruzzone, 1985),
such as laxatives and breath fresheners.
Ail interesting and complicated possibility is that sacrilege may be perceived for
brand-event associations in which the brand is not strictly profane and the event is not
strictly sacred. Some combinations of associates, for reasons which have not yet been
examined, may lead consumers to consciously elaborate on the "inappropriateness" of the
association. This may sometimes go beyond mere elaboration in ones own mind, leading
to actual verbalization. Discussion among consumers about the crass and commercial
nature of the association can then ensue. The impact of such discussion among consumers
about associations is also currently unexplored.
4 The pen was apparently used in signing some of the documents.


143
more likely to attribute a larger role to the supply of bottled water (Evian) to the success
of the event. That is, a small event is helped more by this form of event support than a
large one.
To test this, subjective and objective knowledge were regressed on the perceived
size of the event. The impact of subjective knowledge on perceived size was significant
(F=5.51. p<03). The impact of objective knowledge was nearly significant (F=3.64,
p<065). In both cases, as knowledge decreases, the perceived size of the event decreases
as well.
Therefore, it appears that the somewhat mysterious moderating role of knowledge
on facilitation inferences may have a possible explanation. In addition, this finding can be
seen as supporting H9, which states that as the perceived size of an event decreases,
perceived facilitation by a sponsoring brand increases. While indirect, this finding is of
some importance to the development of overall sponsorship strategy. This topic is
covered in depth in chapter eleven.
This chapter completes the initial exploration of the first stage of the inference
based model. A good deal has been learned about how event, associational and audience
factors impact inference making about a sponsoring brand. How these inference impact
choice is covered in the following chapter.


CHAPTER SEVEN
ASSOCIATION-ENHANCING COMMUNICATIONS (AEC)
As alluded to earlier, sponsorship associations do not exist in a vacuum. Firms can
engage in a variety of communication activities in order to maximize the impact of the
association for which they have paid. These communications are referred to here as
association-enhancing communications, or AEC. The expenditures above and beyond
the fee paid to be the sponsor are genetically referred to by practitioners as
leverage(Aiidrews 1996). Leverage can attempt to enhance awareness of the
sponsorship among consumers through advertising. It can also, however, take the form of
tie-in promotions, sales promotions, case lot discounts, and other perks to the channel
members in conjunction with the event.
Leverage has received some attention in the sponsorship literature. Meenaghan
(1991a) defines leveraging as the additional effort, largely promotional, which must be
invested by the sponsor in order to properly exploit the opportunity provided as a result of
securing particular sponsorship rights (p. 43). He goes on to say that it is generally
agreed that an amount at least equal to the sponsorship fee should be allocated to
leveraging the sponsorship. Otker (1988 ) presents a somewhat higher estimate, stating
that for each dollar spent on buying, another three dollars is spent on exploitation (p.
77).
80


CHAPTER SIX
SACRILEGE.
Tottenham Hotspur, Liverpool, Manchester United: these names have magical
associations for those -who -worship the great football god Singing, swaying
masses on Wembley terraces. Muddy mayhem in the twilight gloom. The glory
and the gory. Red vs. Blue. Catholic vs. Protestant. North vs. South. And now...
AVCO vs. Crown Paints. JVC vs. Sharp. Holsten vs. Home Ales. Wright, (1988,
p. 104)
No discussion of sponsorship would be complete without acknowledging the fact
that some consumers may think that it is going too far. In an age of increasingly bold
associations, the problem should only continue to grow. One acknowledgment that
people may be getting fed up is evidenced by Taco Bells April Fools Day prank in
which the company placed an ad in the April 1, 1996 New York Times proclaiming:
TACO BELL BUYS THE LIBERTY BELL
In an effort to help the national debt, Taco Bell is pleased to announce that we
have agreed to purchase the Liberty Bell, one of our countrys most historic
treasures. It will now be called the Taco Liberty Bell and will still be accessible
to the American public for viewing ....
Clearly, at least among the advertising creatives at Taco Bell, there may be a problem.
The idea that maybe there are some things which are sacred, and therefore
should not be involved in commercial association is an intriguing one. The idea that some
products may be so profane that any association with a non commercial entity would
74


88
in their goodwill position, are expected to be differentially impacted by each type of
sponsorship-based inference.
FIGURE 3
THE INFERENCE-BASED MODEL OF BUILDING BRAND EQUITY THROUGH
SPONSORSHIP
The remainder of this chapter details a series of three studies which investigate the
impact of event size, brand/event fit, duration of association, and share of presence
(defined below) on consumer inference-making about a sponsoring brand. The following


125
Since the effect of event size on perceived quality appeared to be bigger under title
sponsorship, the interaction of title sponsorship with event size was tested to see if this
difference was significant. A model which regressed size, title sponsorship and their
interaction on quality was run. In this combined model, event size still has a significant
impact (F=5.54, p<02), but neither the main effect of title sponsorship (F=1.86, p>. 15)
nor its interaction with event size (F<1) was significant. Therefore no statistical support is
found for an amplification of the impact of event size on perceived quality under title
sponsorship.
Open Ended Measures
As in la and lb, two raters independently coded all of the thoughts. Initial
agreement between raters was high. Out of a total of 561 thoughts, the coders matched
on 519 (92.5%). Discrepancies were again reconciled through discussion and mutual
agreement. Of the 62 subjects, 24 made a total of 43 inferences about the sponsoring
brand. The other 38 subjects listed no thoughts which could be considered inferences
about the brand.
Once again the small number of overall inferences made it unlikely to be able to
detect effects of event size and duration of association on specific inferences. As
expected, none was detected. As before a series of loglinear models were used to look for
effects of the predictors on total inferences. No impact of either fit or size was detected
on total inference making.


119
Closed Ended Inference Measures
Impact on perceived size of the brand
As in study la, the size of the event had a measurable impact on one of the
perceived size variables Event size significantly predicted absolute perceived size
(F=4.74, p< 05). Once again, brands associated with large events were perceived to be
bigger than brands associated with small events. As in study la and in contradiction to
original prediction, brand/event fit appeared to have a weak, moderating role (F=2.72,
p< 11, (Di=.02). The power of this experiment to detect this small effect was 0.18. In
order to obtain a power level of .80 to detect this effect, sample size would need to
increase from 80 to 224. Unlike in study la, duration of association had no impact on
perceived brand size.
Impact on brand legitimacy and event facilitation
Unlike in study la, event size had no measurable impact on perceived quality or
legitimacy. Duration of association also had no impact on brand legitimacy inferences.
The impact of duration of association on event facilitation did not replicate. Event size
also had no measurable impact on event facilitation inferences.
Open Ended Measures
Once again two raters independently coded all of the thoughts. Initial agreement
between raters was again quite high. Out of a total of 728 thoughts, the coders matched
on 701 (96.3%). The remaining discrepancies were reconciled through discussion and


189
EXPERIMENT 1C ALL LOW
The Rombouts
Student Fun-Run
Leuven, Belgium
Sunday October 19th 1997 at 9:30 am
Proud to be a Sponsor for 5 Years


102
in event sponsorship. Twenty subjects rated perceived fit between a running event and
each category. Means and standard deviations are provided in Table 3. The two brands
chosen are shown in italics. They are coffee (mean perceived fit=2.75), and sports drink
(mean perceived fit=6.9). This difference in perceived fit is statistically significant
(F=157.5, p<001). Coffee and sports drink are both in the beverage category, and both
categories have brands which sponsor running events.
Having selected the categories, it was necessary to select the actual brands.
Preference was given to actual foreign brands, provided they were very unfamiliar to the
subject population. To find such brands of coffee and sports drink, three European sports
drink brands and twelve European coffee brands were pretested for familiarity. These
brands were embedded with other European brands being pretested for an other study. In
all, twenty subjects rated 40 European brands for familiarity. Results of this pretest are
shown in Table 4. On a seven point scale, anchored by never heard of it and very
familiar, Rombouts Coffee had a mean rating of 1.05, and Extran Sports Drink had a
mean rating of 1.35. These differences are not statistically significant, even at a
conservative p=.10.


94
the perceived similarity or congruence between two associates. This is what was defined
in chapter five as generic fit. While other, more complex conceptualizations of the fit
construct were provided in that chapter, the simplicity of perceived fit, and its frequent
usage in the extant commercial association literature make it an appropriate construct for
investigation in the emerging domain of sponsorship.
Duration of Association
Duration of association is also expected to impact inference making about a
sponsoring brand. Duration of association is defined as the length of time that the brand
has been associated with the event. A new or first-time sponsor represents extremely
low duration of association. For most events, duration of association is measured in years.
Consumers are more likely to be aware of long term associations. Since awareness
is a necessary condition for inference making, long term associations should lead to more
consumer inference making than short term associations. In addition, long term
associations may imply greater commitment on the part of the brand to the event. It is
expected, then, that long term associations should lead to greater consumer inference
making about the commitment of the company to the event than do short term
associations. It is also expected that long term associations will lead to more event
facilitation inferences than short term associations.
Share of Presence
The share of presence a brand has at an event relative to all other sponsors is also
expected to impact inference making. Share of presence is a measure of how much the


CHAPTER EIGHT THE IMPACT OF EVENT AND
ASSOCIATIONAL FACTORS ON INFERENCES
ABOUT THE BRAND 86
Conceptual Overview 89
Hypotheses About Specific Inferences 95
Study la 99
Study lb 115
Study lc 122
Additional Cross Study 127
Overall Discussion and Limitations of Studies la-lc 128
CHAPTER NINE THE IMPACT OF AUDIENCE FACTORS
AND FIT ON BRAND INFERENCES 131
Study 2a 133
Study 139
Overall Discussion 141
CHAPTER TEN THE IMPACT OF SPONSORSHIP-BASED
INFERENCES ON BRAND CHOICE 144
Study 3a 145
Study 3b 149
Study 3c 152
Overall Discussion of Studies 3a, 3b, and 3c 157
CHAPTER ELEVEN CONCLUSION, IMPLICATIONS AND
FUTURE RESEARCH 159
Implications 162
Future Research 165
Conclusion 169
APPENDDC A OPEN ENDED MEASURES AND MANIPILATION
CHECKS FOR EXPERIMENT 1A, IB, AND 1C 170
APPENDDC B OPEN ENDED INFERENCE MEASURES AND
MANIPULATION CHECKS FOR EXPERIMENT 2A AND 2B 173
APPENDIX C INVOLVEMENT AND KNOWLEDGE MEASURES
USED IN EXPERIMENT 2A 175
APPENBDC D OPEN ENDED THOUGHT LISTING BOXES 179
APPENDIX E SAMPLE CHOICE PROFILE USED IN
EXPERIMENT 3 A 180


53
by communications, the details of this finding will be discussed in the association
enhancing communications chapter. If, however, one supposes that frequency of
instantiation of the brand/event pairing might spontaneously lead people to consider the
similarity, then it may be possible that over several encounters with the pairing, not only
is fit facilitated, but the more abstract, relational dimensions of fit might become more
prominent. These ideas are formally stated:
F42: Frequent attendance at an event leads to greater increase in the salience
of abstract relational common attributes than in the salience of more
concrete common attributes.
P43: Recent attendance at an event leads to greater increase in the salience of
abstract relational common attributes than in the salience of more concrete
common attributes.
Within the broader research stream involving how people fit things into categories,
Cohen and Basu (1987) suggest that there is a real possibility... that the consumer
environment favors category definition in terms of specific exemplars rather than category
defining features.. . (p. 470). This is in reference to a perceived uncertainty as to when
categories will be defined by exemplars as opposed to being defined in a probabilistic
fashion (see also Smith and Medin, 1981). The critical issue here is that because brands
are not incredibly important to most people, the mechanism associated with categorizing
brands is not likely to involve capacity-depleting probabilistic rules. This seems to suggest
that when it comes to the fit between a brand and an event, the large bulk of the fit will be
determined by a perceived match between exemplary attributes, with less of a role for
attribute matches which do not immediately or spontaneously come to mind.


92
The marketing literature provides other evidence of the benefits of perceived size.
Kirmani and Wright (1989) propose three possible mechanisms through which perceived
expenditure on advertising might lead to inferred quality. One is the notion that
confidence in quality leads companies to spend more on promoting a product. Another is a
perceived relationship between promotional spending and quality in some markets. Finally,
they cite the idea that perceived advertising expense is a signal of a firms financial
strength, probable social acceptance or some other factor that defines quality in some
markets (p. 344). Kirmani (1990) proposes that perceived expenditure will be used as
cues to quality when they are more diagnostic than other cues.
Associational Factors
Associational factors, or attributes of the association between the event and the
brand, are also expected to impact inference making about the brand. Associational
factors include fit between the event and the brand, the duration of association between
the event and the brand, and the share of presence a particular brand has relative to all the
brands at the event. Brand/event fit was discussed at length in chapter five. Duration of
association refers to the event-brand associational history (number of years the brand
has sponsored the event). Share of presence is a measure of consumer perception of how
large a presence a brand has at an event relative to the other sponsors.
The fit between a brand and an event is expected to be a strong determinant of the
number and kinds of inferences a consumer makes about a sponsoring brand. As
previously discussed, fit can be directly measured as a unidimensional construct which is
perceived by consumers. Operationally, it has been defined in several ways by researchers


Abstract of Dissertation Presented to the Graduate School
Of the University of Florida in Partial Fulfillment of the
Requirement for the Degree of Doctor of Philosophy
AN INFERENCE-BASED MODEL OF BUILDING BRAND
EQUITY THROUGH SPONSORSHIP
By
John W. Pracejus
August 1998
Chairperson: Richard J. Lutz
Major Department: Marketing
Sponsorship, or the intentional association of a brand or company with a sports
teams or sporting event, entertainment tour or attraction, festival, fair or annual event,
charitable cause, the arts, or cultural institutions, is emerging as an important component
of the marketing mix. While there is general agreement that sponsorship is effective at
exposing consumers to a brand name, little is known about how sponsorship can build
brand equity.
A two stage model of building brand equity through sponsorship is proposed. It is
proposed that in the first stage, event, associational and audience factors lead consumers
to draw different inferences about the sponsoring brand. These inferences include brand
size (i.e. this must be a big brand to be able to sponsor this event), brand legitimacy
(i.e. this brand must be OK in order to be allowed to sponsor this event), and event
facilitation (i.e. this brand is making this event possible, or at least making it better). In
the second stage, these inferences are proposed to differentially impact choice
probabilities for different categories and brands.
vii


148
The key attributes of interest were company size and event facilitation.
Perceptions of both were shown in previous studies to be impacted by sponsorship
associations. Company size was manipulated at two levels. At the low level, the profile
read small company. At the high level, it read large company. Event facilitation was
also manipulated at two levels. The attribute was given in the profile as company
involvement with favorite local events. At the low level of event facilitation, the profile
read not involved. At the high level of event facilitation the profile read helps make
them possible.
Finally, customer service hours were manipulated at two levels. At the low level,
the profile read 8AM to 6PM, Mon.-Sat. (60 hours per week). At the high level, the
profile read 24 hours/day, 7 days/week (168 hours per week).
Results
A logistic regression was used to test the impact of the five independent variables
on auto insurance choice. The logit model regressed price, deductible, company size,
event facilitation, and service hours on choice. Parameter estimates for each variable are
provided in Table 8 below.


6
provide corporate hospitality was rated below the midpoint of the five point importance
scale.
Withcer et al. (1991) found that companies used sponsorship of cultural and art
institutions to reach opinion leaders and to achieve community relations goals, whereas
sports sponsorship was used primarily to reach the general public directly. Scott and
Suchard (1992) factor analyzed the results of a 23 question survey of 108 Australian
businesses. The resultant four factors were regressed on the amount spent on sponsorship
by each company. Two of the four factors, performance and client relationships, were
found to be significantly correlated to expenditure on sponsorship. The significant
correlation of client relationship and expenditure is said to be somewhat counter to
previous studies (i.e. Marshall and Cook, 1992) which found hospitality to be only a
marginally important goal.
Another issue which has received some attention in the sponsorship literature is
how do companies go about choosing an event, once they have decided to engage in
sponsorship? Gardner and Shuman (1987) surveyed Fortune 500 companies about their
choice of sponsorship activities. Eighty-four percent reported using in-house expertise for
screening, while 24% percent utilized outside counsel to evaluate and choose particular
events.
In a survey of companies sponsoring football (soccer) in the UK, Thwaites (1995)
found that sponsors of Premier League Clubs, which have high profile and cost, were
more proactive in securing sponsorship opportunities, compared more clubs before the
decision and negotiated longer term contracts, than did sponsors of Division One Leagues,
which have moderate profile and cost (all significant at p<05). This seems to suggest that


APPENDIX A
OPEN ENDED INFERENCE MEASURES AND MANIPULATION CHECKS FOR
EXPERIMENT 1A,1B AND 1C
Now we would like to get your impressions about one of the sponsoring brands:
Extran Sports Drink (or Rombouts coffee)
Do you think Extran Sports Drink (Rombouts Coffee) is:
One of the smallest sports One of the largest sports
drinks (coffees) in Belgium | | | | | | | | drinks (coffees) in Belgium
How large do you think Extrans (Rombouts) market share is, compared to other sports
drinks in Belgium?
Very small share | | | | | 1 | | very large share
How legitimate of a sports drink (coffee) do you perceive Extran Sports Drink
(Rombouts Coffee) to be?
A fly-by-night brand | | | | | | | | A legitimate brand
How real of a brand do you perceive Extran Sports Drink (Rombouts Coffee) to be?
Not very real | | | | | | | | Very real
170


201
Rajaretnam, J. (1994) The long term effects of sponsorship of corporate and product
image: findings of a unique experiment. Marketing and Research Today. 22, 1,
62-74
Rangaswamy, Arvind, Raymond R. Burke, and Terrence A. Oliva (1993) Brand equity
and the extendibility of brand names. International journal of Research in
Marketing. 10, 61-75
Rao, Akshay A., and Robert W. Ruekert (1994) Brand alliances as signals of product
quality. Sloan Management Review. Fall, 87-97
Reddy, Srinivas K., Susan L. Holak, and Subodh Bhat (1994) To extend or not to
extend: success determinants of line extensions Journal of Marketing Research.
31 (May), 243-262
Rosch, Eleanor (1976) Basic objects in natural categories. Cognitive Psychology, 8, 3,
382-439
Roslow, Sydney, J.A.F Nicholls, and Henry A. Laskey (1992) Hallmark events and
measurement of reach and audience characteristics. Journal of Advertising
Research. 32, 4, 53-59
Rosner, Hillary (1995) Dream team's sisters. Brandweek. 36, 42, 24-26
Ryssel, Christian, and Erich Stamminger (1988) Sponsoring world-class tennis players
European Research. May. 110-116
Sanbonmatsu, David M. and Frank R. Kardes (1988) The effects of physiological
arousal on information processing and persuasion Journal of Consumer Research.
15, 3, 379-385
Sandler, Dennis, and David Shani (1989) Olympic sponsorship vs. ambush marketing:
who gets the gold? Journal of Advertising Research. 29, 4, 9-14
Sawyer, Alan G. and Danial J. Howard (1991) Effects of omitting conclusions in
advertisments to involved and uninvolved audiances. Journal of Marketing
Research. 28, 4, 467-474
Scott, D. R., and Hazel T. Suchard (1992) Motivations for australian expenditure on
sponsorship-an analysis. International Journal of Advertising. 11, 325-332
Sedikides, Constantine, Nils Olsen, and Harry T. Reis (1993) Relationships as natural
categories. Journal of Personality & Social Psychology. 64, 1,71-82


177
Playing tennis lets people see me as I ideally would like them to see me
strongly agree | | | | | | | |strongly disagree
Tennis helps me to attain the type of life I strive for
strongly agree | | | | | | | |strongly disagree
I can make many connections or associations between tennis and experiences in my life
strongly agree | | | | | | | |strongly disagree
Tennis is of high personal importance to me
strongly agree | | | | | | | |strongly disagree
Tennis helps me express who I am
strongly agree | | | | | | | |strongly disagree
How many tim es did you play tennis last week?
How many times did you play tennis last month?
How many times did you play since last May?
How many times did you watch tennis last week?
How many times did you watch tennis last month?
How many times did you watch tennis since last May?


98
tournament should be more concerned about sponsorship by a poor quality brand of tennis
balls than about sponsorship by a poor quality bank. Therefore:
H8: Brand/event fit moderates the relationships in H5-B7, such that high fit
amplifies these effects
Event Facilitation
Event facilitation inferences occur when a consumer comes to the conclusion that a
company or brand is helping to make possible, or improve the quality of an event. Given
that small events are more likely to be completely dependent on the goodwill of the
sponsor:
H9: Sponsoring smaller events leads to more event facilitation inferences on the
part of consumers
Since long term associations should imply more commitment on the part of the sponsor:
H10: Longer duration of association between brand and event leads to more event
facilitation inferences on the part of consumers
Since major sponsors should be viewed as doing more to insure the success of the event:
Hll: Engaging in a sponsorship with a large, as opposed to small, share of
presence leads to more event facilitation inferences on the part of consumers
Since sponsors may contribute technical knowledge and expertise to an event:
H12: Brand/event fit moderates the relationships in H9-H11, such that high fit
amplifies these effects


120
mutual agreement. Of the 80 subjects, 20 made a total of 26 inferences about the
sponsoring brand. The other 60 subjects listed no thoughts which could be considered
inferences about the brand.
Once again the small number of overall inferences made it unlikely to be able to
detect effects of event size and duration of association on specific inferences. As
expected, none was detected. As before, a series of loglinear models were used to look
for effects of the predictors on total inferences.
Just as in study la, fit seemed to have an impact on overall inference making (Chi
Sq=3.64, p<06). The low fit group made an average of .45 inferences, whereas subjects
in the high fit group made an average of .20 inferences, meaning the effect of fit found in
la had reversed! Therefore, it should be immediately obvious that the increased inference
making in la was not necessarily the result of fit, but some other aspect of Extran Sports
Drink.
Discussion
Study lb provided a good bit of additional insight into the consumer inference
making process, and how it can be influenced by event and associational factors. By
replicating the significant impact of event size on perceived brand size, greater confidence
in the robustness of this effect is obtained. As mentioned previously, such an effect should
be of great interest for categories in which perceived brand size strongly impacts choice.
Secondly, and of considerable importance, is the complete reversal of the impact of
fit on the total amount of inference making. This relationship is depicted below in Figure
6.


s
viewers of a sponsored public television program (MacNeil/Lehrer News Hour) were
more aware that AT&T sponsored the News Hour than were nonviewers.
Crimmins and Horn (1996) point out the need to take into account the number of
people who mistakenly believe that a competitor is a sponsor, in addition to actual
awareness that your brand is the sponsor of an event. They point out that while 35% of a
consumer mail panel could correctly identify Coke as the official soft drink of the NFL,
that 34% thought it was Pepsi, and note that Coke had paid $250 million to become the
official soft drink.
Awareness levels of sponsorship have also been studied in an experimental setting.
Hoek, Gendall and Stockdale (1993) exposed New Zealand secondary school students to
tapes containing a cricket match with a cigarette brand as the title sponsor (the 1990
Rothmans Cricket Series) and control subjects to a match without cigarette sponsorship.
The only significant effect found was that treatment subjects subsequently had a higher
awareness for Rothmans than did a control group. No significant impact on attitude
towards smoking was found.
Some papers have focused on awareness levels that a company is a sponsor.
Wright (1988), for example reports awareness levels for sponsors of different football
(soccer) teams in Great Britain. Time series data are provided which demonstrates the
various awareness levels over time.
Rajaretnam (1994) reports the findings of a unique experiment in which a large
Indian tire maker stopped all product message advertising and spent almost its entire
marketing budget on sponsorship from 1984-1987. While the author does not consider in
detail other potential factors impacting awareness and image variables (e.g., increased


33
the brand. Image transfer is the mechanism requiring the least cognitive elaboration in
which the association can be formally stated as a message. The message of image transfer
is that, along some dimension, this brand is like this event.
The idea that the image, or personality of an event can impact upon the image or
personality of a brand has been alluded to several times in the context of sponsorship.
Parker (1991), for example, describes the relationship between sponsorship and brand
image. Marshall and Cook (1992) found that 14% of respondents listed image of sport
as the most important determinant in selecting an event to sponsor. Wolton (1988)
contends that art sponsorship is used to alter or modify the image of a company, lending a
human face to cold, technologically-oriented companies, and providing a positive
association with companies such as insurers which are usually associated with unpleasant
situations.
McDonald (1991) describes what he refers to as product relevance (p. 36) which
can be either direct or indirect. Indirect relevance is proposed to imply association at an
abstract, more image-based level. A strong and traditional bank sponsoring the traditional
sport of football (soccer) is given as an example of indirect relevance, where the link is
tradition. Relating back to the discussion above, in this example the message would be
that this bank is like the soccer league in that they are both traditional.
Perhaps the best anecdotal evidence of how the image transfer mechanism might
be used for a specific marketing objective, like repositioning a brand, is given by
Meenaghan (1991a). He states that sponsorships can achieve specific objectives through
what he calls the image by association effect(p. 42). He gives as an example of how a
very American company, Gillette, was made to seem more British through its


200
Nicholls, J.A.F., Henry A. Laskey, and Sydney Roslow (1992) A comparison of
audiences at selected hallmark events in the united states. International Journal of
Advertising. 11, 215-225
OConnor, Robert (1996) Thumbs up for cobranding Credit Card Management. 9, 6,
150-154
Ortony, Andrew (1979) Beyond literal similarity. Psychological Review. 86, 3, 161-180
Otker, Ton (1988) Exploitation: the key to sponsorship success European Research.
May, p.77-86
Park, C. Whan, Bernard J. Jaworski, and Deborah J. Maclnnis (1986) Strategic brand
concept-image management. Journal Of Marketing. 50, 4, 135-146
, Sandra Milberg, and Robert Lawson (1991) Evaluation of brand extensions:
the role of product feature similarity and brand concept consistency Journal of
Consumer Research. 18, 2, 185-193
, Bernard J. Jaworski, and Deborah J. Macinnis (1986) Strategic brand concept-
image management. Journal of Marketing. 50, 4, 135-145
, Sung Youl Jun, and Allen D. Shocker (1996) Composite branding alliances: an
investigation of extension and feedback effects Journal of Marketing Research.
33, 4, 453-466
Parker, Ken (1991) Sponsorship: the research contribution. European Journal of
Marketing. 25, 11, 22-30
Petty, Richard E., John T. Cacioppo, and Rachel Goldman (1981) Personal involvement
as a determinant of argument-based persuasion. Journal of Personality & Social
Psychology, 41, 5, 847-855
Pham, Michel Tuan (1993) Cue representation and selection effects of arousal on
persuasion Journal of Consumer Research. 22, 4, 373-387
Pracejus, John W. (1995) Is more exposure always better? Effects of incidental exposure
to a brand name on subsequent processing of advertising Advances in Consumer
Research. Vol. 22, Frank Kardes and Mita Sujan, eds. Provo, UT: Association for
Consumer Research, 319-322
Proctor, James (1996) Why L.A. Gear is relying on sponsorship as the first step in its
repositioning. Paper presented at the 1995 IEG Event Marketing Conference,
Chicago Illinois.


184
EXPERIMENT 1A ALL HIGH
The Euro-Capital
International Marathon
Brussels, Belgium
Sunday October 19th 1997 at 9:30 am
Proud to be a Sponsor for 5 Years


95
brand is perceived to be sponsoring the event. Things like the amount and prominence of
brand signage at the event and on promotional materials for the event are expected to
moderately impact perceived share of presence. Naming the event after the brand is
expected to have a strong, positive impact on perceived share of presence.
Sole sponsorship of an event would represent 100% share. As the number of co
sponsors at the same level of sponsorship goes up, an individual brands share of presence
at the event goes down. Like duration of association, share of presence is expected to
imply commitment of the brand to the event. Therefore, consumers are expected to make
more event facilitation inferences about brands with a high share of presence than about
brands with low share of presence.
Hypotheses About Specific Inferences
Review of the extant literature on sponsorship and other commercial associations
led to the conceptualization of three specific types of inference that might be generated
from an exposure to a sponsorship association and might impact brand choice. These
inferences are: brand size, brand legitimacy, and event facilitation. These three
types of inference, and their a priori relationship to event and associational factors are
described below. Other consumer inferences which emerged from the data are described
in the results section.
Brand Size
Inferences about brand size occur when a consumer changes (or forms) his/her
perception of the size of a firm or brand based on a sponsorship association. As


101
TABLE 2
AVERAGE PRECE1VED SIZE OF RUNNING EVENTS* (N=20)
Runnina Event
Mean
St Dev
AUCKLAND MARATHON
2.90
1.59
BERLIN MARATHON
5.15
1.46
COEUR D' ALEE MARATHON
3.80
1.15
COLOGNE 10K
3.80
2.14
COMRADES MARATHON
3.45
1.90
DUBUN MARATHON
4.75
1.52
DUSSELDORF MINI-MARATHON
2.75
1.16
EURO-CAPITOL INTNL MARATHON
5.35
1.53
FIRENZE MARATHON
3.85
1.39
FLETCHER CHALLENGE MARATHON
3.95
1.57
HELSINKI CITY MARATHON
4.30
1.59
JUNGFRAU-MARATHON
2.90
1.07
LEUVEN STUDENT FUN-RUN
2.50
1.19
LIVERPOOL HALF MARATHON
3.20
1.06
LONDON MARATHON
6.20
0.95
LUXEMBORG 10K
4.00
1.89
MARATHON DES SABLES
2.05
1.05
NEWCASTLE 10K
3.45
1.54
PARIS MARATHON
6.10
1.25
REYKJAVIK MARATHON
3.55
1.47
ROMA CITY MARATHON
5.65
0.93
ROTTERDAM MARATHON
3.60
1.27
SHANNON CITY FUN-RUN
2.35
1.23
STOCKHOLM MARATHON
4.95
1.05
TILLBERG CITY 10K
2.95
1.36
VIGARANO MARATONA
3.30
1.53
* 1=VERY SMALL; 7=VERY LARGE
Brand/event fit
Fit was manipulated by selecting two brands which were equally unfamiliar to the
student population and varied substantially in degree of perceived fit with a running event.
Ideally, these would be from categories which actually engage in running event
sponsorship. Pretesting was performed on thirty-nine brand categories currently involved


138
p< 001) domain knowledge (as measured by the interaction term in a model containing fit,
the knowledge variable, and the interaction of fit and the knowledge variable).
To see the nature of the fit by subjective knowledge interaction, the data was split
at the mean of subjective knowledge. Separate follow up analysis then tested the effect of
fit on event: facilitation for each of the two subjective knowledge groups. This analysis
revealed that for the high subjective knowledge subjects, fit had no impact on event
facilitation inferences (F<1). For low subjective knowledge subjects, however, Fit had a
strong, positive impact on event facilitation inferences (F=l 1.99, p< 01)
In the same way, the data was split at the mean of objective knowledge. Here
again, for high knowledge subjects, fit had no impact on event facilitation inferences
(F<1). For the low knowledge subjects, fit had a strong and significant effect on
perceived event facilitation (F=25.66, p<0001).
Open Ended Bata
Open ended inference measures were coded as before. Overall initial agreement
between the two raters was quite high. Out of 316 total thoughts, they agreed on 293
(92.7%). Discrepancies were resolved as before. Inferences about the sponsor were made
by 12 out of 38 subjects. A total of 12 brand inferences were made. No impact of fit or
any audience factors was detected on any of the four specific inferences. Neither was any
impact of fit or the audience factors detected on total inference making. Given the small
number of inferences, however, the lack of significant results may be due to low power.
It appears that the main effect of fit on event facilitation inferences was being
driven by subjects with low knowledge levels about tennis. Before discussing the


APPENDIX C
INVOLVEMENT AND KNOWLEDGE MEASURES USED IN EXPERIMENT 2A
Please indicate your level of agreement with the following statements by placing an X in the appropriate
box.
The message in the poster I just saw was important to me
strongly agree | | | | | | | | strongly disagree
The poster didnt have anything to do with me or my needs
strongly agree | | | | | | | | strongly disagree
Please take a moment (approximately 5-10 seconds) to think about tennis and what it means to you. Then
complete these scales.
important 1
I
1
1
1
!
1
J unimportant
of no concern (_
1
1
1
1
1
1
_J of great concern
irrelevant |_
|
J
1
1
1
1
_| relevant
means a lot to me |_
I
I
I
1
1
1
I means nothing to me
useless |_
|
1
I
I
1
1
_J useful
valuable |_
|
1
I
I
1
I
_| worthless
trivial 1_
|
1
1
1
1
1
J fundamental
beneficial |_
|
1
I
I
1
I
_J not beneficial
matters to me [_
1
1
1
I
1
I
_J doesnt matter
uninterested |_
|
1
I
1
1
1
_] interested
significant 1
1
1
I
!
1
1
J insignificant
vital |_
1
1
1
1
1
1
I superfluous
boring |
1
I
1
1
1
1
1 interesting
unexciting |
|
I
I
1
1
1
I exciting
appealing |
I
1
1
1
1
1
1 unappealing
mundane (_
I
1
I
!
1
1
1 fascinating
essential |
_L_
I
I
1
_l_
1
J nonessential
175


132
option two, measuring involvement with an extant activity, was chosen as the method of
choice.
Well known, popular events were chosen so as to ensure that some subjects would
exhibit measurable levels of involvement with the events. Because real events are used, it
was thought that associating fictitious or unfamiliar brands with them would lead to
unacceptably low levels of mundane realism. Rather, real brands were chosen for
association with the event. In the key study (2a), the brands chosen are actual sponsors of
the event.
Because real brands are used, it is unlikely that a single experimental exposure will
influence perceived brand size, legitimacy or quality. Event facilitation and commercial
motive inferences, however, might still be influenced by associational variables. It is these
inferences, therefore, that are the focus of this pair of studies.
Extant commercial association research, such as brand extension and celebrity
endorsement has not addressed the issue of involvement with the provider. This makes
the current study of how involvement with an event impacts inferences about a sponsoring
brand largely exploratory. For this reason, no formal hypotheses are provided.
This set of studies looked at the impact of audience involvement on inference
making. Study 2a is the main study. Study 2b is used to check for a possible confound of
fit with brand. The manipulation of fit in study 2a is reversed in study 2b (as was done in
la and lb). This allows any impact of fit in the key study (2a) to be checked in 2b, to
make sure the effect is the result of fit, and not some other aspect of the two sponsoring
brands. The flipped fit is depicted in figure 8 below.


124
large event size mean was 5.2, and the small event size mean was 3.7 on the size check
measure. This difference is statistically significant as well (F=13.43, p<001).
Closed Ended Measures
Once again, sponsoring a large event impacted the perceived absolute size of the
brand (F=3.05, p<086). As before, associating a brand with a large event led subjects to
infer that the brand is large. This effect may again have been moderated by fit (F=2.8,
p< 10, 2= 02). The power of the experiment to detect this effect was 0.47. To obtain
power of .80, sample size would need to increase from 62 to 172. Event size once again
also had a strong positive impact on perceived quality (F=12.09, p<001) While this effect
was found in study la, the impact appears to be much stronger under title sponsorship
(this possibility is tested below). No other significant effects were detected.
To test whether title sponsorship impacted inference-making, the two data sets
from la and lc were combined. This allowed a comparison of the same brands associated
with the same events. The only difference was whether or not the target sponsor was a
title sponsor.
This analysis revealed that title sponsorship can aid event facilitation inferences.
The difference in response to the event enabling measure was significantly higher in study
lc than in study la (F=5.54, p<03). This effect was not moderated by event size (F<1).
Therefore, it appears that being a title sponsor leads to more event facilitation inferences
regardless of whether the event sponsored is large or small. No other significant effects of
title sponsorship on inference making were found.


158
market structure (domestic competitor), but not on another (foreign competitor). This
finding supports the notion that managers need to consider the competitive environment
when setting sponsorship strategy.
Taken together, these three findings suggest that in order to evaluate the value of a
sponsorship opportunity, a manager must consider the needs of the specific brand and the
category within the current competitive environment. It is not necessary, however, to test
multiple scenarios as has been done in this chapter. An individual decision maker need
only calculate the relative value of each sponsorship-based inference for that brand. Once
these are known, selecting from among various sponsorship opportunities becomes a
matter of considering the likely inferences generated by each available association, and
picking sponsorships that can generate the inferences most desirable for the particular
brand
Basic strategies for developing sponsorship strategies are covered in the following
chapter, which discusses the combined findings of the eight studies, and how they can be
used by managers to make sponsorship decisions. The next chapter also discusses future
research which will eventually lead to more rigorous mathematical screening criteria for
selecting sponsorship associations for brands.


2
sponsorship is increasing faster than spending on advertising or sales promotion
(Andrews 1996).
While the bulk of the growth in sponsorship activities has taken place over the last
twenty years, the idea of creating an exploitable association between a commercial and a
noncommercial entity is not a new one. In fact, an association between the product Bovril
and the Nottingham Forest Football Club dates from 1898 (Marshall and Cook, 1992). In
1927, Charles Lindbergh named his plane the Spirit of St. Louis because his financial
sponsors wanted to promote St. Louis, Missouri, as an aviation city. Thus, the first
transatlantic flight was, in fact, a sponsored event (Hildreth and Nalty, 1969; Lindbergh,
1953). When Jesse Owens won four gold medals in the 1936 Olympics, he did so in
Adidas shoes (Ryssel and Stamminger,1988, p. 111). Finally, in the first known example
of a sporting event being named after a brand, The Whitbread (Ale) Gold Cup, a horse
racing event, first took place in 1956 (Marshall and Cook, 1992).
][n spite of this long history as a phenomenon and rapidly growing levels of
expenditure, the mechanisms through which sponsorship affects brand equity, preference
and choice, are not well understood. The purpose of this dissertation is to begin
systematically exploring the ways in which sponsorship works. It begins with a review of
extant sponsorship research and is followed by a summary of relevant research from
related commercial association domains, like celebrity endorsement and brand and line
extension. Measurable marketing variables upon which sponsorship may have an impact
are then discussed.
Exploratory research and managerial intuition as to how sponsorship works are
then developed into seven testable mechanisms through which sponsorship may function in


AN INFERENCE-BASED MODEL OF BUILDING BRAND EQUITY THROUGH
SPONSORSHIP
By
JOHN W. PRACEJUS
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA


CHAPTER EIGHT
THE IMPACT OF EVENT AND ASSOCIATIONAL FACTORS ON INFERENCES
ABOUT THE BRAND
The preceding chapters have provided an overview of what is currently known
about sponsorship. They then went on to map out the somewhat vast unexplored
territory of how sponsorship may work. While the wide array of unexplored issues
makes the general area of sponsorship an exciting domain in which to work, it also makes
it necessary to limit to a subset the number of issues which can reasonably be addressed
within the scope of a dissertation. This first empirical chapter, therefore, begins with a
brief discussion of which issues were selected for empirical investigation and why they
were chosen.
The core unexplored issues in sponsorship are how and when does associating a
brand with an event build equity for the brand, and, ultimately, increase choice probability
for a segment of consumers. Given that the mechanisms at the lower end of the
necessary elaboration continuum (i.e. awareness, affect transfer) are not expected to
differ functionally from their counterparts in the domain of advertising, it seems most
fruitful to pursue those mechanisms at the upper end of the continuum. Very little is
known about mechanisms like implied size, implied endorsement and reciprocity. These
mechanisms are also likely to be more greatly impacted by sponsorship than by other
forms of promotional activity.
86


APPENDIX D
OPEN ENDED THOUGHT LISTING BOXES
179


93
in other areas of commercial associations. In the domain of brand extensions, Aaker and
Keller (1993) operationally defined fit to be the degree to which the brand extension is a
substitute or a complement for the core brand. Broniarczyk and Alba (1994) seek out
brands with strong, unique associates. If the core brand and the extension possess the
same strong and unique association, they are defined to have relatively high fit. This
suggests that consumer perception of fit is high when the event and brand share unique,
salient attributes.
Many examples of the impact of fit on the success of endorsers have been found.
Friedman and Friedman (1976) found that fit between the type of product being endorsed
and the type of endorser significantly predicted ad evaluation, such that ads for jewelry
were best evaluated when done by a celebrity, whereas ads for cookies were better
received when done by a homemaker. Kamins and Gupta (1994) found that increased
congruence between the spokesperson and the product resulted in a more favorable
product attitude. Speck, Schumann and Thompson (1988) found that when the
endorsers relationship to the product was more obvious, information could be processed
in a shorter time.
Fit has been defined at even higher levels of abstraction, and has been found to
impact marketing success. Park, Milberg and Lawson (1991), for example, define high fit
as occurring when both the core brand and its extension are either prestige or
functional. They found that prestige extensions worked best for prestige core brands
and functional extensions worked best for functional brands.
While definitions of fit vary widely in the extant commercial association literature,
certain commonalties can be found. In most of the examples cited above, fit is defined as


31
be proposing that the positive affective responses to the Hallmark event may transfer to
the sponsoring brand.
Crimmins and Horn (1996) also address the idea of affect transfer. They discuss
the transfer of positive feelings from the event to the brand in terms of Heiders (1946)
elementary human calculus. They posit that, given the relative weakness of many brand
beliefs, the impact of associating a lowly regarded brand with a highly regarded event
should usually lead to an upward estimation of the brand. While this seems plausible, the
authors provide no data to support the assertion. The idea, however, is that the
positiveness of the event can transfer to the sponsor.
Stipp and Schiavone (1996) attempt to demonstrate that attitude toward the
Olympics can actually rub off onto Olympic sponsors. Through multiple regression,
they find that attitude toward Olympic sponsorship in general has a significant impact, net
of ad recall and ad evaluation, upon the sponsors image. While this provides some
support for the affect transfer mechanism, it should be noted that the survey methodology
employed somewhat stacked the deck in favor of their finding.
Respondents were first asked a series of agree-disagree statements like official
Olympic sponsors provide money to: Help athletes train, Make it possible for the
American teams to go to the Olympics,. (p. 24). Setting up the survey with these
questions at the front end certainly had the potential to bias responses toward a strong
association between attitude toward sponsorship and attitude toward a specific sponsor (a
demand effect).
Indication of an affect transfer mechanism is also found in other commercial
association domains. In the context of celebrity endorsers, for example, Misra and Beatty


89
chapter describes two experiments which explore the roie of audience factors on consumer
inference making. Chapter ten explores how sponsorship-based inferences differentially
impact different brands and categories.
Conceptual overview and construct definition for studies la, lb and lc
Studies la,b and c investigate how event and associational factors impact inference
making. Specifically, they explore how event size, brand/event fit, share of presence, and
duration of association impact inferences about brand size, brand legitimacy and event
facilitation. These constructs are defined below.
Event Factors
Event Factors, or attributes of the event being sponsored, are expected to
directly impact consumer inference making about the sponsoring brand. Many aspects of
an event, such as perceived size and prestige, might influence inferences about sponsoring
brands. Pretesting was undertaken to determine the relatedness of these constructs. One
hundred eleven actual sponsored events were evaluated by independent groups of subjects.
Perceived prestige was evaluated along a seven point scale anchored by very prestigious
and not prestigious at all (n=21). Perceived size was evaluated along a seven point scale
anchored by very large and very small (n=18). Mean scores for each construct were
calculated. The means for both constructs, for all 111 events, were then submitted to
simple correlational analysis. The correlation coefficient between the two constructs was
found to be .785.


79
promotional effort which explains the association to the public. The potential benefit of
such additional communication about the association is presented in the following chapter.


105
Duration of association
The duration of association variable was manipulated directly. In the short
duration of association conditions, the target sponsor (Rombouts or Extran) was described
as Proud to be a New Sponsor. In the long duration of association condition, the target
sponsor was described as Proud to be a sponsor for 5 Years.
Share of presence
Share of presence was also directly manipulated. In the low share-of-presence
condition, the target sponsor was simply listed as one of three sponsors of the event. In
the high share-of-presence condition, the target (Extran or Rombouts) was labeled as a
Key Sponsor, whereas the other two sponsors (always Adidas and Generate Bank) were
listed as Associate Sponsors. Also, in the high share of presence condition, the target
sponsors logo was approximately twice as large as the associate sponsors logos. In the
low share-of-presence condition, the logo of the target brand was the same size as the
logos of the other two sponsors.
Experimental Procedure
Subjects entered the lab in groups of four to sixteen. On each desk was a packet
which contained one of sixteen posters which exposed the subject to one of sixteen
combinations of size, fit, duration and share. Each subject saw only one poster. Subjects
were instructed to open the packet to the page containing the poster. They were given 60
seconds to look at the poster as you would if you were to see it while waiting for a bus
or train.


127
Additional Cross-Study Analysis
Because event size and fit were manipulated across all three experiments, effects of
these two variables are tested on the pooled data from all three studies. Impact of
duration, which was manipulated in la and lb, but held constant in lc, is tested on pooled
data from studies la and lb only. Since share was manipulated only in la, no additional
analysis is possible. The impact of title sponsorship is also tested further. Since title
sponsorship was explored only in the context of a running event, it is confounded with
event type. Therefore, title sponsorship was tested on the pooled data from la and lc (the
two running event studies) only.
Total inferences and the four classes of spontaneous inferences were regressed on
fit and size. This was done using the pooled data from all three studies. No additional
significant effects were detected.
Total inferences and the four classes of spontaneous inferences were regressed on
duration of association using the pooled data from la and lb. One significant effect was
found. Here duration of association significantly predicted the number of brand size
inferences made (Chi Sq.=3.9, p<05). Therefore, the support for H2 found with the
closed ended data in study one is replicated with the spontaneous inference data.
Finally, the four classes of inference, as well as total inference making, were
regressed on title sponsorship using pooled data from la and lc. This analysis also
revealed one significant effect. Title sponsorship significantly predicted the number of
commercial association inferences made (Chi Sq.=4.3, p<05). It appears, therefore, that
while naming an event after a brand has positive effects (facilitation inferences) it also may
have a downside (more inferences about the commercial intent of the sponsor).


126
Discussion
This study further supports the impact of sponsoring large events on perceived size
of the brand. It also replicates the significant impact of event size on perceived quality.
While this effect initially looked to be bigger under title sponsorship, the formal statistical
test did not support this conclusion. Nonetheless, these replications do add support to the
previous finding that the size of a sponsored event can influence important inferences
about the sponsor.
Comparing the data in this study with that in la also revealed a strong impact of
title sponsorship on event facilitation inferences. Being a title sponsor led to a much
greater perception that the event was being made possible by the sponsor. Given the
costs associated with title sponsorship, this should be a most interesting finding to those
engaged in tactical sponsorship decision making (i.e. being the title sponsor at one event
vs. being an associate sponsor at many).
Once again, while over one third of the subjects generated spontaneous inferences
about the brand, the inadequate number of inferences prevented the detection of any
effects of the independent variables on specific inferences. One way to further explore the
data with greater statistical power was to combine the open ended data from the three
studies. Since subjects did not have equal probability of inclusion in each of the three
studies, results of this analysis are not based on a true experiment. The potential for
additional insight into the spontaneous inference-making process, however, makes such
an analysis worthy of consideration.


156
TABLE 11
ALTERNATIVE SPECIFIC IMPACT OF MANIPULATED VARIABLES ON AUTO
INSURANCE CHOICE: FOREIGN VS. DOMESTIC
Parameter SE of Asymptotic
Estimate
Parameter
t-Stat
Pr(Z>|t|)
intercept
.27265667
.17080180
1.596
.1104
price
-.17256471E-01
.15446752E-02
-11.172
. 0000
deductible
-.15683230E-02
.21380779E-03
-7.335
. 0000
domestic
facilitation
57844041E-01
.79390851E-01
-.729
.4662
foreign
facilitation
. 17589399
.80882673E-01
2.175
.0297
domestic
brand size
.94754983E-01
.16374539
.579
.5628
foreign
brand size
.96620186
.17070479
5.660
. 0000
service hours
.92624914E-02
.10521120E-02
8.804
. 0000
As was the case in 3a, choice decreased as price and deductible increased. Also as
in 3a, as service hours increased, choice increased. All of these effects are as would be
expected.
Turning to the brand specific parameter estimates, however, the results are quite
different from those found in 3 a. Event facilitation, which previously had no impact on
auto insurance choice, now has a significant impact (p<03) on choice for the foreign
brand. For the domestic brand, no influence of event facilitation on choice is found. It
appears, therefore, that event facilitation can differentially impact different brands within
the same category.
Brand size as well shows differential impact on the foreign and domestic brands.
In this study, brand size has a significant positive impact on choice for the foreign brand,
but no measurable impact on choice for the domestic brand. Once again, evidence is
found that even within the same product category an attribute of the brand has a different


141
(91.3%). Discrepancies were resolved as before. Inferences about the sponsor were made
by 14 out of 41 subjects. A total if 19 brand inferences were made. No impact of fit or
any audience factors was detected on any of the four specific inferences. Neither was any
impact of fit or the audience factors detected on total inference making.
Replicating the impact of brand/event fit on event facilitation inferences lends
credibility to the assumption that this effect is truly being driven by fit, and not anomalous
factors associated with the two brands. Instead of revealing a confound as was the case in
the comparison between la and lb, the possibility that this effect is being driven by brand
and not fit is strongly refuted.
Unfortunately the spectator-only nature of Indy Car racing made replicatory
measures of many audience factors seem unrealistic. Asking subjects, for example, their
level of agreement with Indy Car racing helps me attain the type of life I strive for
seemed unreasonable. In hindsight, it is clear that measures of subjective, and objective
knowledge of Indy Car racing could have been taken. While the objective measures
would have been somewhat difficult to calibrate, the objective measure alone would have
been valuable. This shortcoming, however, should not diminish the interesting finding of
study 2a.
Overall Discussion
The purpose of this chapter was to explore the role of audience factors on
inference making about sponsoring brands, and to examine the inference making process
which occurs when familiar brands are associated with large, well known events. Two
main findings were discovered. First, evidence that brand/event fit enhances event
facilitation inference making was found in 2a and replicated in 2b. The fact that this effect


63
poorly evaluated. While the result obtained by these authors is not entirely intuitive, its
existence suggests some possibilities for the impact of attribute fit in sponsorship.
P65: In some situations, extremely high attribute fit results in lower
evaluation of the association than does moderate fit3
No discussion of a construct like attribute fit would be complete without an
examination of possible methods of operational measurement. Several interesting
possibilities are presented in the literature. Others are developed below.
Johnson (1986) performs a direct test of the contrast model (Tversky 1977) in a
consumer product domain. Specifically, the study finds that similarity judgments are
predicted by a linear combination or contrast of the average number of common and
distinctive features of the products being judged (p. 56). This suggests that attribute fit
might be measured by free elicitation of attributes (abstract and concrete) from
independent groups of subjects for an event and a brand. The attribute fit measure could
then be calculated as a function of the number of common attributes and unique attributes
of each associate. Given the possibility that events may generally elicit more abstract, and
fewer concrete attributes than brands, it may be necessary to normalize elicited attributes
along an abstract-concrete continuum before calculating the attribute fit measure.
Similar techniques have been suggested within the domain of brand equity
evaluation Farquhar and Herr (1992) point out the distinction between strong category to
brand associations (category dominance) and brand to category associations (instance
3 When this is the case remains an empirical question.


34
sponsorship of cricket, a traditionally British sport. He also cites the case of IVECO
trucks, which were viewed as weak European vehicles, compared to more macho,
American trucks. Through a sponsorship of the macho sport of heavyweight boxing,
IVECO was able to shift this image in a highly successful way.
The image transfer mechanism is also found in other domains of commercial
association. In the context of celebrity endorsers McCracken (1989), points out that the
cultural meaning with which celebrities are endowed is the root of their success in product
endorsement. Clearly from his arguments, the image of the celebrity (and the meaning of
this image within the culture) are key to an effective meaning transfer process. This
process of transferring meaning from the celebrity to the product is more complex than
explanations which consider attributes of a celebrity, such as credibility and attractiveness,
as unidimensional, linear constructs.
In an empirical study of a transfer mechanism, Walker, Langmeyer and Langmeyer
(1992) have subjects rate brands and endorsers along several abstract dimensions. They
compare before association scores with after association scores along these dimensions to
see how products pick up some of the personality of the endorser. They find that
previously unendorsed products pick up more of the personality of the endorser than did
a previously endorsed product This leads to the following testable proposition with regard
to image transfer in sponsorship:
F22: Brands involved in few sponsorship activities acquire more of the image
and abstract associations of an event than brands which sponsor many
events.


100
Subjects
Eighty-nine undergraduates in an introductory marketing class at a major
southeastern university took part. Each received course extra credit for participation.
Independent Variables:
Four independent variables were orthogonally manipulated. They were event size,
brand/event fit, duration of association and share of presence.
Event size
The event class chosen was a running event. Pretesting was undertaken to find
two events, which while largely unfamiliar (to make sure subjects would not be aware of
the actual sponsors), yet varied substantially in perceived size within the population to be
studied. Both actual and fictitious foreign running events were tested for perceived size
by twenty subjects. Results of this pretest are shown in Table 4 below. From this data, it
was determined that a clean manipulation of perceived size could be accomplished by
using two fictitious running events which ostensibly take place in Belgium. These two
events, The Euro-Capitol International Marathon (mean perceived size 5.35) and The
Leuven Student Fun Run (mean perceived size 2.50) differed significantly (F=43.14,
p<001). The use of fictitious, unfamiliar events had the additional advantage of allowing
maximum flexibility in manipulating the sponsorship association (see study lc).


77
An example of this might be the case of the Tostitos Fiesta Bowl, which was
played to determine the national college football champion of the 1995 season. While the
Fiesta Bowl may not have been strictly sacred and com chips may not have been strictly
profane, discussions which were seemingly the result of perceived sacrilege could be
heard among Florida Gator fans around the time of the game. Comments like Thats not
right.. .the taco chip national championship and Good grief, the Tostitos logo on the
sweatshirts is bigger than the words Florida Gators were common.
Using a small, convenience sampling of comments, it also seemed that the people
who were bigger fans were more likely to express disdain for the association. While there
is absolutely no empirical work in this area, it is possible to propose the following:
P69: The more mundane the brand, the more likely the association to result
in perceived sacrilege.
P70: The more sacred the event, the more likely the association to result in
perceived sacrilege.
P71: For associations which run the risk of sacrilege, greater consumer
involvement with the event results in a greater likelihood of perceived
sacrilege.
P72: For associations which run the risk of sacrilege, lower levels of attribute
fit lead to a greater likelihood of perceived sacrilege.
P73: The more blatant the perceived commercial intent of the association, the
greater the likelihood of perceived sacrilege.
The most frequent response to perceived sacrilege may well be for the consumer to
elaborate that those two things dont belong together. This elaboration may provide the
necessary condition for a dissimilarity frame. Given that people attend more to the


18
P8: Events which are highly arousing cause higher evaluation of more
positive affect for a sponsored brand than do equally evaluated events which
sire less arousing.
Given the intrinsically more arousing nature of live events over televised ones,
P9: The differential effect proposed in P8 is greater for live spectators than
for people who watch an event on TV.
Another moderating factor of commercial associations is the symbolic value of the
provider. In general, providers with greater symbolic value are more capable of having an
impact on the acceptor through the association. This is the foundation for many of the
ideas put forth by McCracken (1989) with regard to celebrity endorsers. In the context of
brand extensions, Reddy, Holak, and Bhat (1994) propose that brands with high symbolic
value will be more readily extended than others. Using statistical analysis on time series
sales data, they find evidence that highly symbolic brands extended better than less
symbolic brands. This is true using both expert ratings of symbolism and consumer ratings
of symbolic value. From these findings in the general area of commercial associations, it
can be argued that for sponsorship,
P10: The higher the symbolic value of an event for a consumer, the more
impact the association will have upon the evaluation of the sponsoring brand.
The type of good involved as the acceptor has also been shown to impact the
outcome of an association. Smith and Park (1992) find that in general, brand strategy
(use of extensions vs. not) accounts for only 4% of market share, whereas for goods
with mostly experience attributes it accounts for 9% of the variance (p. 309). This
increase in the impact of brand extension upon share for experience goods over all


140
Independent Variable
In order to find an event for which Evian had high fit and Texaco low fit, a pretest
was conducted in which 13 subjects were asked to rate perceived fit between both brands
and several large events. The Indianapolis 500 was chosen. This event had a mean
perceived fit with Evian of 5.07, and a mean perceived fit with Texaco of 6.38. These
differences are statistically significant (F=7.41, P<02)
Results
Manipulation Check
Using the same manipulation check scales as in previous studies, mean perceived
fit was 5.9 for the high fit brand and 2.62 for the low fit brand. These differences are
statistically significant (F=56.03, p<0001).
Closed End Inference Measures
Mean perceived event facilitation, as measured by the summation of the three,
seven point scales, is 7.52 for the low fit group and 10.45 for the high fit group. This
difference is significant (F=5.28, p<.03), and in the same direction as in 2a. This supports
the conclusion that fit, and not extraneous brand factors, had amplified event facilitation
inferences. No impact of fit on commercial motive inferences were found (F<1).
Open Ended Inference Measures
Open ended inference measures were coded as before. Overall initial agreement
between raters was reasonably high. Out of 276 total thoughts, they agreed on 232


Copyright 1998
by
John W. Pracejus


CHAPTER TEN
THE IMPACT OF SPONSORSHIP-BASED INFERENCES ON BRAND CHOICE
The preceding chapters have examined how and when sponsorship associations
lead to inference making about the brand. While interesting from a theoretical
perspective, this knowledge is managerially relevant only if these inferences impact brand
choice. The purpose of this chapter is, therefore, two-fold. The first goal is to
demonstrate that sponsorship-based inferences can impact brand choice. The second goal
is to demonstrate that this impact can vary from category to category and from brand to
brand.
All four types of sponsorship-based inferences could easily be investigated with the
method used in this chapter. In order to keep the number of attributes in the choice task
manageable for the subjects, however, the number of sponsorship-based inferences was
limited to two. Because the previous five studies have provided the strongest evidence of
sponsorships ability to impact brand size and event facilitation inferences, the impact of
these inference on choice is investigated in this series of studies.
In order to demonstrate differential impact of the inferences, it was necessary to
select two product classes which varied along some dimension. One such dimension
which might lead to differential impact of each inference is the ambiguity of product
144


16
extension.. .That is, the perceived quality of the core brand influenced brand extension
evaluation only when there was some basis of fit between the core brand and proposed
extension products (p. 37).
Sheinin and Schmitt (1994) find complex interactions between brand affect,
category breadth, and attribute congruity between the original and the extended brand.
Sunde and Brodie (1993) find that consumer acceptance of a brand extension will be
higher if the original brand is of high quality and there is a perceived fit between core
brand and extension.
McCracken (1989), in the context of celebrity endorsers, points out that the ability
of celebrity endorsers to act as providers for a brand comes from the cultural meaning
with which they are endowed. His argument is that if an endorser does not fit with a
brand, there will be no overlap in cultural meaning and, therefore, no positive impact upon
the acceptor.
Perhaps the best example of the interaction between quality of provider and
associate similarity in the domain of endorsers is found by Friedman and Friedman (1976).
They investigated whether the effectiveness of endorser-type is dependent on the type of
product being endorsed. They find a significant product by endorser-type interaction such
that ads for jewelry were evaluated highest when done by a celebrity, whereas ads for
cookies were better received when done by a homemaker. Thus, the study seems to show
the importance of the fit between endorser and product type.
The common finding of an interaction between provider quality and associate
similarity leads to the following proposal for sponsorship:


30
P18: The more frequently a sponsored brand name is encountered by a
consumer in the context of an event, the more positive feelings s/he has
toward the brand.
P19: The effect of P18 is greater for less familiar brands than for more
familiar brands.
Affect Transfer
Affect transfer in a sponsorship context is analogous to affect transfer in an
advertising context, whereby ones attitude toward the ad transfers to ones attitude
toward the brand, with little cognitive mediation (MacKenzie, Lutz and Belch 1986;
MacKenzie and Lutz 1989). As a mechanism through which sponsorship may function,
affect transfer refers to positive feelings toward an event transferring to the sponsoring
brand through the sponsorship association. This mechanism requires no real cognitive
elaboration. It does not even require conscious awareness of the association, although
such an awareness might be beneficial to the transfer taking place.
Affect transfer is a mechanism which has received some mention in the sponsorship
literature, although much of it is rather vague. In a paper which reports significant
differences in demographic and lifestyle variables among several Hallmark events (major
city-wide destination events), Roslow, Nicholls and Laskey (1992) note that, Hallmark
events are attended by diverse audiences who have come together to enjoy a specific
experience or entertainment. This heightened level of expectation may provide a more
beneficial exposure content than traditional media (p. 54). They elaborate that, It is
probable that this form of promotion would benefit the advertiser from increased levels of
excitement and interest associated with Hallmark events (p. 58). These authors seem to


90
While this correlation was quite high, further investigation of the data attempted to
determine if the two constructs could be orthogonally manipulated experimentally. While
it was possible to find events which differed in prestige and size, these events fell into very
specific event categories. That is, nearly all of the high prestige, low size events were arts
related, while nearly all the high size, low prestige events were dirt-track motor sports or
rodeos. Rather than limit the initial test of the inference-based model to these rather
unique categories of event, where size and prestige were poorly correlated, a decision was
made to include only one of the two constructs. This was done with the knowledge that
the chosen construct is most likely correlated with the other. Given the interesting
theoretical predictions about the impact of event size on the perceived size mechanism, the
perceived event size construct was chosen for exploration.
Perceived Size
Perceived size of the event is expected to impact inferences about the size of the
sponsoring firm. Here, consumers are expected to infer that only a large brand could
afford to sponsor a large event. Perceived size of the event is also expected to affect
inferences about the commitment of the brand (or firm) to supporting the class of activity.
For example, an athletic shoe sponsoring a large tennis match might be perceived as more
committed to tennis than an athletic shoe which sponsors a smaller, less important tennis
match. Perceived size of the event is also expected to impact consumer inference making
about the necessity of the sponsor to the event taking place (event facilitation). Here,
sponsoring a small event is more likely to lead consumers to infer that without the support
of the brand, the event would not take place.


37
(Deshpande, Hoyer and Donthu, 1986) on consumer behavior. At a broader level,
Westbrook and Black (1985) find that desires for affiliations with others is an
important determinants of some peoples shopping behavior. In addition to these
general impacts of affiliation with a group on consumer behavior, the possibility that an
affiliation with a brand or organization can be beneficially developed through the use of
marketing variables has been explored.
Macchiette and Abhijit (1992) describe affinity marketing as the use of group
affiliation to produce a strong promotional program. The resultant affinity groups are
said to have high levels of social bonding and cohesiveness. Members of affinity groups
are also said to be are committed to the objectives of the group. The authors propose
marketing generated affinity as one source of group centeredness.
An example of how marketing can take advantage of peoples existing affiliations
is the affinity credit card. Such cards bear the name and symbol of an organization to
which the card holder belongs. High levels of card holder loyalty have been claimed for
these cards (Worthington and Horn, 1992).
Bhattacharya, Rao and Glynn (1995) investigate the correlates of social
identification with an organization. By defining a museum as the product and members
as customers, the study investigates the determinants of affiliation with the museum.
They find that perceived prestige is positively correlated with ones affiliation with the
organization. Visiting frequency is also found to predict affiliation. Finally, participation
in similar organizations is negatively correlated with affiliation. Based upon these
findings, it can be proposed that


73
fit, such as attribute fit, audience fit and positioning fit were then discussed and
defined. The chapter concluded with a discussion of measurement issues. While this
chapter did address the idea of lack-of-fit, it did not explore the potential outcomes of
associations which are viewed by consumers as inappropriate. This issue is explored in the
next chapter.


APPENDIX I
ONE POSSIBLE FORM OF A SPONSORSHIP VALUATION FUNCTION FOR A
SPECIFIC BRAND
Value = MVSBI + MVE
Where:
Value=Change in brand equity from sponsorship
MVSBI= Marginal Value of Sponsorship Based Inferences
MVE= Marginal Value of Exposure Through Sponsorship
Assume that efficient markets lead to similar costs per 1000 exposures for various
sponsorships. Therefore, exposure per dollar becomes a constant, and deciding which
sponsorship opportunity is best for a given brand reduces to a comparison of the MVSBI
associated with each sponsorship opportunity.
4
MVSBI=Marginal value of sponsorship based inferences
R=Reach (number of people aware of the sponsorship association)
M=Margin (net present value of marginal revenue per customer)
%i=% inference (percentage of R who make inference x)
CF=Choice factor (change in choice probability given inference x)
Inferences:
x=l: size
x=2: legitimacy
x=3: facilitation
x=4: commercial motive
192


150
Independent Variables
Five independent attribute variables were manipulated orthogonally. They were
price (four levels), shelf life (two levels), company size (two levels), event facilitation (two
levels), and calories (two levels). Each is described below (see Appendix F).
Price was set by surveying actual prices for single 6-8 oz. cups of yogurt at a local
supermarket. Four equally spaced prices within the observed price-range were selected.
These prices were $0.47, $0.59, $0.71, and $0.83. Shelf life was manipulated at two
levels. The high level of shelf life was given as 15 days. The low level was 8 days.
Company size and event facilitation were manipulated exactly as in 3 a. Calories were
manipulated at two levels. The low level was 110, and the high level was 160. These are
both within the range observed in the supermarket.
Results
As in 3 a, logistic regression was used to test the impact of the five independent
variables on choice. The logit model regressed price, calories, company size, event
facilitation, and shelf life on choice. Parameter estimates for each variable are provided
below in Table 9.


40
confidence in quality leads companies to spend more on promoting a product. Another is a
perceived relationship between promotional spending and quality in some markets. Finally,
they cite the idea that perceived advertising expense is a signal of a firms financial
strength, probable social acceptance or some other factor that defines quality in some
markets (p. 344). This last mechanism is closest to the idea of perceived size. The idea
is that a company couldnt afford to sponsor a prominent event unless it is a large, stable
company. The authors do not investigate the perceived financial strength possibility,
however, stating that the default attribution is the most interesting possibility and the
one we pursue most directly in this research (p. 346).
Following up this work, Kirmani (1990) proposes that perceived costs will be used
as cues to quality when they are more diagnostic than other cues. She hypothesized that
this should take place under low involvement with the product class. The hypothesis,
however, was not supported. No effect of involvement was found.
Kirmani and Wright (1989) also list several undermines to the positive
relationship between perceived advertising expenditure and perceived quality. They
describe the desperation undermine as occurring when expenditures are so high that the
consumer assumes that the company is desperate to move a bad product. The immunity
undermine occurs when the consumer perceives that the company has nothing to lose, as
is the case with a fly-by-night operation. The no pain undermine is said to occur when a
high expenditure is perceived to be a small part of a companys budget. Finally, they
present the basic premise undermine which occurs when a consumer is confronted with
information which argues against the default attribution.


116
unaided and cued open ended inference measures were taken, followed by closed ended
inference scale measures, manipulation checks and demographic information.
Design
The study consisted of a 2(big/small event size) by 2(high/low brand-event fit) by
2(long/short duration of association) between subjects design. This design is depicted
graphically below:
Big Event
Small Event
Low
Fit
duration
short
duration
High
long
duration
Fit
short
duration
FIGURE 5
8 CELL DESIGN FOR STUDY IB (N=80)
Subjects
Eighty undergraduates in an introductory marketing class at a major southeastern
university took part. Each received course extra credit for participation.
Independent variables:
Three independent variables were orthogonally manipulated. They were
brand/event fit, event size, and duration of association.


BIOGRAPHICAL SKETCH
John W. Pracejus was bom in Bedford, Ohio, a suburb on the southeast of
Cleveland. He was raised with three younger sisters, Joan, Anne, and Sue by his father
Walter, a civil engineer, and his mother Rosemary, a primary school teacher. He spent
his first eight years of school at St. Marys of the Immaculate Conception where he
formed many friendships that last to this day.
He graduated from Bedford Senior High School where he was a member of the
swim team, ski club and National Honor Society. He earned two bachelors degrees from
Miami University, one in business and one in communications. He then worked selling
advertising space and video production services before deciding to return to college to
pursue a masters degree at the University of Illinois. It was there that he was exposed to
the stimulating environment of an active research faculty. To learn more about a career
in academic research and teaching, he asked the advice of Tom OGuinn, who
encouraged John to pursue a Ph.D. in marketing. After completing his masters degree at
Illinois, John joined the University of Florida as a doctoral student. He is currently an
Assistant Professor of Marketing at the University of Alberta in the city of Edmonton.
204


115
however, was not great enough to detect effects on specific effects. The open ended data
did reveal that fit between brand and event might facilitate overall inference making about
the brand. That is, sponsorship association with high fit (Extran) led to more inferences
about the brand than did the association with low fit (Rombouts).
Generalizability of these findings beyond a single type of event is, of course,
unwarranted. By choosing to manipulate fit by holding event constant and varying the
sponsor, fit is essentially confounded with brand. One way to increase the generalizability
of the findings would be to choose both another event and a set of two new brands. Here
again fit would be confounded with brand. Another approach is to find another class of
event which has a high fit with coffee, and a low fit with sports drinks. If the results
replicate in this case, both the overall generalizability, and the internal validity of the fit
construct, will be enhanced. Study lb, then, attempts to replicate the findings of study la
with an event for which coffee has a high fit and sports drinks have a low fit. Since no
evidence of the share of presence manipulation was found, this factor is eliminated from
the design of study lb. A stronger manipulation of share is attempted in study lc.
Study lb
Overview
This study attempts to replicate the findings of study la using a different event
category for which fit is reversed (see above). It again tests the impact of event size,
brand/event fit and duration of association on inferences about a sponsoring brand. A
completely between subjects 2x2x2 factorial design exposed subjects to an unfamiliar
brand in the context of a poster of a similar format to that used in study la. As before,


52
order to make sense out of a group of objects which could not easily be separated by
preexisting, taxonomic categories. Barsalou found that frequency of instantiation predicted
graded structure in both goal derived and taxonomic categories to a large and
approximately equal extent. This provides evidence that, to the extent to which seeing a
brand paired with an event leads one to create an ad hoc category (Barsalou 1983), in
order to make sense out of what is being seen, frequently and recently seeing this brand-
event pairing should enhance attribute fit. Noting the continuous nature of this
relationship, Barsalou explains that the strength of association is free to vary
continuously as a function of how frequently and recently an association has been active in
working memory (p. 212).
P40: Frequent attendance at an event leads to greater salience of common
attributes.
F41: Recent attendance at an event leads to greater salience of common
attributes.
Markman and Gentner (1993) argue that mental representations (which could
include degree of fit between two objects) are formed through structural alignment
which is derived from analogical reasoning. They find support for the proposition that
the likelihood of a structural alignment between two stimuli is greater after a similarity
comparison than before (p. 435). It seems, then, that consciously considering the
similarity between two objects can shift attention away from the obvious, surface
dimensions of fit, to the more relational dimensions of fit. Since conscious consideration
of the similarity between the sponsor and the event is something which may be facilitated


81
Gardner and Shuman (1987) empirically investigate the amount spent on leverage.
They survey 300 Fortune 500 companies about their sponsorship activities. Eighty-nine
percent of responding companies engaged in sponsorship report some form of project
support, such as advertising tie-ins. The authors find that within their sample, an average
of forty percent of sponsorship cost was spent on leveraging the sponsorship (including
advertising, promotion, and employee tie-ins). They also find that less than ten percent of
companies spend more money leveraging the sponsorship than the sponsorship actually
costs.
Although the optimal level of spending on leverage seems to be a matter of some
dispute, the facts seem to support the idea that most sponsorships involve some additional
spending on communications about the association. The interesting question then
becomes what should be said and to whom? Crimmins and Horn (1996) propose that
when the link between an event and a brand is clear, as with basketball and athletic shoes,
consumers can make the link themselves. When the link is harder to forge, however, the
link must be made for them, as with Visa driving home the message that the Olympics
dont take American Express. While this advice seems to be face valid, it does not answer
the biggest question. When does the meaning of the association need to be driven
home? When does this help? When might it hurt? When might it be merely a waste of
money?
In the related area of brand extensions, Aaker and Keller (1990) find that
elaborating on brand extension attributes improves evaluations of the extension. They
propose that this occurs by alleviating negative inference making. Explaining which


61
Another consequence of attribute similarity may be a tendency to infer that one
associate has the attributes of the other associate. In other words, attribute fit between
sponsor and event along several attributes may lead to inferences that a brand possesses an
additional attribute of the event, which the brand is not known to possess. There is
evidence to support the idea that two things which are strongly associated can lead to
inferences about shared attributes which are not known to be shared. Sedikides, Olsen
and Reis (1993) investigate whether providing information about the relationship between
two individuals (that they are married) would lead to greater spontaneous transference of
attributes (e.g., is vegetarian) during recall than for pairs of non-married individuals.
They do, in fact, find that for married couples, more attributes of one spouse are
mistakenly attributed to another spouse than is the case for non-married pairs of people
(p<05). The implications of these findings for sponsorship might be
P61: The tendency to transfer attributes from event to sponsor occurs
more frequently under high attribute fit than under low attribute fit.
In addition to investigating possible consequence of fit, it is necessary to consider
possible consequence of lack-of-fit. Within the domain of investigating complex social
conjunction categories, Hastie, Schroeder and Weber (1991) cite evidence which suggests
likely char acteristics of attributions made about compound categories which are seemingly
incongruent. In this paradigm, the authors compare conjunctive categories formed by
pairing either congruent or incongruent attributes of a person. They find that attributions
about the incongruent compounds are more extreme along trait adjective scales than are
either the original components of the pairing or the congruent compounds. They also find


9
distribution and word of mouth), the numbers are impressive. During the period of the
experiment, top of mind awareness went from 4% to 21%, which is a 425% increase.
Unaided recall went from 38% to 68%, and aided awareness went from 92% to 99% (p.
65). The author notes that while the impact of sponsorship on awareness was almost
immediate, the impact on brand preference took a longer time.
One paper attempted to address determinants of sponsorship awareness. Hansen
and Scotwin (1995) reported that in an experimental setting, those interested in soccer
(measured) had higher awareness of the team sponsors than did those who were not
interested. The authors, however, did not report whether these differences are statistically
significant.
Other impacts of sponsorship have also been reported such as beliefs about the
company. Javalgi et al. (1994) studied perceptions of five local companies and awareness
of their sponsorship activities. They hypothesized that persons who are aware of corporate
sponsorship have a more favorable view of the sponsoring company than persons who are
unaware. This was not supported. Awareness of the sponsorship activity was only related
to the statements is involved in the community and only wants to make money
(reversed), and neither was significant overall. Is involved in the community was
significantly correlated with awareness of sponsorship for only one of the five companies.
In a study which did not directly involve sponsorship, Nebenzahl and Jaffe (1991)
found that evaluation of Korean products by Israeli consumers was higher after the 1988
Olympics in Seoul than it was before the games among heavy Olympic viewers, but not
among a low viewing control group. This suggested to them that an association with
the Olympic games had raised peoples evaluations of Korea and in turn Korean products.


136
alpha for these three items was over .85, so the three items were combined into a single
measure of perceived event facilitation.
Because commercial motive inferences emerged in previous open ended data, a
closed ended measure of commercial motive inference-making was added. A seven point
scale measured agreement with the statement Evian (Texaco) is just trying to get a lot of
people to see their name. This scale was anchored by strongly disagree and strongly
agree. A manipulation check was then taken for fit between the brand and the US Open,
as was a measure of the perceived size of the event.
Subjects then rated their involvement level with tennis using a twenty item form of
the Personal Involvement Inventory, or PII (Zaichkowsky, 1985). Instructions for
completion of the PH were as follows:
Now, we would like you to please take a moment (approximately 5-10 seconds) to
think about tennis and what it means to you. Then complete these scales with
regards to tennis.
The scales were anchored by items like important/unimportant, trivial/fundamental,
and baring/fascinating.
Subjects then rated their subjective knowledge about tennis. They expressed their
level of agreement with the statement I know a great deal about the game of tennis. The
scale was anchored by strongly agree and strongly disagree.
Subjects then completed the intrinsic source of personal relevance, or ISPR scale
(Celsi, Chow, Olson, and Walker, 1992), adapted to tennis. Agreement was measured
with five items along a seven point strongly agree/strongly disagree scale. The five
items were playing tennis lets people see me as I would ideally like to be seen, tennis
helps me attain the type of life I strive for, I can make connections or associations


149
TABLE 8
IMPACT OF FIVE MANIPULATED VARIABLES ON BINARY LOGIT MODEL OF
AUTO INSURANCE CHOICE
Parameter
Estimate
SE of
Parameter
Asymptotic
t-Stat
Pr(Z>|t|)
intercept
.21532644
. 86447026E-01
2.491
.0127
price
-.23668224E-01
. 21119801E-02
-11.207
.0000
deductible
-. 27158952E-02
.2593348IE-03
-10.473
.0000
facilitation
.68801263E-01
.57565562E-01
1.195
.2320
company size
.50404886
.12435944
4.053
.0001
service hours
.10788367E-01
.11847266E-02
9.106
. 0000
As can be seen in the table, four of the attributes had a significant impact on
choice. The sign of the parameter estimate indicates direction of this impact. Therefore,
in this data set, as price and deductible increase, choice decreases. As service hours, and
company size increase, choice increases. Event facilitation, however, had no detectable
impact on choice for auto insurance.
Study 3b
Overview
Subjects made a series of eight choices between two brands of yogurt (1,248 total
choices). Each choice profile consisted of five brand attribute (including size and event
facilitation) for two yogurt brands. As in 3a these attributes were manipulated
orthogonally. Each subject completed half of a 16 cell, radical fractional factorial design.
Subjects
One hundred fifty-six undergraduates in an introductory marketing class at a major
southeastern university took part. Each received course extra credit for participation.


56
things, do this very thing (e g., IEG Sponsordex software). While such a database
might prove useful as a decision aid, it seems quite reasonable (and not unimportant) to
propose that
P48: Attribute fit between sponsors and events is not transitive.
A final property of similarity proposed and demonstrated by Tversky is its
relationship with dissimilarity. Similarity has generally been found to be strongly
negatively correlated with dissimilarity, but people attend more to the common features
in judgment of similarity than in judgment of difference (Tversky, 1977, p.340). As a
result, prominent pairs (e.g., East Germany and West Germany) are found to be more
similar than non-prominent pairs (e g., Ceylon and Nepal) in a similarity condition and
more dissimilar than non-prominent pairs in a dissimilarity condition. This effect has
also been demonstrated in a marketing context. Johnson (1981) finds that Coke and Pepsi
are very similar in the context of similarity, and very dissimilar in a dissimilarity context.
This suggests that
P49: When an event has many salient associations which are similar to
associations of the brand as well as many salient associations which are
incongruente the magnitude of similarity depends on the frame (fit vs. lack-
of-fit) such that the sponsorship association is perceived as having either high
fit or high lack-of-fit.
Before considering the consequences of attribute fit, it should be noted that
similarity has received some criticism for representing at best an intermediate measure for
predicting category membership, and at worst a midpoint in a circular argument. Murphy
and Medin (1985) point out that any two entities can be arbitrarily similar or dissimilar by


50
possess incongruous attributes. By attributes are meant associations with the brand or
event which are commonly elicited in the minds of consumers when thinking about the
event or brand. These associations can vary along a concrete-abstract continuum. An
example of a concrete attribute shared by sponsor and event is automotive. Speed is
a more abstract association, and macho is more abstract still.
This definition closely resembles Tverskys (1977) definition of similarity, which
increases with the addition of common features or the deletion of distinctive features.
More will be said of the relationship between similarity and attribute fit. First,
however, possible antecedents of attribute fit are discussed.
One thing which should influence attribute fit is the frequency and recency with
which two things are experienced together. Barsalou (1982), for example, investigated
the ways in which context-dependent properties of an object might be made to be more
context-independent. In other words, he explored how situation-dependent associates of
an object can be generated more spontaneously. He found that repeated simultaneous
exposure to the context-dependent property of an object can make the property more
context-independent.
An example of a context dependent property which can be made to be more
context independent is the ability of a basketball to float (Barsalou 1982). Most people,
when asked to lists things associated with basketball would not list its ability to float.
People who had frequently and or recently seen a basketball floating in a pool, however,
would be much more likely to come up with this association. This leads to the following
proposition:


185
EXPERIMENT 1A ALL LOW
The Leuven
Student Fun-Run
Leuven, Belgium
Sunday October 19th 1997 at 9:30 am
Proud to be a New Sponsor


165
TABLE 12
TWO HYPOTHETICAL SPONSORSHIP PROPOSALS
ProDosal
Total # exDosed
Cost
#1-1 event perceived to be large
10,000,000 people
$110,000
#2 20 events perceived to be small
10,000,000 people
$100,000
Given the two proposals above, a brand in a similar situation to one of the domestic auto
insurance companies in study 3 a should probably choose proposal #1, since the brand size
variable had a significant impact on choice while the event facilitation had no measurable
impact in that market. Given the lower cost, however, a brand in the situation of the
yogurt brands in study 3b, should probably choose proposal #2. A brand which is strongly
impacted by event facilitation inferences, but not by brand size, would clearly choose #2.
This example should make two things clear. The first is that a good deal has been
learned about how sponsorship impacts brand choice. The second, however, is that there
is still much work to be done on quantifying this impact. Potential avenues for improving
quantitative rigor are presented below.
Future Research
The research presented in this dissertation represents a first step toward
understanding how to optimally include sponsorship in an overall marketing campaign. It
has been shown that sponsorship, the promotional medium that conveys no overt
message, is in fact capable of generating inferred messages in the minds of consumers. It
has also been demonstrated that these inferences can impact brand choice, and that this


84
one another. Specifically they state that since categorization produces a reduction in
uncertainty, positive affect may result from a successful fit and negative affect from an
inability to categorize an item, particularly if the resulting judgments or inferences are
important (p. 470). Combining this idea with the findings of Linder and Worchel (1970),
it can be proposed that
Pill: For those who determine how a sponsor is like an event on their own,
the more difficult the association is to understand, the more positive affect
will be generated for the association.
P82: The effect predicted in P81 is greater for consumers who are highly
involved with the event than for those who are less involved with the event
The above propositions must be qualified, however, by the generally accepted idea
that for most people, figuring out what a sponsor has to do with an event is not an activity
likely to receive much cognitive effort. Therefore, any association that requires more
than the simplest consideration to understand will likely be ignored. This fact does not,
however, make the above discussion irrelevant. On the contrary, it allows some specific
predictions about the effectiveness of certain strategies for association-enhancing
communications. Some specific AEC strategic proposals are as follows:
P83: Communications which explain obvious associations harm the
association.
P84: Communications which all but explain the associations that are not
oibvious are superior to communications which actually do explain the
association.


44
almost always purchase a sponsors product over a closely priced competitor. Reported
reciprocity for the Chicago Gospel Festival was even more impressive. As measured by
it makes me more likely to buy, fan reciprocity here was 82%. The authors go on to
point out that it is unlikely to get these effects from casual observers or uninvolved
spectators. Simply stated, they claim that Fans are grateful for sponsorship. Others are
not (p. 17). From these observations, it is possible to propose:
F34: The more involved consumers are with an event, the more likely the
sponsorship is to engage reciprocity.
Conclusion
Any attempt to study sponsorship must consider the multiple ways that an
association of a brand with an event can impact consumers. Seven mechanisms and their
psychological underpinnings were proposed and discussed in this chapter By breaking
sponsorship into these component mechanisms, it becomes possible to posit and test how
various aspects of a sponsorship association will impact success. Since there are multiple
ways in which sponsorship may work, considering which mechanisms are likely to be
engaged under various types of sponsorship associations allows many specific hypotheses
to be developed and tested. Empirical testing of numerous hypotheses relating to how
audience, event, and associational factors impact the functioning of these mechanisms is
presented in chapters eight, nine and ten. A thorough exploration of one important
associational factor, fit, is presented in the following chapter.


48
not seen as logically incompatible, the resultant composite brand will have a good
likelihood of success. Key to their definition of good fit are complementary attributes
which are not logically inconsistent. In other words, sharing attributes is not necessary.
Here they argue that so long as the attributes are logically consistent, non-overlapping
attributes help the combination. Implications of this idea are discussed in detail in the
attribute fit section.
Fit has also received attention in the celebrity endorsement literature. While several
empirical outcomes of fit have been documented, little attention has been devoted within
this donum to defining or measuring things which lead to fit. Kamins and Gupta (1994),
for example, found that increased congruence between the spokesperson and the product
resulted in a more favorable product attitude. Speck, Schumann and Thompson (1988)
found that when the endorsers relationship to the product is more obvious, information
can be processed in a shorter time.
Ryssel and Stamminger (1988) report that Adidas has engaged in some interesting
measurements of the perceived personalities of various tennis stars, and how these
measures fit with the Adidas personality. They do not report in this paper whether or not
they measure brand personality along similar scales to those used to evaluate the celebrity,
but the article gives the impression that the company uses some sort of mechanism for
comparing celebrity personality with brand personality for fit. This points out the
interesting possibility that fit can include more abstract attributes, like personality
characteristics.
Despite its frequency of mention in the above literatures, there has been almost no
serious attempt to systematically define or measure the fit construct. There is little


123
Big Event
Small Event
Low
Fit
High
Fit
FIGURE 7
4 CELL DESIGN FOR STUDY 1C
Independent Variables
Two independent variables, event size and brand/event fit, were orthogonally
manipulated. Fit was manipulated exactly as in study la Associating a running event with
a coffee represented low fit and associating a running event with a sports drink
represented high fit. The small size event was again a student ftm run, but this time it
was either the Rombouts student fim-run or the Extran student fun-run The large size
events were the Rombouts Euro-Capital International Marathon or the Extran Euro-
Capital International Marathon.
Results
Manipulation Checks
Using the same scales as before, the high fit mean was 5.56 and the low fit mean
was 3.13 on the fit check measure. This difference is significant (F=43, p<001). The


169
Conclusion
The research reported here provides support for an impact of sponsorship on brand
perceptions, and ultimately on brand choice. Such findings have not been demonstrated
previously. Of special interest, evidence was found that the perceived size of a sponsoring
brand can be influenced by the size of the sponsored event and by the duration of the
association. Perceptions that the brand is helping the event, however, are shown to be
enhanced by title sponsorship, sponsoring small events, and being a sponsor for a long
time. These findings are especially important because it also was found that perceptions of
brand size and perceptions of helping the event are not equally important to all brands.
Taken together, this set of findings suggests the sponsorship association which will
be most beneficial can only be determined once the brand situation is known. Once the
value of changing different brand perceptions has been determined, the research presented
in this dissertation can provide some general guidelines for optimal sponsorship choice.
Future research in this area should make these recommendations much more specific.
Some academic researchers and managers still perceive sponsored events as little
more than billboards for brand logos. They consider sponsorship to be exclusively an
exposure vehicle which is only capable of building top-of-mind awareness. The
conceptual framework and empirical evidence presented in this dissertation, however,
suggest much richer possibilities for the promotional medium of sponsorship.


135
important to me. Item two was The poster didnt have anything to do with me or my
needs.
TABLE 8
PERCEIVED FIT BETWEEN THE US OPEN AND 19 ACTUAL SPONSORS
(N=46)
Brand
mean
stdev
Tyco
2.46
1.38
Texaco
2.80
1.42
Mass Mutual
3.02
1.53
Tiffany
3.13
1.92
Heineken
3.57
2.08
Prudential
3.59
1.94
Chase
3.80
1.64
NY Times
4.11
1.72
Infiniti
4.33
1.94
IBM
4.41
1.94
Citizen
4.87
1.73
AmEx
5.35
1.59
Pepsi
5.80
1.24
Canon
5.91
1.26
Fuji
5.98
1.58
Fila
6.57
0.93
Evian
6.80
0.45
Tennis Mag
6.91
0.35
Wilson
6.93
0.33
Subjects then performed a set of thought listing tasks identical to those in la-lc.
Following this, they filled in the closed ended inference measures. Because event
facilitation inferences were key to this study, a three item scale was developed to measure
perceived event facilitation. All three items were anchored by not important at all and
very important. The three items were how important do you think the support of
Evian (Texaco) is to the US Open Taking Place, how important do you think the
support of Evian (Texaco) is to the success of the US Open, how important do you
think the support of Evian (Texaco) is to the quality of the US Open. The Cronbach


APPENDIX H
SAMPLE EXPOSURE STIMULI FOR EXPERIMENTS 1 A, IB, 1C,2A AND 2B
This appendix contains example of the posters used to expose subjects to the
sponsorship associations in the first five experiments. Two examples are given for
experiments la, lb and lc. One representing the condition in which manipulated
variables are all high (i.e. large event size, high fit, long duration of association and
high share of presence) and one condition in which all manipulated variables are low (i.e.
small event size, low fit, short duration of association and low share of presence). The
high fit conditions are given for 2a and 2b.
183


163
The most robust finding of sponsorships impact on brand perception is that
associating a brand with a larger event leads to perceptions that the brand is large. This
result is found across three studies (la,lb and lc), for both sports (running) and cultural
(literature) events. These results replicate experimentally what Rajaretnam (1994) found
with a pre-test/post-test field study. Taken together, these findings seem to give strong
support to the notion that associating a brand with large events can lead to the perception
that brand itself is large. This impact of sponsorship on brand perceptions is likely to be
more helpful to some brands than others (see below).
Another result to replicate across multiple studies (la and lc) is that associating a
brand with a larger event leads to an increase in perceived product quality. While this
effect initially appeared to be stronger under title sponsorship, follow-up analysis showed
no significant enhancement from the event being named after the brand. This impact of
event size on perceived brand quality obtained across both brand categories (coffee and
sports drinks) and both event categories (running and literature).
Perceived brand size appears to also be affected by duration of association
between the brand and the event. In study la, duration had a significant, positive impact
perceived brand size. A comparison of the pooled, open ended data from studies la and
lb replicated this effect. This is important, because it implies that the benefit to the brand
of sponsoring an event is enhanced over time. While media departments in advertising
agencies are constantly looking for new, more cost effective means to reach target
audiences, these results suggest that for sponsorship, a longer term approach is more
beneficial.


190
EXPERIMENT 2 A HIGH FIT
Win a Trip to
The 1998 US Open
August 24th-September 5th
In New York City
See Bottle for Details
Evian is proud to be an official sponsor of the 1998 US Open


113
commercial motive inferences included negative inferences about the brand because of the
association, such as this is a waste of money or probably raise price to pay for event.
Each rater independently coded all of the thoughts. Initial agreement between
raters was quite high. Out of a total of 826 thoughts, the coders matched on 784 (94.9%
of the time). The remaining discrepancies were reconciled through discussion and mutual
agreement. Of the 89 subjects, 29 made a total of 46 inferences about the sponsoring
brand. The other 60 subjects listed no thoughts which could be considered inferences
about the brand.
The fact that about one third of subjects who saw a poster for an event listed an
inference about a sponsoring brand is encouraging. With 46 total inferences made,
however, it was unlikely that enough power would be available to test the impact of the
four independent variables on each of the four kinds of inferences. Given that number of
inferences is count data, a series of log-linear models were run to determine if any impact
of fit, size, duration, share and the three fit interactions could be detected on each of the
four inferences. As expected, no significant results were found.
In order to test whether any impact of the independent variables on inference
making could be detected, the data was re-coded to create a new variable total
inferences. This variable was a simple summation of the four individual types of
inference. Individual loglinear models tested the impact of fit, size duration and share on
the total number of inferences made. This analysis revealed a strong main effect of fit on
overall inference-making (Chi Sq.=7.02, p<01). The direction of this effect was that
subjects exposed to high fit associations made more inferences (Mean =.72) than did


99
Study la
Overview
This study tests the impact of event size, brand/event fit, duration of association
and share of presence on inferences about a sponsoring brand. A completely between
subjects 2x2x2x2 factorial design exposed subjects to an unfamiliar brand in the context of
a poster for an event where the brand was a sponsor (see Appendix H). Unaided and cued
open ended inference measures were taken, followed by closed ended inference scale
measures.
Design
The study consisted of a 2(big/small event size) by 2(high/low brand-event fit) by
2(long/short duration of association) by 2(high/low share of presence) between subjects
design. This design is depicted graphically below.
Big Event
Small Event
high
share
low
share
high
share
low
share
Lowt
long
duration
Fit
short
duration
High
Fit
long
duration
short
duration
FIGURE 4
16 CELL DESIGN FOR STUDY 1A (N=89)


199
, Scott B., Richard J. Lutz, and George E. Belch (1986) The role of attitude
toward the ad as a mediator of advertising effectiveness : a test of competing
explanations Journal of Marketing Research. 23, 2, 130-143
Markman, Arthur B., and Dedre Gentner (1993) Structural alignment during similarity
comparisons Cognitive Psychology. 25, 4, 431-467
Marshall, D. W., and G. Cook (1992) The corporate (sports) sponsor. International
Journal of Advertising. 11, 307-324.
McCracken, Grant (1989) Who is the celebrity endorser? cultural foundations of the
endorsement process. Journal of Consumer Research, 16, 3, 310-321
McDonald, Colin, (1991) Sponsorship and the image of the corporate sponsor.
European Journal of Marketing. 25,11,31-38
Meenaghan, Tony (1991a) The role of sponsorship in the marketing communications
mix. International journal of Advertising. 10, 35-47
(1991b) Sponsorship-legitimizing the medium European Journal of Marketing.
25, 11, 5-10
(1994) Ambush marketing: immoral or imaginative practice? Journal of
Advertising Research. 34, 5, 77-88
Meerabeau, Elizabeth, Roy Gillett, Michael Kennedy, Johnson Adeoba, Michael Byass,
and Kingsley Tabi (1991) Sponsorship and the drinks industry in the 1990s.
European Journal of Marketing. 25,11, 39-56
Mervis, Car olyn B., and Eleanor Rosch (1982) Categorization of natural objects.
Annual Review of Psychology. 32, 89-115
Meyers-Levy, Joan, and Alice M. Tybout (1989) Schema congruity as a basis for product
evaluation. Journal of Consumer Research. 16, 1, 39-54
Misra, Shekhar, and Sharon E. Beatty (1990) Celebrity spokesperson and brand
congruence: an assessment of recall and affect. Journal of Business Research. 21,
2, 159-173
Murphy, Gregory L., and Douglas L. Medin (1985) The role of theories in conceptual
coherence. Psychological Review. 92, 3, 289-316
Nebenzahl, Israel D., and Eugene D. Jafife (1991) The effectiveness of sponsored events
in promoting a countrys image. International Journal of Advertising. 10, 223-237


32
(1990) found that the transfer of affect from spokesperson to brand was facilitated when
the two were matched. In the domain of brand extensions, Rangaswamy, Burke and
Oliva (1993) propose affect transfer through categorization as a method by which
utility of the brand name impacts utility of the product.
Affect-transfer-like mechanisms are also found in the retail assortment domain.
Jacoby and Mazursky (1984), in a factorial design, associate high and low quality brands
with high and low quality stores. They attempt to determine how differential affect (i.e.
positive affect toward one, negative affect for another) in the associate pair interact. Weak
support is found for an averaging process.
Like simple awareness, the cognitively simple mechanism of affect transfer allows
basic propositions to be made:
F20: The more positive the affective response to the an event, the more
positive the affective response to a sponsoring brand.
P21: The stronger the associative linkage between the brand and the event,
the stronger is the impact posited in P20.
Image transfer
The aim of sponsorship is image transfer: the transfer of the personality, aura
and competence... to the brand Ryssel and Stamminger (1988, p. Ill)
While affect transfer is the process of assigning event-based affective response to
the brand, image transfer refers to the assignment of abstract associations of the event to


97
Legitimacy
Inference about the legitimacy of a company or brand occur when a consumer changes (or
forms) his/her perception of whether this brand is reasonable or of some minimum
quality because of its association with an event. Sponsorship associations are expected to
lead to such legitimacy inferences when a consumer perceives some gatekeeping function
on the part of the event, so as to not allow shoddy, or substandard products or companies
to be sponsors. Assuming large events have more to lose from an association with a
shoddy product:
H5: Sponsoring larger events leads to more brand legitimacy inferences on the
part of consumers
Event gatekeeping should be tighter in the long run than in the short run. As a
result, long term associations should imply greater endorsement of the brand by the event.
Therefore:
H6: Longer duration of association between brand and event leads to more brand
legitimacy inferences on the part of consumers
Events should be more scrutinizing gatekeepers for major, as opposed to minor sponsors.
Therefore:
H7: Engaging in a sponsorship with a large, as opposed to small, share of
presence leads to more brand legitimacy inferences on the part of consumers
Event gatekeeping should be higher for more related products than for less related
products. That is, an event should be more careful not to allow sponsorship by a shoddy
event-related product than by a shoddy, non-event-related product. For example, a tennis


This dissertation was submitted to the Graduate Faculty of the Department of
Marketing in the College of Business Administration and to the Graduate School and
was accepted as partial fulfillment of the requirements for the degree of Doctor of
Philosophy.
August 1998
Dean, Graduate School


14
endorser domain, Kahle and Homer (1985) find that the better (more attractive) the
provider (the endorser) the more the acceptor (product) is liked. This same effect has been
found when the provider is a speaker, and the acceptor is an argument (Horria, Naccari
and Fatoullah, 1974; Chaiken, 1979).
As stated above, there are many different mechanisms through which provider
quality impacts acceptor quality. The specifics of these mechanisms are discussed in
future chapters. In the most general terms, however, it can be proposed:
PI: The more highly regarded the event, the more highly a consumer will
evaluate the sponsor.
P2: The more well liked the event, the more well liked will be the sponsor.
An equally frequent finding in the literature in commercial associations is that the
more qualitatively similar the two associates, the more benefit will come to the acceptor
from the association. Aaker and Keller (1990) find this to be true in the brand extension
domain Beyond simple qualitative similarity, however, they define fit dimensions along
which extended brands (acceptors) may match up with the original branded product
(provider). Specifically, they find that if the acceptor is a complement of, or a substitute
for, the provider, the acceptor will be better received. Similarity between provider and
acceptor, however, can take place at a much higher level of abstraction. Park, Milberg
and Lawson (1991), for example, find that extensions work best when both core brand
and extension are either prestige or functional in nature. Benefits to the acceptor of
qualitative similarity with the provider have also been frequently demonstrated for


APPENDIX B
OPEN ENDED INFERENCE MEASURES AND MANIPULATION CHECKS FOR
EXPERIMENT 2A AND 2B
Now we would like to get your impression about EVIAN (TEXACO) (TEXACO) and the US Open
(Indy 500) (Indy 500)
How important do you think the support of EVIAN (TEXACO) (TEXACO) is to the US Open (Indy 500)
(Indy 500) taking place?
Not at all important I I I 1 | | 1 | Very important
How important do you think the support of EVIAN (TEXACO) is to the success of the US Open?
Not at all important I I I 1 | | | | Very important
How important do you think the support of EVIAN (TEXACO) is to the quality of the US Open?
Not at ail important 1 I I I I 1 I I Very important
Please indicate your level of agreement with these statements:
EVIAN (TEXACO) must be a good water, or it wouldnt be able to be a sponsor of this event
strongly disapee I I I I I I I | strongly apee
EVIAN (TEXACO) is demonstrating good corporate citizenship through its support for this event
strongly disapee I 1 i 1 I I 1 | strongly agree
EVIAN (TEXACO) is just sponsoring this event to get a lot of people to see their name,
strongly disagree I I 1 1 I I I I strongly agree
173


39
Phillips and Cadbury). He posits that had the sample included more smaller companies,
an impact of sponsorship on perceived size might have been found.
Evidence that sponsorship can influence perceived size was found by Rajaretnam
(1994). In the evaluation of the unique experiment where a tire maker spent almost its
entire marketing budget from 1984-1987 on sponsorship, perceived size of company
increased by 19.7% during the period. Perceived financial health of the company, a
related measure, increased by 24.6% (p. 68). While this field experiment with no control
group does not prove that sponsorship can impact the perceived size of the firm, it
certainly allows for the following proposals:
P28: Event sponsorship influences the perceived size of a sponsoring firm.
P29: The larger the perceived size of the event, the greater is the impact of
the sponsorship on the perceived size of the firm.
Anecdotal evidence of the size of the sponsor affecting the perceived size of the
event have also been reported. Parker (1991) proposes that Mars sponsorship of the
London Marathon was the key to getting people to believe that it was a real event. The
following can therefore be proposed:
P30: Sponsorship of an event by a large sponsor increases the perceived size
of the event.
The marketing literature provides evidence of the benefits of perceived size.
Kirmani and Wright (1989) propose three possible mechanisms through which perceived
expenditure on advertising might lead to inferred quality. One is the notion that


68
target market for a brand X with event Z of 100,000, regardless of the absolute size of Y
and Z, or the relative size of event audience and brand market (see Figure 2 below). For
several of the seven mechanisms, intersection is probably the only component of audience
fit that matters.
The second component of audience fit is subsumption, or the degree to which
one audience is subsumed by the other. Here the relative audience size of the two events
does make a difference, as does the percentage of each audience which is represented in
the intersection. A few examples of this construct are given in Figure 2 for clarification.
/ GMC (
\ indy \ Equivalent
l Trucks l
/ 500 J Subsumption
/ GMC f
Y\-MudBog Event
1 Trueles
jj~'Truck luce Subsumption
Brand
Holly
Carburetors V\_
) ^ j Subsumption
FIGURE 2
TYPES AND AMOUNTS OF SUBSUMPTION
As can be seen in the above figures, the type and degree of subsumption depends
upon the area of intersection between the brand and the event, the size of the event, and
the size of the brand. Equivalent subsumption is defined by the intersection representing
the same proportion of the event audience as the brand audience. Event subsumption is


25
Park, Jaworski, and Maclnnis (1986) have conceptualized positioning as
management of brand image. They present a framework for managing brand image at
different times in a brands life cycle. They define brand image as the understanding
consumers derive from the total set of brand-related activities engaged in by the firm
(p. 135). They stress the importance of various brand associations on brand image, stating
that each element of the (marketing) mix can affect the inferences consumers develop
about the brand (p. 138). Therefore,
P15: Whenever sponsorship of an event causes a change in the associations a
consumer has with the brand, brand image changes, and positioning within
multidimensional space is altered.
Crawford (1985) proposes that the primary function of product positioning is to
differentiate among products. He defines a product positioning typology in which three
things, product features, benefits, and surrogates, can be used for positioning. Of
relevance to sponsorship is the surrogates positioning strategy. He lists coming from
respected source or endorsed by a respected personality as tactical means to employ a
surrogates positioning strategy. While sponsorship is not mentioned per se, it seems
reasonable that an association with a respected event or organization should also aid ones
surrogate positioning strategy. Therefore,
P16: To the extent that a sponsorship represents a unique association, it is
exploitable as a differentiating positioning strategy.
Finally, DiMingo (1988) points out that product positioning is a process of
distinguishing ones company or product from competitors along real dimensions which


5
extent and 36% said it was not altruistic (p. 309). Apart from this one study, however,
the issue of altruistic motives in sponsorship is unexplored.
Armstrong (1988) presents the results of a case study analysis of the sponsorship
activities of 17 large international electronics companies. Interviews were conducted in
four countries (United States, Great Britain, Germany, and France) with corporate and
regional marketing managers from the companies, representatives of the sports being
sponsored and any intermediary agencies involved (80 interviews total). Within the
sample, American firms more frequently listed business managers as the primary target for
the sponsorship, while European and Japanese firms tended to use sponsorship to reach
the consumer audience. Interestingly, the demographics of these two target audiences
were nearly identical (male, upper management, age 35-55). The interviews also revealed
that firms which have been involved with a sponsorship for several years tended to have
image enhancement (as opposed to awareness) goals. It is unclear from the data presented
in the paper, however, whether companies changed their focus over time, or if companies
with awareness goals stopped engaging in sponsorship after awareness goals were met.
Marshall and Cook (1992) asked why 200 British companies from the Times 1000
engaged in sponsorship. Reasons given were reach target audience (49%), product
linked (34%), image of sport (14%) and other(3%) (p. 319). They also asked for
importance ratings for a number of potential sponsorship goals. Mean importance ratings
for building corporate image, national link with product or service, potential
advertising exposure, potential TV coverage, potential press coverage, and covers
target audience indicated that all were important or very important. Only ability to


I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and
quality, as a dissertation for the degree of Doctor of Philosophy
Richard J. Lutz, Chairman
Professor of Marketing
I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and
quality, as a dissertation for the degree of Doctor of Philosophy
Barton A. Weitz
J. C. Penney Eminent Scholar
of Marketing
I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and
quality, as a dissertation for the degree of Doctor of Philosophy
Associate Professor of Marketing
jCCi.'i, .'Csl
Christopher^. Jarus^ewski
I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and
quality, as a dissertation for the degree of Doctor of Philosophy
Michael F. Weigold
Associate Professor of
Journalism and Communications


112
at all. The other, called event helping, attempted to measure inferred importance of the
brand to the event being a success.
Event enabling was measured by asking how important do you think the support
of EXTRN SPORTS DRINK (ROMBOUTS COFFEE) is to the sponsored event taking
place. The scale was anchored by not important at all and very important. The
independent variables had no significant impact on brand enabling.
Event helping was measured by asking How important do you think the support
of EXTRAN SPORTS DRINK (ROMBOUTS COFFEE) is to the success of the
sponsored event. The scale was anchored by not important at all and very important.
The independent variables were regressed on brand helping. Weak support was found for
H10. A long duration of association with an event seems to have some effect on event
helping inferences. This effect was not statistically significant at traditional levels (F=3.02,
p<09). Counter to prediction, this effect was not moderated by fit (F=1.22, p>25)
Therefore, HI2 is not supported. No support was found for H9 or HI 1. Neither event
size nor share of presence had any measurable impact on event facilitation inferences.
Open Ended Inference Measures (Thought Listings)
The thoughts in the thought listing boxes were coded by two independent raters
who were naive to both experimental condition and hypothesized relationships. Initially,
the thoughts were to be coded only for brand size, brand legitimacy and event facilitation
inferences. An iterative process of looking at subsets of the data and updating led to an
additional class of inference-making which was included in the final coding scheme. The
additional inference class included was termed commercial motive. Thoughts coded as


APPENDIX F
SAMPLE CHOICE PROFILE USED IN EXPERIMENT 3B
Brand G
Yogurt
Brand K
Yogurt
Shelf Life
8 Days
15 Days
Company Size
Large
Company
Small
Company
Company involvement
with favorite local events
Not involved
Helps make
them possible
Calories
160
110
Price for 8 oz. container
$0.59
$0.71
U
JJ
Please indicate your
choice (check only one)


181


10
Another impact of sponsorship reported in the literature is purchase intention. In a
survey of consumers, Gardner and Shuman (1987) asked people what they perceived to be
the impact of sponsorship on their purchasing behavior. Fifty-three percent of
respondents said that sponsorship makes them more likely to buy. Eight percent of
respondents said sponsorship makes them less likely to buy, and 39% said it had no
impact.
Another topic which has received some attention in the sponsorship literature is
the concept of ambush marketing (Sandler and Shani, 1989; Meenaghan, 1994). Sandler
and Shani (1989) identify and define ambush marketing as a practice whereby a non
sponsoring brand tries to associate itself with a sponsored event. This is done to either
gain the benefits of being associated with an event without paying for the privilege, or to
limit the impact of a competitors legitimate sponsorship effort.
Their research found that ambushers did achieve greater false awareness of being
sponsors than did non-ambushing, non-sponsoring companies.1 The ambushers, however,
did not approach the awareness levels achieved by the actual sponsors.
Meenaghan (1994) defines various types of ambush. He outlines several ways
legitimate sponsors might prevent negative effects of such actions. He concludes by
presenting the question of whether ambush is moral. The answer to this question, he
proposes, is a matter of both the form the ambush takes and ones perspective.
The final issue addressed in the extant sponsorship literature relates to how sponsorship
actually works. Most of the papers which address this issue rely almost exclusively upon
1 As measured by recall data, the combined recall-recognition measure showed no
difference.


7
larger, more sophisticated companies spend more time and effort choosing what they will
sponsor than do smaller, less sophisticated firms.
The types of evaluation employed by firms to determine sponsorship success has
also been investigated. Gardner and Shuman (1987) found that 17% of firms using
sponsorship engaged in audience research. Among consumer firms, 27% used sales or
share to determine success while only 8% of commercial firms used this measure.
Somewhat startling, however, is that fully 47% of responding firms involved in
sponsorship reported no post-event evaluation.
There is one other paper which reports the results of an empirical investigation of
sponsorship evaluation. In a survey of UK soccer sponsors, Thwaites (1995) found a
heavy reliance upon amount of generated media exposure as a measure of sponsorship
success. He also found that this exposure was more carefully measured by firms which
sponsored large and prominent teams.
Research has also reported on amounts of expenditure on corporate sponsorship.
For example, Meerabeau et al. (1991) reported $5 billion in global expenditure on sports
sponsorship during 1989. This study also details the amount spent by different segments of
the drinks industry, as well as percentages of marketing budgets spent on sponsorship by
each segment. Armstrong (1988) surveyed 17 large international electronics companies.
He found that larger companies, while spending more on sponsorship in absolute numbers,
spent a smaller percentage of their marketing budget on sponsorship than did smaller
companies.
Studies have also reported the impact of sponsorship on consumers. By far, the
most widely reported measure is awareness. Chew (1992), for example, found that


55
distinctive features is judged as less similar to an object with fewer distinct features than is
an object with few distinct features to an object with many. Put another way, a small (less
salient) brand is perceived to be more similar to a big (more salient) brand than is the big
brand to the small one.
Tversky (1977) also claims and demonstrates that at similar levels of salience,
similarity asymmetry is reduced, or nonexistent. An example of this case is a comparison
of the similarity between Hungary and Czechoslovakia. Since neither country represented
a prototype, the perceived similarity of Hungary to Czechoslovakia was about the same
as was the perceived similarity of Czechoslovakia to Hungary. The implications for the
perceived fit between sponsor and event is:
P47: For sponsors and events of equivalent perceived size (salience) there is
little or no fit asymmetry.
Another property Tversky (1977) finds for similarity is that it is not transitive. For
example, while Cuba is similar to Jamaica (both Caribbean islands), and Cuba is similar to
Russia (both Communist countries at the time) Jamaica is not perceived to be similar to
Russia at all. While this seems incredibly intuitive in the context of countries, it may be
less so in the context of the fit between sponsor and event.
For example, it seems entirely possible that a marketing manager might develop
rough heuristics by which to evaluate potential sponsorship opportunities. One possible
strategy might be to consider looking for other sponsors of events which his company
currently sponsors, then examine which other events those companies are sponsoring. In
fact, databases for sale to companies involved in sponsoring are set up to, among other


128
The additional across-studies analysis further supported one effect and uncovered
a new one. Replicating the positive impart of duration of association on making the
inference that the brand is large is important. Finding such confirmation from data
gathered before any mention had been made to the subjects about the size of the brand is
indeed exciting. It not only replicates a previous finding, it demonstrates that attributes of
a sponsorship associations can influence inference making about brand size without ever
mentioning the size of the brand to the subjects. Subjects in the long duration of
association condition made inferences about brand size at more than three times the rate of
those in the short duration-of -association condition.
The new finding from this analysis is also quite interesting. The fart that being a
title sponsor may also have a downside is an intriguing possibility which has implications
for tactical sponsorship decisions. For example, if a brand is attempting to engage
reciprocity through a sponsorship which it hopes will lead to event facilitation inferences,
it may want to be a title sponsor. Being a title sponsor, however, might also lead to more
negative inferences, such as they are just trying to sell me something. When, and for
whom, each of these inferences is more likely is certainly fertile ground for further
research
Overall Discussion and Limitations of Studies la-lc.
The series of three studies presented in this chapter have yielded quite a bit of
information about the impart of event and associational factors on consumer inference
making about sponsoring brands. A summary of findings from these three studies is
presented below in Table 5.


15
celebrity endorsers (Speck, Schumann and Thompson, 1988; Misra and Beatty, 1990;
Frieden, 1984).
The details of how different kinds of similarity can impact sponsorship receive
detailed coverage in chapter five. By adhering to the goal of fitting sponsorship into the
broader category of commercial associations, however, the above findings on the impact
of similarity between providers and acceptors lead to the following general proposals
about sponsorship:
P3: The more a sponsor is like an event, the more positively consumers
evaluate the sponsor.
P4: The more a sponsor is like an event, the more positive affect is
generated about the sponsor through the association.
A third, extremely common finding in the commercial association literature is the
interaction between quality of provider and associate similarity. In general, it has been
found that the fit between associates interacts with the positive relationship between the
quality of the provider and the evaluation of the acceptor. Aaker and Keller (1993) for
example, conclude that high quality brands can be stretched farther and extended into
more dissimilar categories (p.56). Broniarczyk and Alba (1994) agree that, all things
being equal, the best scenario for success involves extension of a highly regarded brand to
a very similar product class. (pg. 215).
With regard to brand extensions, Keller and Aaker (1992) state that when
similarity is very low and the extension seems to be very unrelated to the core brand,
extensions will not be evaluated favorably regardless of the nature of the core brand
associations because those associations will not be seen as relevant to judging the


66
Many authors have espoused the commonsense idea that having overlapping
audiences is a good idea. Varadarajan and Menon (1988) suggest that any cause-related
marketing effort should share the demographics of the target market for the product.
Nicholls, Laskey and Roslow (1992) suggest that understanding the differences in
audience from one Hallmark Event to another may allow marketers to develop optimal
messages. Otker (1988) states that the chance of an effective sponsorship is maximized
when there is a perfect marriage of target groups.. .of sponsor and the sponsored body
(p. 85). Marshall and Cook (1992) found that 49% of respondents listed reaching a
specific target audience as the most important determinant in selecting an event to
sponsor. Thomas (1985) states that the first concern of a company with sponsorship in
mind should be to select the sport, the art or the kind of music which is.. .most appealing
to its consumer/audience (p. 322).
Other authors have written about the ability of sponsorship to reach different
and/or prized audiences. Crowley (1991) points out that sponsorship reaches many
audiences, including current and potential customers, suppliers, employees, the general
public and local community, the business community, and shareholders. He presents the
results of a survey which shows the relative importance of each of these audiences by firm
type, and orientation. Withcer et al. (1991) found that companies used sponsorship of
cultural and art institutions to reach opinion leaders and to achieve community relations
goals, whereas sports sponsorship was used primarily to reach the general public directly.
Wolton (1988) states that arts sponsorship is often employed in order to reach decision
makers, government leaders, and opinion leaders (p. 89).


155
auto insurance company faces additional obstacles to acceptance. This fact may mean that
sponsorship-based inferences may have a different impact on the foreign brand
Kobayama, than on the domestic brand Wilson. This possibility is tested formally
with a logit model that includes brand-specific parameters for both company size and
event facilitation
Results
Once again, a logistic regression was used to test the impact of the independent
variables on choice. To test whether the impact of brand size and event facilitation was
different for the foreign and domestic brands, brand-specific parameters were estimated
for these two variables. This resulted in seven total parameter estimates of interest: the
impact of deductible, price, and service hours on overall choice, as well as the impact of
company size on foreign and domestic brands (two parameters), and the impact of event
facilitation on foreign and domestic brands (two parameters). The parameter estimates for
this model are presented in Table 11 below.


38
P25: Affiliation with an event is positively correlated with perceived
prestige of the event
P26: Affiliation with an event is positively correlated with frequency of
attendance.
P27: Affiliation with an event is negatively correlated with ones
participation in other activities.
In a related area, the manipulation of affiliation motivation has been cited as a
frequently employed advertising tactic (Zinkhan, Hong and Lawson, 1990). These
authors content analyzed a sample of magazine ads from selected years between 1935
and 1985. They found that over the time period studied, affiliation motivational appeals
increased in frequency. The authors contrast this increase with a decline in the use of
achievement appeals.
Implied size
The implied size mechanism also has a message which can be stated in words.
The message of implied size is that this company is big. The cognitive elaboration
required for this mechanism to occur is considerably higher that those previously listed.
A consumer has to, at some level, think to himself this must be a big company to have
this event named after it.
The idea that sponsoring an event can influence the perceived size of the
sponsor has been examined and has received some empirical support. McDonald (1991)
explored this possibility, but did not find evidence of an effect. He states, however, that
the reason sponsorship did not appear to affect perceptions of size was that most of the
companies studied were already clearly perceived to be very big and important (e g.,


134
Design
The simple design consists of a single, two level variable (high/ low brand-event
fit), which is manipulated between subjects. Covariates relating to involvement with and
knowledge about tennis are included for a planned analysis using ANCOVA.
Independent Variable
The manipulated independent variable is brand/event fit. The event chosen was the
US Open tennis tournament. A list of the 19 actual sponsors of the US Open were
pretested for perceived fit with the event. Forty-six subjects rated the fit between these
brands and the US Open (see Table 8). The high fit brand selected was Evian. Along the
same seven point scale used previously, Evian had a mean fit rating of 6.8. The low fit
brand chosen was Texaco, which had a mean fit rating of 2.8. This difference in perceived
fit is significant (F=329, p<0001).
Several audience factors, or attributes of the relationship between the audience and
the event, were measured. The intrinsic personal relevance of the event, and felt
involvement with tennis were measured. Domain knowledge about tennis (both objective
and subjective) was then assessed.
Felt involvement is a point-in-time measure of a consumers involvement with an
event. It includes both enduring and situational components, and represents an
instantaneous measure of the level of meaning of an event in a persons life. It was
measured immediately after exposure to the poster using two seven point scales anchored
by strongly agree and strongly disagree Item one was The message I just saw was


106
Subjects were then instructed to turn to the next page which instructed them as to
how to complete the unaided thought listing task. The instructions read:
On the following page, we would like you to give us your thoughts about the
poster you just saw, by writing one thought in each box. Please list in these boxes
the thoughts that went through your mind as you looked at the poster. Please list
only one thought in a box. You might have positive, negative or neutral thoughts.
ALL ARE FINE.
You may ignore spelling, grammar and punctuation, and you may use phrases.
Simply write down the first thought that comes to your mind in the first box, the
second thought in the second box and so on. Dont worry if you dont fill in every
box. Just write down as many thoughts, feelings and reactions as occur to you.
Subjects then turned to the following page which contained seven thought listing
boxes. Each box was 1.8 high by 13 cm. long. All boxes were centered horizontally, and
placed one below the other on the page (see Appendix D). Subjects were given 90
seconds to list their thoughts. At the end of 90 seconds, they were instructed to finish
working on your current thought, but do not go on to an additional box. When all
subjects had recorded their last thought, all were instructed to turn to the following page,
which contained the instructions for the brand cued thought listing task.
The instructions for the brand cued thought listing task read :
Now, on the following page, we would like you to give us any thoughts you may
have had about the sponsor ROMBOUTS COFFEE (or EXTRAN SPORTS
DRINK). As before, please list in these boxes any thoughts that went through
your mind about ROMBOUTS COFFEE (or EXTRAN SPORTS DRINK)as you
looked at the poster. Please list only one thought in a box. You might have
positive, negative or neutral thoughts. ALL ARE FINE.
Again, you may ignore spelling, grammar and punctuation, and you may use
phrases. Simply write down the first thought that comes to your mind in the first
box, the second thought on the second box and so on. Dont worry if you dont
fill in every box. Just write down as many thoughts, feelings and reactions as
occur to you.


172
Have you ever been to Europe? Y N
Have you ever been to Belgium? Y N
Have you ever heard of Extran
Sports Drink (Rombouts Coffee)
Before today? Y N
Have you ever run in a marathon
(published a poem)? Y N
Have you ever run in a 1 OK
Race(gone to a poetry reading)? Y N


110
H4, which predicted no moderating effect of brand/event fit was not fully
supported. The moderating role of fit on the impact of duration on perceived brand share
(as measured by the interaction of fit and duration) was non-significant (F<1). This is as
predicted. It appears, however, that fit may have some moderating role in the impact of
event size on perceived market share. While the interaction of fit with event size did not
reach traditional levels of significance, it was close (F=3.52, p<07, 2=.02). While this is
not strong support for a moderating role of fit, neither is it strong support of H4 which
predicted no interaction. The effect size associated with this interaction falls within what
Cohen (1977) refers to as a small effect. The power of this experiment to detect an effect
of this size is only .37. In order to obtain a power of .80 to detect such an effect, the
sample size for this experiment would need to more than double from 89 to just over 200.
To further explore this interaction, the data was split into a high fit and a low fit
group. Event size was then regressed on perceived share for each of the two fit groups.
In the high fit group, event size had a significant impact on perceived market share
(F=8.13, p<01). In the low fit group, event size had no measurable impact on perceived
brand share. Caution is necessary in interpreting these results, however, remembering that
the original interaction term was not significant.
Brand Legitimacy
An attempt was made to measure subjects perception of the legitimacy of the
sponsoring brand using two seven point scales. The first asked How legitimate of a
sports drink (coffee) do you perceive Extran (Rombouts) to be?. The scale was
anchored by A fly-by-night brand and a legitimate brand. The second asked How


CHAPTER NINE
THE IMPACT OF AUDIENCE FACTORS AND FIT ON BRAND INFERENCES
To explore only the effects of the event and the association on consumer inference
making would ignore one crucial term in the equation: the consumer. This chapter looks
at how audience involvement with an event impacts the kinds of inferences that are drawn
from an association. Specifically, it examines how the intrinsic personal relevance of the
event, felt involvement with the event, and event domain knowledge impact inference
making about a sponsor of the event.
The first set of studies (la, lb and lc) explored inference making about unfamiliar
brands associated with unfamiliar events. While experimental realism (i.e. task
involvement) was high, subject involvement with the events and the brands was quite low.
The brands used were unfamiliar to the subjects and the events used were fictitious. For
these reasons, it would be quite difficult to examine the role of audience involvement with
the event using one of the events from the previous studies.
When studying the impact of audience involvement on inference making, two
possible approaches are available. One option is to attempt to manipulate involvement
with a real or an imaginary event. The other option is to select a real event, and measure
subjects extant involvement and knowledge. Because exploring how sponsorship works
is a largely substantive endeavor, external validity is a key consideration. For this reason
131


58
P50: The greater the attribute fit, the greater the recall for the association.
P51: The greater the attribute fit, the greater the recognition for the
association.
P52: The greater the attribute fit, the greater the recall for the event when
the brand is cued.
P53: The greater the attribute fit, the greater the recall for the brand when
the event is cued.
P54: The greater the attribute fit, the greater the recognition for the event
when the brand is cued.
PS5: The greater the attribute fit, the greater the recognition for the brand
when the event is cued.2
Murphy and Medin (1985) propose another interesting consequence of attribute
fit (or conceptual coherence). They propose that such fit can facilitate inference-making
about the relationship. Specifically, they present the possibility that coherent or useful
categories are the ones that allow the most inferences to be made. They continue by
proposing that a vague category... enables few if any inferences to be made( p. 293).
While they do not present conclusive evidence to support these assertions, it is certainly
possible to imagine how strong attribute fit might facilitate the kind of inferences
necessary for the functioning of mechanisms such as implied size and implied
endorsement. Therefore, it is tentatively proposed that
P56: Strength of attribute fit is positively correlated with the number of
inferences made about the association.
2 It should be noted that the linear form assumed is only proposed, and that other research
indicates other possible functional forms. The functional form of the relationship between
attribute fit and memory, therefore, is an empirical question.


24
measure or manipulate magnitude, accessibility and relevance of associations between
brand and event, and use these variables to predict impact on brand equity in a variety of
situations. Specific possibilities are explored in later chapters.
Another critical marketing goal upon which sponsorship may have an impact is
positioning. One anecdotal account of a company using sponsorship to reposition their
brand is Tanqueray gin. Tanqueray decided to sponsor the California AIDS Ride, a
bicycling event which raises money for AIDS (New York Times, July 24,1995).
Tanqueray was in the process of trying to reposition the brand away from being a classic
martini component for the 50+ crowd, toward being a more versatile beverage for the 25-
40 crowd. They saw sponsorship of the AIDS Ride as a bold way to associate themselves
with an issue of great importance to their new target market.
Several authors have described brand positioning mechanisms which may be
susceptible to the influence of sponsorship associations. Blackston (1995), for example,
conceives of repositioning as a change in brand meaning. He asserts that because value
depends on the meaning, changing brand meaning is equivalent to changing the value of
the brand. He proposes a brand relationship model where the objective brand is
distinguished from the subjective brand. He defines the objective brand as consisting of
a set of associations, images and personality characteristics. Within this framework,
changing the set of associations should change the meaning of the brand and, therefore,
change its positioning. In cases where associating a brand with an event leads to a change
in the mix of brand associates, sponsorship should be able to aid in the repositioning of a
brand.


59
FS7: Strength of association is positively correlated with the confidence in
ones inferences about an association.
Cohen and Basu (1987) reach some similar conclusions about categorization and
inferences. They elaborate on the potential of affect to serve as an intermediate variable in
the inference-making process. They propose that since categorization produces a
reduction in uncertainty, positive affect may result from a successful fit and negative affect
from an inability to categorize an item particularly if the resulting judgments or
inferences are important (p. 470). They go on to say that by relating the assignment of
affect to more fundamental categorization processes we are also more likely to anticipate
the type of inferences that consumers use to fill in information gaps when direct product
knowledge is absent or when motivation to seek such information is low (Cohen and
Basu 1987, p. 471).
These observations are relevant to sponsorship because they point out that the
affect generated by successfully fitting things into a category may moderate the number
and type of inferences generated. It can therefore be proposed that
P58: The affect generated by the association of a brand with an event
moderates the relationship between attribute fit and inference making
proposed in P56 and P57.
Ortony (1979) also provides some insight into the consequences of attribute fit.
This work attempts to go beyond literal similarity to include what he refers to as
metaphorical similarity. He defines many properties of metaphors, and relates these
properties to Tverskys (1977) concept of similarity. Key to this discussion is the idea that


APPENDIX G
SAMPLE CHOICE PROFILE USED IN EXPERIMENT 3C
Kobayama Auto
Insurance
Wilson Auto
Insurance
Customer service
Hours
8AM-6PM
Mon-Sat
24 Hours 7
days/week
Company Size
Large
Company
Small
Company
Company involvement
with favorite local events
Not involved
Helps make
them possible
Deductible
$1000
$500
Price for 6 months
coverage
$570.00
$615.00
U
u
Please indicate your
choice (check only one)


182


47
added). They then qualify these statements by explaining that this is not news to brand
managers, and that it misses the obvious differences in the brand specific associations
which are not common to all brands within a category. While they go into some detail on
why brand specific associations are an important consideration, they do not develop a
method for determining the brand specific associations for each brand, and assessing their
strength or importance.
Rather, they use pretests to identify brands with at least one strong, unique
association. They then test to see whether these brands are more successful at entering
product classes for which their unique associations are relevant than are brands which do
not possess these unique associations. While the authors make a significant point about
the unique associations possessed by brands (as opposed to categories), their method of
seeking out brands with strong, unique associates still does not answer the question what
is fit?. It does, however, suggest that if both the provider and the acceptor posses the
same strong and unique association, fit will be relatively high.
While specific antecedents of different kinds of fit are discussed below, the
implications for sponsorship of the Broniarczyk and Alba (1994) study can be formally
stated as:
P3S: Perception of generic fit by consumers will be high when the event
and brand share unique, salient attributes.
Within the context of composite branding alliances, Park, Jun and Shocker
(1996) propose structural determinants of fit between two brands. Specifically, they
propose that when the header brand (first brand in composite name) and the modifier
brand (the brand which follows by in the name) have complementary attributes, and are


49
detailed consideration of its potential antecedents, properties and consequences. There
are few details on how one might measure it. The rest of this chapter represents a first step
toward remedying this situation.
There are at least three kinds of (or approaches to) fit. While the three are likely
to be highly correlated in many situations, there are reasons to believe that they have
different impacts on the seven mechanisms discussed in the previous chapter. The three
types are attribute fit, audience fit, and positioning fit. This chapter first describes
attribute fit in terms of various psychological constructs, such as similarity (Tversky, 1977;
Tversky Gati, 1982) and categorization (Mervis and Rosh, 1982; Rosch, 1976) which give
insight into possible antecedents, properties and consequences of attribute fit, as well as
suggestions for its measurement. Audience fit and positioning fit are then presented.
Issues relevant to their special impact on several of the seven mechanisms are discussed,
as are issues related to their measurement. Specific predictions on how these types of fit
impact the seven mechanisms, however, are left for a future chapter.
Attribute Fit
Although similarity is recognized as fundamental to marketing and marketing
strategy, little attention has been paid to the way similarity judgments are
produced Johnson (1986, p. 59)
Attribute fit is the most general type of fit. It is not specific to sponsorship in any
way. All of the antecedents, properties and consequences are, in theory, as applicable to
any other form of commercial association as they are to sponsorship. Attribute fit is
defined as the degree to which two associates share common attributes, and do not


AN INFERENCE-BASED MODEL OF BUILDING BRAND EQUITY THROUGH
SPONSORSHIP
By
JOHN W. PRACEJUS
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA

Copyright 1998
by
John W. Pracejus

ACKNOWLEDGEMENTS
I would first like to thank my chair, Rich Lutz, for guidance, support, and for
allowing me to find my own way. I would also like to thank Chris Janiszewski for
helping me to think more conceptually about the topic, Bart Weitz for helping me to
focus on my strengths, and Mike Weigold for helpful comments and suggestions. Special
thanks go to Joffre Swait for providing me with a new perspective on this project and
research in general, and for tremendous help with development and analysis of the choice
studies.
Others to whom I am grateful for contributing to my intellectual development
include .Alan Sawyer, Joel Cohen, Joe Alba, John Lynch, and especially Tom OGuinn,
without whom I would never have started this process.
I would also like to acknowledge all of the doctoral students who have made my
life at Florida the wonderful experience it has been. Many thanks go to Luk Warlop,
Michel Pham, Susan Fournier, and Prasad Niak for providing maps which allowed me to
successfully follow the path, and to Norma Mednoza, Andre Menck, and Stijn Van
Osselaer, for being so supportive while walking that path with me. Thanks also go to
Corinne Faure, Francis Hollman Anne Stringfellow, Stacy Wood, Americus Reed, Lisa
Bolton, Kevin Bradford, Tom Meyvis and Velitchka Kaltcheva for ideas, talk and coffee.
Finally, I am thankful to my parents, Walter and Rosemary, without whom none
of this would have been possible.
iii

TABLE OF CONTENTS
page
ACKNOWLEDGEMENTS iii
ABSTRACT vii
INTRODUCTION 1
CHAPTER ONE EXTANT SPONSORSHIP RESEARCH:
WHERE ARE WE NOW? 4
CHAPTER TWO SPONSORSHIP WITHIN A BROADER
CATEGORY OF COMMERCIAL ASSOCIATIONS 12
CHAPTER THREE THE POTENTIAL IMPACT OF
SPONSORSHIP ON BRAND EQUITY AND POSITIONING 21
CHAPTER FOUR SEVEN MECHANISMS THROUGH WHICH
SPONSORSHIP MAY FUNCTION 27
Simple Awareness 28
Affect Transfer 30
Image Transfer 32
Affiliation 36
Implied Size 38
Implied Endorsement 41
Reciprocity 43
Conclusion 44
CHAPTER FIVE DIMENSIONS OF FIT 45
Attribute Fit 49
Audience Fit 65
Positioning Fit 70
CHAPTER SIX SACRILEGE 74
CHAPTER SEVEN ASSOCIATION-ENHANCING
COMMUNICATIONS (AEC 80
iv

CHAPTER EIGHT THE IMPACT OF EVENT AND
ASSOCIATIONAL FACTORS ON INFERENCES
ABOUT THE BRAND 86
Conceptual Overview 89
Hypotheses About Specific Inferences 95
Study la 99
Study lb 115
Study lc 122
Additional Cross Study 127
Overall Discussion and Limitations of Studies la-lc 128
CHAPTER NINE THE IMPACT OF AUDIENCE FACTORS
AND FIT ON BRAND INFERENCES 131
Study 2a 133
Study 139
Overall Discussion 141
CHAPTER TEN THE IMPACT OF SPONSORSHIP-BASED
INFERENCES ON BRAND CHOICE 144
Study 3a 145
Study 3b 149
Study 3c 152
Overall Discussion of Studies 3a, 3b, and 3c 157
CHAPTER ELEVEN CONCLUSION, IMPLICATIONS AND
FUTURE RESEARCH 159
Implications 162
Future Research 165
Conclusion 169
APPENDDC A OPEN ENDED MEASURES AND MANIPILATION
CHECKS FOR EXPERIMENT 1A, IB, AND 1C 170
APPENDDC B OPEN ENDED INFERENCE MEASURES AND
MANIPULATION CHECKS FOR EXPERIMENT 2A AND 2B 173
APPENDIX C INVOLVEMENT AND KNOWLEDGE MEASURES
USED IN EXPERIMENT 2A 175
APPENBDC D OPEN ENDED THOUGHT LISTING BOXES 179
APPENDIX E SAMPLE CHOICE PROFILE USED IN
EXPERIMENT 3 A 180

APPENDIX F SAMPLE CHOICE PROFILE USED IN
EXPERIMENT 3B 181
APPENDIX G SAMPLE CHOICE PROFILE USED IN
EXPERIMENT 3C 182
APPENDIX H SAMPLE EXPOSURE STIMULI FOR
EXPERIMENTS 1A, IB, 1C, 2A, AND 2B 183
APPENDIX I ONE POSSIBLE FORM OF A SPONSORSHIP
VALUATION FUNCTION FOR A SPECIFIC BRAND 192
REFERENCES 193
BIOGRAPHICAL SKETCH 204
vi

Abstract of Dissertation Presented to the Graduate School
Of the University of Florida in Partial Fulfillment of the
Requirement for the Degree of Doctor of Philosophy
AN INFERENCE-BASED MODEL OF BUILDING BRAND
EQUITY THROUGH SPONSORSHIP
By
John W. Pracejus
August 1998
Chairperson: Richard J. Lutz
Major Department: Marketing
Sponsorship, or the intentional association of a brand or company with a sports
teams or sporting event, entertainment tour or attraction, festival, fair or annual event,
charitable cause, the arts, or cultural institutions, is emerging as an important component
of the marketing mix. While there is general agreement that sponsorship is effective at
exposing consumers to a brand name, little is known about how sponsorship can build
brand equity.
A two stage model of building brand equity through sponsorship is proposed. It is
proposed that in the first stage, event, associational and audience factors lead consumers
to draw different inferences about the sponsoring brand. These inferences include brand
size (i.e. this must be a big brand to be able to sponsor this event), brand legitimacy
(i.e. this brand must be OK in order to be allowed to sponsor this event), and event
facilitation (i.e. this brand is making this event possible, or at least making it better). In
the second stage, these inferences are proposed to differentially impact choice
probabilities for different categories and brands.
vii

A series of five studies investigates the role played by event, associational and
audience factors on inference making about a sponsoring brand. Evidence is found that
brand size inferences are best created through sponsoring large events for a long time,
whereas event facilitation inferences are best created through being the title sponsor of
smaller events. Three additional studies demonstrate that these two inferences (brand
size and event facilitation) have a different impact on the choice probabilities of different
categories and brands.
viii

INTRODUCTION
As a promotional tool, sponsorship is unique. It is, perhaps, the only means of
communicating with customers without the benefit of a verbal message. Any message
received must be derived from the association in the mind of a consumer. While
traditional advertising often employs simple visual associations between objects and
products, the headline, body copy and/or audio usually help to explain what the
association means. With sponsorship, however, the association is the entire message. For
this reason, the study of sponsorship involves the development and testing of constructs
which ar e quite different from those employed in the study of traditional advertising.
Before going any further, it is necessary to define what is meant by sponsorship.
Sponsorship is a commercial association of a brand with a sports teams or sporting event,
entertainment tour or attraction, festival, fair or annual event, charitable cause, the arts, or
cultural institutions (Andrews 1996). While purchasing advertising time on TV and radio
has sometimes been referred to as sponsoring, buying space in traditional media (i.e.
TV, radio, print) is not what is meant by sponsorship here.
Spending on sponsorship (as defined above) has grown more than fivefold in the
last ten years, and is expected to exceed $5 billion this year in North America alone
(Andrews, 1996). Currently, the bulk of this sum (66%) is spent on sports sponsorship,
but arts and cultural sponsorships are growing at the fastest rate. Overall, spending on
1

2
sponsorship is increasing faster than spending on advertising or sales promotion
(Andrews 1996).
While the bulk of the growth in sponsorship activities has taken place over the last
twenty years, the idea of creating an exploitable association between a commercial and a
noncommercial entity is not a new one. In fact, an association between the product Bovril
and the Nottingham Forest Football Club dates from 1898 (Marshall and Cook, 1992). In
1927, Charles Lindbergh named his plane the Spirit of St. Louis because his financial
sponsors wanted to promote St. Louis, Missouri, as an aviation city. Thus, the first
transatlantic flight was, in fact, a sponsored event (Hildreth and Nalty, 1969; Lindbergh,
1953). When Jesse Owens won four gold medals in the 1936 Olympics, he did so in
Adidas shoes (Ryssel and Stamminger,1988, p. 111). Finally, in the first known example
of a sporting event being named after a brand, The Whitbread (Ale) Gold Cup, a horse
racing event, first took place in 1956 (Marshall and Cook, 1992).
][n spite of this long history as a phenomenon and rapidly growing levels of
expenditure, the mechanisms through which sponsorship affects brand equity, preference
and choice, are not well understood. The purpose of this dissertation is to begin
systematically exploring the ways in which sponsorship works. It begins with a review of
extant sponsorship research and is followed by a summary of relevant research from
related commercial association domains, like celebrity endorsement and brand and line
extension. Measurable marketing variables upon which sponsorship may have an impact
are then discussed.
Exploratory research and managerial intuition as to how sponsorship works are
then developed into seven testable mechanisms through which sponsorship may function in

3
various situations. These mechanisms are simple awareness, affect transfer, image
transfer, affiliation, implied size, implied endorsement, and reciprocity. Constructs which
are hypothesized to have an impact upon the functioning of these mechanisms are then
developed. These constructs include various types of fit between sponsor and event, as
well as sacrilege or consumer outrage at an association which is perceived to be
inappropriate. A theoretical discussion of how other communications can be deployed to
maximize the impact of the sponsorship association wraps up the background and theory
chapters.
Three empirical studies of how the event and associational factors (event size,
brand-event fit, duration of association and share of presence) of a sponsorship impact
consumer perceptions of unfamiliar brands are presented in chapter eight. Chapter nine
reports the results of two studies which investigate the role of audience characteristics
(involvement with and knowledge about the event) on consumer response to a pair of
familiar brands. In chapter ten, the impact of sponsorship-based inference on brand choice
is investigated for several categories of brands. Chapter 11 provides an overview of the
findings of chapters eight through ten and suggests some interesting areas for further study
in this emerging area.

CHAPTER ONE
EXTANT SPONSORSHIP RESEARCH: WHERE ARE WE NOW?
Since 1984, there have been at least 35 journal articles which deal, either directly
or indirectly, with sponsorship. Hundreds more can be found in the trade press. Despite
these facts, very little work beyond describing the occurrence frequency of sponsorship
activities has taken place. What has gone beyond description of frequency often involves
the reporting of levels of awareness, attitude or purchase intention following a
sponsorship. A few tangential topics have also received some attention. Finally, there
have been a good number of papers which put forth some ideas of how sponsorship
works. For the most part, though, these papers rely entirely upon anecdotal evidence for
support of their assertions. These assertions, which come largely from papers written by
practitioners and consultants, however, do provide a wealth of insight into potential
mechanisms through which sponsorship may work. The systematic development of such
mechanisms follows in a subsequent chapter.
One of the questions most frequently addressed in the sponsorship literature is
why do companies engage in sponsorship? Abratt, Clayton, and Pitt (1987) found that
among 60 sponsoring organizations in South Africa, 11% of respondents claimed that
their sponsorship was altruistic to a large extent, 53% said it was altruistic to some
4

5
extent and 36% said it was not altruistic (p. 309). Apart from this one study, however,
the issue of altruistic motives in sponsorship is unexplored.
Armstrong (1988) presents the results of a case study analysis of the sponsorship
activities of 17 large international electronics companies. Interviews were conducted in
four countries (United States, Great Britain, Germany, and France) with corporate and
regional marketing managers from the companies, representatives of the sports being
sponsored and any intermediary agencies involved (80 interviews total). Within the
sample, American firms more frequently listed business managers as the primary target for
the sponsorship, while European and Japanese firms tended to use sponsorship to reach
the consumer audience. Interestingly, the demographics of these two target audiences
were nearly identical (male, upper management, age 35-55). The interviews also revealed
that firms which have been involved with a sponsorship for several years tended to have
image enhancement (as opposed to awareness) goals. It is unclear from the data presented
in the paper, however, whether companies changed their focus over time, or if companies
with awareness goals stopped engaging in sponsorship after awareness goals were met.
Marshall and Cook (1992) asked why 200 British companies from the Times 1000
engaged in sponsorship. Reasons given were reach target audience (49%), product
linked (34%), image of sport (14%) and other(3%) (p. 319). They also asked for
importance ratings for a number of potential sponsorship goals. Mean importance ratings
for building corporate image, national link with product or service, potential
advertising exposure, potential TV coverage, potential press coverage, and covers
target audience indicated that all were important or very important. Only ability to

6
provide corporate hospitality was rated below the midpoint of the five point importance
scale.
Withcer et al. (1991) found that companies used sponsorship of cultural and art
institutions to reach opinion leaders and to achieve community relations goals, whereas
sports sponsorship was used primarily to reach the general public directly. Scott and
Suchard (1992) factor analyzed the results of a 23 question survey of 108 Australian
businesses. The resultant four factors were regressed on the amount spent on sponsorship
by each company. Two of the four factors, performance and client relationships, were
found to be significantly correlated to expenditure on sponsorship. The significant
correlation of client relationship and expenditure is said to be somewhat counter to
previous studies (i.e. Marshall and Cook, 1992) which found hospitality to be only a
marginally important goal.
Another issue which has received some attention in the sponsorship literature is
how do companies go about choosing an event, once they have decided to engage in
sponsorship? Gardner and Shuman (1987) surveyed Fortune 500 companies about their
choice of sponsorship activities. Eighty-four percent reported using in-house expertise for
screening, while 24% percent utilized outside counsel to evaluate and choose particular
events.
In a survey of companies sponsoring football (soccer) in the UK, Thwaites (1995)
found that sponsors of Premier League Clubs, which have high profile and cost, were
more proactive in securing sponsorship opportunities, compared more clubs before the
decision and negotiated longer term contracts, than did sponsors of Division One Leagues,
which have moderate profile and cost (all significant at p<05). This seems to suggest that

7
larger, more sophisticated companies spend more time and effort choosing what they will
sponsor than do smaller, less sophisticated firms.
The types of evaluation employed by firms to determine sponsorship success has
also been investigated. Gardner and Shuman (1987) found that 17% of firms using
sponsorship engaged in audience research. Among consumer firms, 27% used sales or
share to determine success while only 8% of commercial firms used this measure.
Somewhat startling, however, is that fully 47% of responding firms involved in
sponsorship reported no post-event evaluation.
There is one other paper which reports the results of an empirical investigation of
sponsorship evaluation. In a survey of UK soccer sponsors, Thwaites (1995) found a
heavy reliance upon amount of generated media exposure as a measure of sponsorship
success. He also found that this exposure was more carefully measured by firms which
sponsored large and prominent teams.
Research has also reported on amounts of expenditure on corporate sponsorship.
For example, Meerabeau et al. (1991) reported $5 billion in global expenditure on sports
sponsorship during 1989. This study also details the amount spent by different segments of
the drinks industry, as well as percentages of marketing budgets spent on sponsorship by
each segment. Armstrong (1988) surveyed 17 large international electronics companies.
He found that larger companies, while spending more on sponsorship in absolute numbers,
spent a smaller percentage of their marketing budget on sponsorship than did smaller
companies.
Studies have also reported the impact of sponsorship on consumers. By far, the
most widely reported measure is awareness. Chew (1992), for example, found that

s
viewers of a sponsored public television program (MacNeil/Lehrer News Hour) were
more aware that AT&T sponsored the News Hour than were nonviewers.
Crimmins and Horn (1996) point out the need to take into account the number of
people who mistakenly believe that a competitor is a sponsor, in addition to actual
awareness that your brand is the sponsor of an event. They point out that while 35% of a
consumer mail panel could correctly identify Coke as the official soft drink of the NFL,
that 34% thought it was Pepsi, and note that Coke had paid $250 million to become the
official soft drink.
Awareness levels of sponsorship have also been studied in an experimental setting.
Hoek, Gendall and Stockdale (1993) exposed New Zealand secondary school students to
tapes containing a cricket match with a cigarette brand as the title sponsor (the 1990
Rothmans Cricket Series) and control subjects to a match without cigarette sponsorship.
The only significant effect found was that treatment subjects subsequently had a higher
awareness for Rothmans than did a control group. No significant impact on attitude
towards smoking was found.
Some papers have focused on awareness levels that a company is a sponsor.
Wright (1988), for example reports awareness levels for sponsors of different football
(soccer) teams in Great Britain. Time series data are provided which demonstrates the
various awareness levels over time.
Rajaretnam (1994) reports the findings of a unique experiment in which a large
Indian tire maker stopped all product message advertising and spent almost its entire
marketing budget on sponsorship from 1984-1987. While the author does not consider in
detail other potential factors impacting awareness and image variables (e.g., increased

9
distribution and word of mouth), the numbers are impressive. During the period of the
experiment, top of mind awareness went from 4% to 21%, which is a 425% increase.
Unaided recall went from 38% to 68%, and aided awareness went from 92% to 99% (p.
65). The author notes that while the impact of sponsorship on awareness was almost
immediate, the impact on brand preference took a longer time.
One paper attempted to address determinants of sponsorship awareness. Hansen
and Scotwin (1995) reported that in an experimental setting, those interested in soccer
(measured) had higher awareness of the team sponsors than did those who were not
interested. The authors, however, did not report whether these differences are statistically
significant.
Other impacts of sponsorship have also been reported such as beliefs about the
company. Javalgi et al. (1994) studied perceptions of five local companies and awareness
of their sponsorship activities. They hypothesized that persons who are aware of corporate
sponsorship have a more favorable view of the sponsoring company than persons who are
unaware. This was not supported. Awareness of the sponsorship activity was only related
to the statements is involved in the community and only wants to make money
(reversed), and neither was significant overall. Is involved in the community was
significantly correlated with awareness of sponsorship for only one of the five companies.
In a study which did not directly involve sponsorship, Nebenzahl and Jaffe (1991)
found that evaluation of Korean products by Israeli consumers was higher after the 1988
Olympics in Seoul than it was before the games among heavy Olympic viewers, but not
among a low viewing control group. This suggested to them that an association with
the Olympic games had raised peoples evaluations of Korea and in turn Korean products.

10
Another impact of sponsorship reported in the literature is purchase intention. In a
survey of consumers, Gardner and Shuman (1987) asked people what they perceived to be
the impact of sponsorship on their purchasing behavior. Fifty-three percent of
respondents said that sponsorship makes them more likely to buy. Eight percent of
respondents said sponsorship makes them less likely to buy, and 39% said it had no
impact.
Another topic which has received some attention in the sponsorship literature is
the concept of ambush marketing (Sandler and Shani, 1989; Meenaghan, 1994). Sandler
and Shani (1989) identify and define ambush marketing as a practice whereby a non
sponsoring brand tries to associate itself with a sponsored event. This is done to either
gain the benefits of being associated with an event without paying for the privilege, or to
limit the impact of a competitors legitimate sponsorship effort.
Their research found that ambushers did achieve greater false awareness of being
sponsors than did non-ambushing, non-sponsoring companies.1 The ambushers, however,
did not approach the awareness levels achieved by the actual sponsors.
Meenaghan (1994) defines various types of ambush. He outlines several ways
legitimate sponsors might prevent negative effects of such actions. He concludes by
presenting the question of whether ambush is moral. The answer to this question, he
proposes, is a matter of both the form the ambush takes and ones perspective.
The final issue addressed in the extant sponsorship literature relates to how sponsorship
actually works. Most of the papers which address this issue rely almost exclusively upon
1 As measured by recall data, the combined recall-recognition measure showed no
difference.

11
anecdotal evidence drawn from colorful examples (Hastings, 1984; Meenaghan, 1991b;
Thomas, 1985; Crowley, 1991; Parker, 1991; McDonald, 1991; Otker, 1988; Ryssel and
Stamminger, 1988; Schoch, 1994; Walshe and Wilkinson, 1994). They have titles like
Professional Football SponsorshipProfit or Profligate (Thwaites 1995) and Arts
sponsorship: Harmony or discord? (Wolton 1988). While there is little formal conceptual
development, construct definition or theory, they do contain a good deal of intuition about
the processes involved. These papers are drawn upon heavily for examples of constructs
proposed in the following chapters

CHAPTER TWO
SPONSORSHIP WITHIN A BROADER CATEGORY OF COMMERCIAL
ASSOCIATIONS
As alluded to in the introduction, sponsorship is not unique as a commercial
endeavor which seeks to associate two things. The core idea behind brand extension, for
example, is that a new brand can sometimes benefit from an association with an existing
brand. Likewise for celebrity endorsers, an underlying assumption is that the brand can
sometimes benefit from being associated with a human being. Other areas of study in
marketing which involve commercial associations include co-branding (OConnor, 1996;
Carpenter, 1994), the use of one brand as an ingredient in another (Park, Jun and
Shocker, 1996), brand alliances (Rao and Ruekert, 1994), and retail brand assortment
(Laforet and Saunders, 1994; Bawa, Landwehr and Krishna, 1989; Bultez et al., 1989).
There is even anecdotal evidence of the importance of customers as commercial associates
in terms of providing prestige (Smith and Park, 1992).
The two areas which have received the most attention in terms of understanding
the impact of such associations on consumers are the brand extension and celebrity
endorsement literatures. While brand extensions and celebrity endorsers share with
sponsorship issues related to association, they also differ in some conceptually important
ways. The goal of this chapter is to fit sponsorship into a larger framework of commercial
associations. Due to the different assumptions about the underlying mechanisms of brand
12

13
extension and celebrity endorsers, the issues addressed in this chapter are limited to the
very basic, common denominator issues and predictions.
Since the goal is to set up a broad understanding of commonalties among
commercial associations, the level of precision in this chapter is predictably broad. The
propositions put forth here, therefore, are also predictably broad. The specific details of
how sponsorship is likely to function, as well as detailed predictions about main effects
and moderators are left for later chapters.
In order to facilitate discussion of broad issues relating to commercial
associations, it is necessary to introduce a bit of vocabulary. In the discussion below,
associate can refer to either member of an association. For example, a brand can be an
associate of a celebrity, and a celebrity can be an associate of a brand. An associate of
simply means that an object is associated with something else. Acceptor, however,
refers to the object which an association is attempting to help. In celebrity endorsement,
the acceptor is usually the brand being associated with the celebrity. In a brand extension,
the acceptor is the new brand which is being associated with the old one. The provider,
on the other hand, is the object or person providing the help. In celebrity endorsement,
the celebrity is usually the provider. In a brand extension, the preexisting brand is the
provider.
P erhaps the most basic premise, across domains of commercial associations, is that
the more desirable the provider, the more beneficial the association will be to the acceptor.
Aaker and Keller (1990), for example, find that in the brand extension domain the higher
the quality of the core brand (the provider) the better the consumer acceptance of the
extension. Sunde and Brodie (1993) obtain similar results. Likewise, in the celebrity

14
endorser domain, Kahle and Homer (1985) find that the better (more attractive) the
provider (the endorser) the more the acceptor (product) is liked. This same effect has been
found when the provider is a speaker, and the acceptor is an argument (Horria, Naccari
and Fatoullah, 1974; Chaiken, 1979).
As stated above, there are many different mechanisms through which provider
quality impacts acceptor quality. The specifics of these mechanisms are discussed in
future chapters. In the most general terms, however, it can be proposed:
PI: The more highly regarded the event, the more highly a consumer will
evaluate the sponsor.
P2: The more well liked the event, the more well liked will be the sponsor.
An equally frequent finding in the literature in commercial associations is that the
more qualitatively similar the two associates, the more benefit will come to the acceptor
from the association. Aaker and Keller (1990) find this to be true in the brand extension
domain Beyond simple qualitative similarity, however, they define fit dimensions along
which extended brands (acceptors) may match up with the original branded product
(provider). Specifically, they find that if the acceptor is a complement of, or a substitute
for, the provider, the acceptor will be better received. Similarity between provider and
acceptor, however, can take place at a much higher level of abstraction. Park, Milberg
and Lawson (1991), for example, find that extensions work best when both core brand
and extension are either prestige or functional in nature. Benefits to the acceptor of
qualitative similarity with the provider have also been frequently demonstrated for

15
celebrity endorsers (Speck, Schumann and Thompson, 1988; Misra and Beatty, 1990;
Frieden, 1984).
The details of how different kinds of similarity can impact sponsorship receive
detailed coverage in chapter five. By adhering to the goal of fitting sponsorship into the
broader category of commercial associations, however, the above findings on the impact
of similarity between providers and acceptors lead to the following general proposals
about sponsorship:
P3: The more a sponsor is like an event, the more positively consumers
evaluate the sponsor.
P4: The more a sponsor is like an event, the more positive affect is
generated about the sponsor through the association.
A third, extremely common finding in the commercial association literature is the
interaction between quality of provider and associate similarity. In general, it has been
found that the fit between associates interacts with the positive relationship between the
quality of the provider and the evaluation of the acceptor. Aaker and Keller (1993) for
example, conclude that high quality brands can be stretched farther and extended into
more dissimilar categories (p.56). Broniarczyk and Alba (1994) agree that, all things
being equal, the best scenario for success involves extension of a highly regarded brand to
a very similar product class. (pg. 215).
With regard to brand extensions, Keller and Aaker (1992) state that when
similarity is very low and the extension seems to be very unrelated to the core brand,
extensions will not be evaluated favorably regardless of the nature of the core brand
associations because those associations will not be seen as relevant to judging the

16
extension.. .That is, the perceived quality of the core brand influenced brand extension
evaluation only when there was some basis of fit between the core brand and proposed
extension products (p. 37).
Sheinin and Schmitt (1994) find complex interactions between brand affect,
category breadth, and attribute congruity between the original and the extended brand.
Sunde and Brodie (1993) find that consumer acceptance of a brand extension will be
higher if the original brand is of high quality and there is a perceived fit between core
brand and extension.
McCracken (1989), in the context of celebrity endorsers, points out that the ability
of celebrity endorsers to act as providers for a brand comes from the cultural meaning
with which they are endowed. His argument is that if an endorser does not fit with a
brand, there will be no overlap in cultural meaning and, therefore, no positive impact upon
the acceptor.
Perhaps the best example of the interaction between quality of provider and
associate similarity in the domain of endorsers is found by Friedman and Friedman (1976).
They investigated whether the effectiveness of endorser-type is dependent on the type of
product being endorsed. They find a significant product by endorser-type interaction such
that ads for jewelry were evaluated highest when done by a celebrity, whereas ads for
cookies were better received when done by a homemaker. Thus, the study seems to show
the importance of the fit between endorser and product type.
The common finding of an interaction between provider quality and associate
similarity leads to the following proposal for sponsorship:

17
P5: The impact of perceived quality of an event on evaluation of, and positive
affect toward, a sponsor depends upon the level of fit between sponsor and
event. High fit facilitates the impact of event quality, whereas poor fit
impedes it.
The commercial association literature also provides evidence of many moderators
of the impact a commercial association may have. Examples include previous
associations, arousal, symbolic value, type of good, and priming. Each of these topics is
considered below.
In the case of previous associations, Walker, Langmeyer and Langmeyer (1992)
find that previously unendorsed products (bath towels and VCRs) pick up more of the
personality of the endorser than does a previously endorsed product (jeans). This
suggests the possibility of the following propositions for sponsorship:
1P6: Brands in product classes which have not previously engaged in
sponsorship form stronger associations in the minds of consumers than
brands in categories where sponsorship is common.
F7: Brands which have not previously engaged in sponsorship form stronger
associations with events than brands engaged in multiple sponsorships.
.Another moderator of commercial association effectiveness is arousal.
Sanbonmatsu and Kardes (1988) find that high arousal causes an increase in the
effectiveness of using a celebrity vs. noncelebrity endorser. Pham (1993) replicates and
extends this finding. From these results it is possible to develop the following
propositions:

18
P8: Events which are highly arousing cause higher evaluation of more
positive affect for a sponsored brand than do equally evaluated events which
sire less arousing.
Given the intrinsically more arousing nature of live events over televised ones,
P9: The differential effect proposed in P8 is greater for live spectators than
for people who watch an event on TV.
Another moderating factor of commercial associations is the symbolic value of the
provider. In general, providers with greater symbolic value are more capable of having an
impact on the acceptor through the association. This is the foundation for many of the
ideas put forth by McCracken (1989) with regard to celebrity endorsers. In the context of
brand extensions, Reddy, Holak, and Bhat (1994) propose that brands with high symbolic
value will be more readily extended than others. Using statistical analysis on time series
sales data, they find evidence that highly symbolic brands extended better than less
symbolic brands. This is true using both expert ratings of symbolism and consumer ratings
of symbolic value. From these findings in the general area of commercial associations, it
can be argued that for sponsorship,
P10: The higher the symbolic value of an event for a consumer, the more
impact the association will have upon the evaluation of the sponsoring brand.
The type of good involved as the acceptor has also been shown to impact the
outcome of an association. Smith and Park (1992) find that in general, brand strategy
(use of extensions vs. not) accounts for only 4% of market share, whereas for goods
with mostly experience attributes it accounts for 9% of the variance (p. 309). This
increase in the impact of brand extension upon share for experience goods over all

19
goods indicates that experience goods may be more susceptible to the impact of
commercial associations than are search goods. In the context of brand alliances Rao
and Ruekert (1994) view the search-experience dichotomy as a continuum (p. 95).
They give examples of how brand name ingredients in food products which are likely to
affect taste are better brand alliances than are alliances involving products with mainly
search attributes. These studies seem to indicate that experience goods may be more
easily impacted by commercial associations than are search goods. Therefore it can be
proposed:
PI 1: All else being equal, sponsorship will be more effective for brands with
more experience attributes and fewer search attributes.
P12: Brands which are considered credence goods will be more greatly
impacted by sponsorship than will search or experience goods.
One final issue needs to be touched upon in this review of the commercial
association literature. That is the issue of the acceptor having a negative impact upon
the provider. Aaker (1990) notes that risky associations may backfire and gives
examples of several notable cases where a failed brand extension actually harmed the
image of the core product. Dawar and Anderson (1994) propose and find evidence that
an extension to a particular product may make it more difficult for the brand to extend
in a different direction in the future (p. 124). Loken and Roedder John (1993) explain
how a dilution effect (harming the provider) occurs under a bookkeeping model. This
bookkeeping model dilution effect will occur for extensions which have attributes that
are incompatible with the image of the core brand. They note that this can be avoided if

20
the association can be perceived as either an anomaly or a separate category. With
regard to sponsorship, therefore, it can be proposed:
P13: Effects of an event/hrand association can impact the event as well as
the brand.
The goal of this chapter was to provide a conceptual framework for commercial
associations. The similarities in findings from celebrity endorsement, brand extension and
other commercial association literatures were presented. Likely implications of these
common findings for sponsorship were given as formal propositions. Overall, this chapter
has demonstrated the ways in which sponsorship is like other commercial associations.
The unique aspects of sponsorship and their likely impact on brand equity and positioning
are provided in the following chapter

CHAPTER THREE
THE POTENTIAL IMPACT OF SPONSORSHIP ON BRAND EQUITY AND
POSITIONING.
As stated in the introduction, with sponsorship, the association is the message.
There is no argument with which to measure agreement, no claims, attributes or benefits
on which to measure recall. It seems appropriate, therefore, to examine marketing
variables which may be affected by sponsorship and which do not rely entirely upon formal
argument. Two such variables are brand equity and brand positioning. The goal of this
chapter, therefore, is to investigate what sponsorship can do. It does not deal with how it
does it nor when it will work best. These issues are addressed in subsequent chapters.
While the complex associative mechanisms through which sponsorship may impact
brand equity have received little attention to date, the act of associating a brand with a
socially intrusive activity possessed of its own personality in the eyes of the receiving
audience (Meenhaghan, 1991b, p. 8) should certainly be able to impact the things one
associates with a brand. This is relevant because several authors have conceived of brand
equity as consisting largely of brand associations.
For instance, Keller (1993) describes what he calls customer-based brand equity.
He defines this as the differential effect of brand knowledge on consumer response to the
marketing of the brand (p. 1). The keys to this effect, according to Keller, are familiarity
with the brand and holding favorable, strong, and unique brand associations in memory
21

22
He goes on to propose that marketing activity can enhance or maintain the favorability,
strength, and uniqueness of a brand. Given an association-based definition of brand
equity:
F14: Significant sponsorship activities are capable of impacting a brands
equity.
In his book on managing brand equity, Aaker (1991) devotes a good deal of time
to explaining the importance of associations to brand equity. With regard to this
importance, he suggests that the association not only exists but has a level of strength.
A link to a brand will be stronger when it is based on many experiences or exposures
rather than few. It will also be stronger when it is supported by a network of other links
(p. 109). Clearly, then, by Aakers definition as well, sponsorship associations should
sometimes be able to affect brand equity.
It should be noted, however, that Aaker has not endorsed the ability of
sponsorship to impact brand equity through building favorable associations. In fact, he
states that the primary role of most event sponsorship is to create or maintain awareness
(p. 75). Given, therefore, that sponsorship receives but a single line in the book, it does
not seem likely that Aaker had fully considered the ways in which sponsorship might
impact upon brand equity in great detail.
Other definitions of brand equity take a slightly different perspective. Simon and
Sullivan (1993), for example, have proposed a model of brand equity based upon market
valuation of the stock equities of corporations with substantial brand assets. Similarly,
Farquhar (1989) defines brand equity as the value added with which a given brand

23
endows a product(p. 24). He notes that Coopers and Lybrand calculate brand equity by
comparing the premium price paid for a branded product compared to a generic one. It
should be noted that while these definitions focus on the economic value which comes
from brand equity, they do not preclude an associative component in their formation.
In fact Farquhar et al. (1992) propose ways in which the economic value of brands
can be maximized and exploited through management of associations. They develop the
concept of master brands which are very exploitable. They define these master brands
as those brands which own an association. That is, when the associate is mentioned, the
brand comes to mind. While the associates used in most of the examples are product
categories, they acknowledge that other types of brand associations include celebrity
endorsers, sponsored events, geographic affiliations, user groups, etc. (pp. 32-33,
emphasis added).
In addition to specifically including the management of commercial associations as
an important component of brand equity management, they provide insight into methods
of measurement. They suggest that associations can be determined through free
associative tasks, projective methods, and focus groups. They also claim that the process
of building brand equity consists of developing favorable associations, making these
associations accessible, and developing points of difference which can be easily
remembered by consumers.
These ideas have clear implications for sponsorship. In addition to the awareness
building contribution of sponsorship to brand equity, it seems feasible that more complex
associative mechanisms might be able to play a role as well. It should be possible to
develop functional forms for the impact of sponsorship on brand equity. These might

24
measure or manipulate magnitude, accessibility and relevance of associations between
brand and event, and use these variables to predict impact on brand equity in a variety of
situations. Specific possibilities are explored in later chapters.
Another critical marketing goal upon which sponsorship may have an impact is
positioning. One anecdotal account of a company using sponsorship to reposition their
brand is Tanqueray gin. Tanqueray decided to sponsor the California AIDS Ride, a
bicycling event which raises money for AIDS (New York Times, July 24,1995).
Tanqueray was in the process of trying to reposition the brand away from being a classic
martini component for the 50+ crowd, toward being a more versatile beverage for the 25-
40 crowd. They saw sponsorship of the AIDS Ride as a bold way to associate themselves
with an issue of great importance to their new target market.
Several authors have described brand positioning mechanisms which may be
susceptible to the influence of sponsorship associations. Blackston (1995), for example,
conceives of repositioning as a change in brand meaning. He asserts that because value
depends on the meaning, changing brand meaning is equivalent to changing the value of
the brand. He proposes a brand relationship model where the objective brand is
distinguished from the subjective brand. He defines the objective brand as consisting of
a set of associations, images and personality characteristics. Within this framework,
changing the set of associations should change the meaning of the brand and, therefore,
change its positioning. In cases where associating a brand with an event leads to a change
in the mix of brand associates, sponsorship should be able to aid in the repositioning of a
brand.

25
Park, Jaworski, and Maclnnis (1986) have conceptualized positioning as
management of brand image. They present a framework for managing brand image at
different times in a brands life cycle. They define brand image as the understanding
consumers derive from the total set of brand-related activities engaged in by the firm
(p. 135). They stress the importance of various brand associations on brand image, stating
that each element of the (marketing) mix can affect the inferences consumers develop
about the brand (p. 138). Therefore,
P15: Whenever sponsorship of an event causes a change in the associations a
consumer has with the brand, brand image changes, and positioning within
multidimensional space is altered.
Crawford (1985) proposes that the primary function of product positioning is to
differentiate among products. He defines a product positioning typology in which three
things, product features, benefits, and surrogates, can be used for positioning. Of
relevance to sponsorship is the surrogates positioning strategy. He lists coming from
respected source or endorsed by a respected personality as tactical means to employ a
surrogates positioning strategy. While sponsorship is not mentioned per se, it seems
reasonable that an association with a respected event or organization should also aid ones
surrogate positioning strategy. Therefore,
P16: To the extent that a sponsorship represents a unique association, it is
exploitable as a differentiating positioning strategy.
Finally, DiMingo (1988) points out that product positioning is a process of
distinguishing ones company or product from competitors along real dimensions which

26
are important to customers. By his definition, there is no ideal position for all brands.
Positioning must be based upon advantages of the brand. He contends that these
advantages should be communicated through a consistent, integrated communications
strategy involving brand name, packaging, and advertising. While he does not specify a
role for sponsorship, the idea that positioning should be used primarily for differentiation
from competition suggests that to the extent to which a sponsorship represents a point of
differentiation from competitors, it will be an effective positioning tactic.
This can be formally stated as:
P17: Sponsorship will have a greater impact on perceived brand positioning
when there is less competitor sponsorship of similar events.
This chapter has examined the marketing variables which may be affected by
sponsorship. Specific proposals about the impact of sponsorship on brand equity and
positioning were presented. The mechanisms through which sponsorship can effect these
variables, however, was not addressed. This issue is covered in the following chapter.

CHAPTER FOUR
SEVEN MECHANISMS THROUGH WHICH SPONSORSHIP MAY WORK
Almost no conceptual or empirical studies have, thus far, addressed how
sponsorship works. It is the contention of this dissertation that at least seven distinct
mechanisms exist through which sponsorship may function. As proposed in the previous
chapter, these mechanisms may impact brand equity and brand positioning. They may also
impact other relevant marketing variables of interest. These mechanisms are not
hypothesized to be exclusive of one another. Two or more of them may function in any
given sponsorship association. Together, however, they are proposed to represent a
reasonably complete set of the ways sponsorship works.
The simplest way to order these mechanisms is by the amount of cognitive
elaboration which is required by a consumer for the mechanism to function. In
approximate order of necessary cognitive elaboration, they are simple awareness, affect
transfer, image transfer, affiliation, implied size, implied endorsement, and reciprocity.
The rela tionship of the mechanisms depicted in Figure 1 below.
27

28
Increase in required cognitive elaboration
( Simple J f*? ( JT ( affiliation( implied Y^A
\ awareness^ transfer transfer l size V endorsement J J
FIGURE 1
RELATIONSHIP BETWEEN THE SEVEN MECHANISMS
The only of these mechanisms to receive significant direct acknowledgment are
simple awareness, affect transfer and, to a lesser extent, image transfer. Evidence for the
existence of the affiliation, implied size, implied endorsement, and reciprocity mechanisms
is largely intuitive and/or anecdotal. The seven mechanisms are described below.
Propositions related to the main effects of these mechanisms are also provided. Potential
moderators of these mechanisms are provided in later chapters.
Simple Awareness
For who is going to buy a product or service from a company that one has never
heard of? Wright (1988, p. 104)
Awaireness, as a mechanism through which sponsorship is likely to function, is
widely accepted by practitioners and academics alike. A typical explanation of how this
mechanism might function is as follows. A consumer goes to an event or watches it on
TV and is exposed to the brand name. Since repeated exposure to an object has been
found to lead to positive affect toward the object (Zajonc 1969), the consumer feels better
about the brand. This good feeling then has all sorts of benefits to the brand, including,

29
but not limited to, greater attention paid to subsequent commercial communications (or
less counterarguing, see Pracejus, 1993), and greater chance of inclusion in the
consideration set. While there is little doubt that such things take place, to end here with
awareness ignores the rich possibilities involved in the association. In short, if simple
awareness were the only mechanism in operation, then it would, for example, make no
difference whether a brand name was seen on the side of a race car or on the side of a
bam.
As mentioned above, there is little doubt that awareness functions as an important
mechanism in sponsorship. Examples of simple awareness, however, where there is little
or no chance of cognitive elaboration are not easy to cite. One which comes close is the
practice of soft drink bottlers paying for convenience store signage, provided it
prominently displays the appropriate brand logo. This example meets the criterion for
simple awareness because of the low likelihood of elaboration by a store patron as to the
origins of the large Coke logo on the ABC Beverage sign. Even if there were some
attribution on the part of a customer, it is unlikely that he would have enough involvement
with the store to care that the owners received a new sign courtesy of the local Coke
bottler. The most likely impact (if any) is that the customer may have high top of mind
awareness of Coke as he faces his soft drink selection. There is little or no difference
between the logo appearing on the sign or on the side of a nearby building. The only thing
accomplished is simple awareness of the brand name.
Despite the relative simplicity of this mechanism, two basic propositions can be
offered. Given the reasonably well established mere exposure effect,

30
P18: The more frequently a sponsored brand name is encountered by a
consumer in the context of an event, the more positive feelings s/he has
toward the brand.
P19: The effect of P18 is greater for less familiar brands than for more
familiar brands.
Affect Transfer
Affect transfer in a sponsorship context is analogous to affect transfer in an
advertising context, whereby ones attitude toward the ad transfers to ones attitude
toward the brand, with little cognitive mediation (MacKenzie, Lutz and Belch 1986;
MacKenzie and Lutz 1989). As a mechanism through which sponsorship may function,
affect transfer refers to positive feelings toward an event transferring to the sponsoring
brand through the sponsorship association. This mechanism requires no real cognitive
elaboration. It does not even require conscious awareness of the association, although
such an awareness might be beneficial to the transfer taking place.
Affect transfer is a mechanism which has received some mention in the sponsorship
literature, although much of it is rather vague. In a paper which reports significant
differences in demographic and lifestyle variables among several Hallmark events (major
city-wide destination events), Roslow, Nicholls and Laskey (1992) note that, Hallmark
events are attended by diverse audiences who have come together to enjoy a specific
experience or entertainment. This heightened level of expectation may provide a more
beneficial exposure content than traditional media (p. 54). They elaborate that, It is
probable that this form of promotion would benefit the advertiser from increased levels of
excitement and interest associated with Hallmark events (p. 58). These authors seem to

31
be proposing that the positive affective responses to the Hallmark event may transfer to
the sponsoring brand.
Crimmins and Horn (1996) also address the idea of affect transfer. They discuss
the transfer of positive feelings from the event to the brand in terms of Heiders (1946)
elementary human calculus. They posit that, given the relative weakness of many brand
beliefs, the impact of associating a lowly regarded brand with a highly regarded event
should usually lead to an upward estimation of the brand. While this seems plausible, the
authors provide no data to support the assertion. The idea, however, is that the
positiveness of the event can transfer to the sponsor.
Stipp and Schiavone (1996) attempt to demonstrate that attitude toward the
Olympics can actually rub off onto Olympic sponsors. Through multiple regression,
they find that attitude toward Olympic sponsorship in general has a significant impact, net
of ad recall and ad evaluation, upon the sponsors image. While this provides some
support for the affect transfer mechanism, it should be noted that the survey methodology
employed somewhat stacked the deck in favor of their finding.
Respondents were first asked a series of agree-disagree statements like official
Olympic sponsors provide money to: Help athletes train, Make it possible for the
American teams to go to the Olympics,. (p. 24). Setting up the survey with these
questions at the front end certainly had the potential to bias responses toward a strong
association between attitude toward sponsorship and attitude toward a specific sponsor (a
demand effect).
Indication of an affect transfer mechanism is also found in other commercial
association domains. In the context of celebrity endorsers, for example, Misra and Beatty

32
(1990) found that the transfer of affect from spokesperson to brand was facilitated when
the two were matched. In the domain of brand extensions, Rangaswamy, Burke and
Oliva (1993) propose affect transfer through categorization as a method by which
utility of the brand name impacts utility of the product.
Affect-transfer-like mechanisms are also found in the retail assortment domain.
Jacoby and Mazursky (1984), in a factorial design, associate high and low quality brands
with high and low quality stores. They attempt to determine how differential affect (i.e.
positive affect toward one, negative affect for another) in the associate pair interact. Weak
support is found for an averaging process.
Like simple awareness, the cognitively simple mechanism of affect transfer allows
basic propositions to be made:
F20: The more positive the affective response to the an event, the more
positive the affective response to a sponsoring brand.
P21: The stronger the associative linkage between the brand and the event,
the stronger is the impact posited in P20.
Image transfer
The aim of sponsorship is image transfer: the transfer of the personality, aura
and competence... to the brand Ryssel and Stamminger (1988, p. Ill)
While affect transfer is the process of assigning event-based affective response to
the brand, image transfer refers to the assignment of abstract associations of the event to

33
the brand. Image transfer is the mechanism requiring the least cognitive elaboration in
which the association can be formally stated as a message. The message of image transfer
is that, along some dimension, this brand is like this event.
The idea that the image, or personality of an event can impact upon the image or
personality of a brand has been alluded to several times in the context of sponsorship.
Parker (1991), for example, describes the relationship between sponsorship and brand
image. Marshall and Cook (1992) found that 14% of respondents listed image of sport
as the most important determinant in selecting an event to sponsor. Wolton (1988)
contends that art sponsorship is used to alter or modify the image of a company, lending a
human face to cold, technologically-oriented companies, and providing a positive
association with companies such as insurers which are usually associated with unpleasant
situations.
McDonald (1991) describes what he refers to as product relevance (p. 36) which
can be either direct or indirect. Indirect relevance is proposed to imply association at an
abstract, more image-based level. A strong and traditional bank sponsoring the traditional
sport of football (soccer) is given as an example of indirect relevance, where the link is
tradition. Relating back to the discussion above, in this example the message would be
that this bank is like the soccer league in that they are both traditional.
Perhaps the best anecdotal evidence of how the image transfer mechanism might
be used for a specific marketing objective, like repositioning a brand, is given by
Meenaghan (1991a). He states that sponsorships can achieve specific objectives through
what he calls the image by association effect(p. 42). He gives as an example of how a
very American company, Gillette, was made to seem more British through its

34
sponsorship of cricket, a traditionally British sport. He also cites the case of IVECO
trucks, which were viewed as weak European vehicles, compared to more macho,
American trucks. Through a sponsorship of the macho sport of heavyweight boxing,
IVECO was able to shift this image in a highly successful way.
The image transfer mechanism is also found in other domains of commercial
association. In the context of celebrity endorsers McCracken (1989), points out that the
cultural meaning with which celebrities are endowed is the root of their success in product
endorsement. Clearly from his arguments, the image of the celebrity (and the meaning of
this image within the culture) are key to an effective meaning transfer process. This
process of transferring meaning from the celebrity to the product is more complex than
explanations which consider attributes of a celebrity, such as credibility and attractiveness,
as unidimensional, linear constructs.
In an empirical study of a transfer mechanism, Walker, Langmeyer and Langmeyer
(1992) have subjects rate brands and endorsers along several abstract dimensions. They
compare before association scores with after association scores along these dimensions to
see how products pick up some of the personality of the endorser. They find that
previously unendorsed products pick up more of the personality of the endorser than did
a previously endorsed product This leads to the following testable proposition with regard
to image transfer in sponsorship:
F22: Brands involved in few sponsorship activities acquire more of the image
and abstract associations of an event than brands which sponsor many
events.

35
As mentioned previously within the brand extension domain, Reddy, Holak, and
Bhat (1994) find that brands with high symbolic value extend better than brands with
low symbolic value. Given that in the case of sponsorship, the brand is the acceptor, one
might expect that brands with high symbolic value already would be less likely to have the
image of the event transferred onto them. Therefore it can be proposed:
P23: Transfer of image from event to brand takes place more easily, and
with greater magnitude, for low symbolic brands compared to high
symbolic brands.
Within the domain of combined brands Rao and Ruekert (1994) discuss how an
averaging model of attribute levels would lead to a zero sum game in which brand
alliances would be almost impossible to arrange. They argue in favor of a maximum
rule whereby the combined brand may receive the higher of the two brand
perceptions on each dimension. An example of this is the association between Lays
potato chips and KC Masterpiece barbecue sauce. While the combined brand gains the
desired associations of tangy and hickory smoke flavor from KC Masterpiece, an
average crispness between the chips and the sauce is not perceived. Crispness remains
at the higher of the two levels, the level of regular Lays. This leads to the following
proposition for the domain of sponsorship:
P24: Due to a maximum rule, brand images not fully congruent with an
event will not cause appreciable harm to the image of the event.
In fact, anecdotal evidence of reverse image transfer without negative impact upon
the brand has been cited. Jones and Dearsley (1989 cf. McDonald, 1991) find that the

36
image of football (soccer) was actually improved by a sponsorship by Barclays Bank
during a time when football was receiving a lot of bad press. Apparently this association
transferred stability from Barclays to the soccer league without transferring rowdiness
to Barclays. Perhaps positive associations are more easily transferred than negative ones.
This is certainly an interesting empirical question for future study.
Affiliation
Affiliation, like image transfer, can be stated as a message. The message of the
affiliation component of an association is that this brand is for people like me. The
primary ¡requirement for the functioning of this mechanism is that the event be perceived
as for people like me. The more this statement is perceived to be true, the more the
message about the brand being for me will be facilitated.
While there is no formal conceptual development of the way affiliation might
function as a mechanism of sponsorship effectiveness, some authors have hinted at the
idea. Meenaghan (1991b), for example, points out that the (sponsorship) message is
delivered by association with a socially intrusive activity (p. 8). Parker (1991) posits
that todays consumers, in addition to demanding products with desired attributes at a
reasonable price, also seek products which are for people like me. Sponsorship, he
proposes, is uniquely suited to helping a brand achieve the perception of being for
people like me. No further development of these ideas, however, is presented by either
of these authors.
There have been several studies of the impact of strong, social affiliations such
as political affiliations (Ellen, Wiener and Cobb-Walgren, 1991) and ethnic affiliations

37
(Deshpande, Hoyer and Donthu, 1986) on consumer behavior. At a broader level,
Westbrook and Black (1985) find that desires for affiliations with others is an
important determinants of some peoples shopping behavior. In addition to these
general impacts of affiliation with a group on consumer behavior, the possibility that an
affiliation with a brand or organization can be beneficially developed through the use of
marketing variables has been explored.
Macchiette and Abhijit (1992) describe affinity marketing as the use of group
affiliation to produce a strong promotional program. The resultant affinity groups are
said to have high levels of social bonding and cohesiveness. Members of affinity groups
are also said to be are committed to the objectives of the group. The authors propose
marketing generated affinity as one source of group centeredness.
An example of how marketing can take advantage of peoples existing affiliations
is the affinity credit card. Such cards bear the name and symbol of an organization to
which the card holder belongs. High levels of card holder loyalty have been claimed for
these cards (Worthington and Horn, 1992).
Bhattacharya, Rao and Glynn (1995) investigate the correlates of social
identification with an organization. By defining a museum as the product and members
as customers, the study investigates the determinants of affiliation with the museum.
They find that perceived prestige is positively correlated with ones affiliation with the
organization. Visiting frequency is also found to predict affiliation. Finally, participation
in similar organizations is negatively correlated with affiliation. Based upon these
findings, it can be proposed that

38
P25: Affiliation with an event is positively correlated with perceived
prestige of the event
P26: Affiliation with an event is positively correlated with frequency of
attendance.
P27: Affiliation with an event is negatively correlated with ones
participation in other activities.
In a related area, the manipulation of affiliation motivation has been cited as a
frequently employed advertising tactic (Zinkhan, Hong and Lawson, 1990). These
authors content analyzed a sample of magazine ads from selected years between 1935
and 1985. They found that over the time period studied, affiliation motivational appeals
increased in frequency. The authors contrast this increase with a decline in the use of
achievement appeals.
Implied size
The implied size mechanism also has a message which can be stated in words.
The message of implied size is that this company is big. The cognitive elaboration
required for this mechanism to occur is considerably higher that those previously listed.
A consumer has to, at some level, think to himself this must be a big company to have
this event named after it.
The idea that sponsoring an event can influence the perceived size of the
sponsor has been examined and has received some empirical support. McDonald (1991)
explored this possibility, but did not find evidence of an effect. He states, however, that
the reason sponsorship did not appear to affect perceptions of size was that most of the
companies studied were already clearly perceived to be very big and important (e g.,

39
Phillips and Cadbury). He posits that had the sample included more smaller companies,
an impact of sponsorship on perceived size might have been found.
Evidence that sponsorship can influence perceived size was found by Rajaretnam
(1994). In the evaluation of the unique experiment where a tire maker spent almost its
entire marketing budget from 1984-1987 on sponsorship, perceived size of company
increased by 19.7% during the period. Perceived financial health of the company, a
related measure, increased by 24.6% (p. 68). While this field experiment with no control
group does not prove that sponsorship can impact the perceived size of the firm, it
certainly allows for the following proposals:
P28: Event sponsorship influences the perceived size of a sponsoring firm.
P29: The larger the perceived size of the event, the greater is the impact of
the sponsorship on the perceived size of the firm.
Anecdotal evidence of the size of the sponsor affecting the perceived size of the
event have also been reported. Parker (1991) proposes that Mars sponsorship of the
London Marathon was the key to getting people to believe that it was a real event. The
following can therefore be proposed:
P30: Sponsorship of an event by a large sponsor increases the perceived size
of the event.
The marketing literature provides evidence of the benefits of perceived size.
Kirmani and Wright (1989) propose three possible mechanisms through which perceived
expenditure on advertising might lead to inferred quality. One is the notion that

40
confidence in quality leads companies to spend more on promoting a product. Another is a
perceived relationship between promotional spending and quality in some markets. Finally,
they cite the idea that perceived advertising expense is a signal of a firms financial
strength, probable social acceptance or some other factor that defines quality in some
markets (p. 344). This last mechanism is closest to the idea of perceived size. The idea
is that a company couldnt afford to sponsor a prominent event unless it is a large, stable
company. The authors do not investigate the perceived financial strength possibility,
however, stating that the default attribution is the most interesting possibility and the
one we pursue most directly in this research (p. 346).
Following up this work, Kirmani (1990) proposes that perceived costs will be used
as cues to quality when they are more diagnostic than other cues. She hypothesized that
this should take place under low involvement with the product class. The hypothesis,
however, was not supported. No effect of involvement was found.
Kirmani and Wright (1989) also list several undermines to the positive
relationship between perceived advertising expenditure and perceived quality. They
describe the desperation undermine as occurring when expenditures are so high that the
consumer assumes that the company is desperate to move a bad product. The immunity
undermine occurs when the consumer perceives that the company has nothing to lose, as
is the case with a fly-by-night operation. The no pain undermine is said to occur when a
high expenditure is perceived to be a small part of a companys budget. Finally, they
present the basic premise undermine which occurs when a consumer is confronted with
information which argues against the default attribution.

41
While both Kirmani (1990) and Kirmani and Wright (1989) report evidence of an
inverted U relationship between perceived advertising cost and perceived quality,
whereby beyond a certain point, quality perceptions actually go down. The data reported
in both cases could, however, simply reflect diminishing returns. Asymptotic functions
could have produced the significant quadratic terms reported. Also, no means or least
squared means were reported which might have shown the true shape of the
expenditure/quality relationship.
While it is not difficult to imagine situations under which perception of firm size
would have an impact on perceived quality, very little research has directly investigated
this effect. Within the context of the banking industry there is evidence that perceived size
of a bank can affect consumer beliefs about the likelihood of receiving a loan (Leonard and
Spencer, 1991). In the arena of legal services, however, there is evidence to suggest that
service providers believe perceived firm size to be more important than do clients
(Gaedeke and Tootelian, 1988).
Given the results of this research, it can be proposed that
P31: The sponsorship impact on perceived size has a positive impact on
perceived quality of the brand.
Implied Endorsement
The implied endorsement mechanism requires the most cognitive elaboration on
the part of the consumer, save reciprocity. The message of the implied endorsement
mechanism is this brand must be OK, or people wouldnt participate in an event named

42
after it The implied endorsement mechanism requires that, from a sponsorship
association, the consumer infers that the event is somehow endorsing the quality of the
brand. Several examples of this can be cited. Seikos sponsorship of track meets at which
it is the official timer is provided by McDonald (1991) as an example of what he calls
direct relevance(p. 36). By direct relevance is meant that one might logically conclude
that the official status of a sponsor implies some measure of quality.
Crimmins and Horn (1996) cite Visa and Seiko as brands which actually affected
consumer perceptions of quality due to their association with the Olympics (as measured
by DDB Needhams SponsorWatch service). Since Seiko was the official timekeeper
and Visa was the only credit card the Olympics would accept for tickets, it seems that an
implied endorsement may have led to the Olympic sponsorship success of these two
brands.
For the implied endorsement mechanism to function, however, does not require
some official sanctioning of the sponsoring brand. As long as a consumer infers that some
minimum quality level is required before an event will let a brand sponsor it, the implied
endorsement mechanism will be engaged. Usage of the brand at an event, however,
should increase the likelihood of this type of inference. Seeing top pool players compete
for large prizes on Brunswick tables, at the Brunswick Shoot Out for example, might
cause some consumers to infer a level of quality for Brunswick tables.
It can therefore be proposed:
P32: Use of the brand in the sponsored event facilitates the inferences
necessary for the implied endorsement to function.
P33: Having the brand name in the event name facilitates the inferences
necessary for the implied endorsement to function.

43
The ways in which implied endorsement can impact perceptions, attitudes, and
choice, have been well delineated in the source effects literature. Factors affecting the
ability of a source to persuade include trustworthiness (Brehm and Lipsher, 1959),
expertise (Petty, Cacioppo and Goldman, 1981), attractiveness (Landy and Sigall, 1974),
and credibility (Bimbaum, Wong and Wong, 1976). While the literature on source effects
is rich with predictive constructs, proposing the impact of these constructs in the context
of sponsorship first requires some empirical investigation. For example, it is not known
whether a credible or trustworthy event exists in the minds of consumers. If such things
do exist, it is not obvious what the determinants of these source characteristics would be.
Further elaboration on the role of source characteristics in sponsorship, therefore, requires
prior empirical investigation.
Reciprocity
Reciprocity is the mechanism which requires the highest level of cognitive
elaboration. The message of reciprocity is the sponsor supports events you care about,
so you should patronize the sponsor. The reciprocity mechanism is engaged when a
consumer makes a conscious decision to go out of her way to support the brands that
support the events which she cares about. While this may not be the most common
response to sponsorship, it is certainly a powerful one.
Surprisingly, there is evidence to support that fans of certain events demonstrate
significant reciprocity (or so they claim). Crimmins and Horn (1996) present what they
call gratitude whereby the fan of an event goes out of his way to buy the products of an
event sponsor. They cite a survey of NASCAR fans of whom 48% say they would

44
almost always purchase a sponsors product over a closely priced competitor. Reported
reciprocity for the Chicago Gospel Festival was even more impressive. As measured by
it makes me more likely to buy, fan reciprocity here was 82%. The authors go on to
point out that it is unlikely to get these effects from casual observers or uninvolved
spectators. Simply stated, they claim that Fans are grateful for sponsorship. Others are
not (p. 17). From these observations, it is possible to propose:
F34: The more involved consumers are with an event, the more likely the
sponsorship is to engage reciprocity.
Conclusion
Any attempt to study sponsorship must consider the multiple ways that an
association of a brand with an event can impact consumers. Seven mechanisms and their
psychological underpinnings were proposed and discussed in this chapter By breaking
sponsorship into these component mechanisms, it becomes possible to posit and test how
various aspects of a sponsorship association will impact success. Since there are multiple
ways in which sponsorship may work, considering which mechanisms are likely to be
engaged under various types of sponsorship associations allows many specific hypotheses
to be developed and tested. Empirical testing of numerous hypotheses relating to how
audience, event, and associational factors impact the functioning of these mechanisms is
presented in chapters eight, nine and ten. A thorough exploration of one important
associational factor, fit, is presented in the following chapter.

CHAPTER FIVE
DIMENSIONS OF FIT.
The better the fit of all elements, the better the chance that the sponsorship is
effective, because the relations between the sponsor and sponsored activity are
logical, natural, and easy to understand Otker (1988, p. 83)
The notion of fit is seemingly ubiquitous in the extant sponsorship
literature.(Roslow, and Laskey,1992; Thomas,1985; Crimmins and Hom,1996; Otker,
1988; Walshe and Wilkinson, 1994). This chapter attempts to define fit between an event
and a brand. Generic or respondent-reported perceived fit requires no development or
investigation of elaborate antecedents and component structures. It is simply a measure of
whatever a respondent thinks it is. Operationally, generic fit is measured by a subjects
responses to a question like how good is the fit between event X and brand Y. While
this measure may be a bit crude, the importance of quickly getting a handle on the fit
construct makes it useful.
A brief review of how the fit construct has been presented in sponsorship and other
commercial association domains is reviewed. Other, more technically defined types of fit,
such as attribute fit, audience fit and positioning fit are then discussed and defined.
The chapter concludes with a discussion of measurement issues.
Most sponsorship research refers to fit as the degree to which an association
makes sense. Walshe and Wilkinson (1994), for example, report that IBMs
45

46
sponsorship of a Pompeii museum exhibit was received much better than corporate
sponsorship of the arts had typically been received up to that point. Because IBM was
able to use its computers to bring a computerized archive of the dig into the museum, the
attendees not only saw a fit between the exhibit and IBM, they actually were able to
perceive a value added to their experience from the association. Negative reactions to the
association (measured by polling attendees) were, as a result of the association making
sense, much lower than negatives toward other similar art sponsorships.
Within the context of evaluating sponsorships, McDonald (1991) argues that
synergy between the brand and the event is essential. He proposes that assessing
synergy would be useful, as opposed to doing simple favorability tracking. He does not,
however, specifically define synergy nor does he propose how to measure it.
The idea of fit is also found throughout the rest of the commercial associations
literature. In the domain of brand extensions, Aaker and Keller (1993) define various types
of fit. They state that it is clear that the interactive relationship between quality and
the fit variables is sensitive to the levels of perceived quality and fit (in addition, possibly,
to different types of fit) (p. 56). These authors operationally defined fit to be the degree
to which the brand extension is a substitute or a complement for the core brand. Such an
operational definition, however, cannot be applied directly to the domain of sponsorship.
Broniarczyk and Alba (1994) reassert the importance of fit in brand extensions,
stating an assumption common to virtually all research, however, is that brand affect and
product category similarity play important roles. (p. 214, emphasis added). They
continue, we agree that, all things being equal, the best scenario for success involves
extension of a highly regarded brand to a very similar product class (pg. 215, emphasis

47
added). They then qualify these statements by explaining that this is not news to brand
managers, and that it misses the obvious differences in the brand specific associations
which are not common to all brands within a category. While they go into some detail on
why brand specific associations are an important consideration, they do not develop a
method for determining the brand specific associations for each brand, and assessing their
strength or importance.
Rather, they use pretests to identify brands with at least one strong, unique
association. They then test to see whether these brands are more successful at entering
product classes for which their unique associations are relevant than are brands which do
not possess these unique associations. While the authors make a significant point about
the unique associations possessed by brands (as opposed to categories), their method of
seeking out brands with strong, unique associates still does not answer the question what
is fit?. It does, however, suggest that if both the provider and the acceptor posses the
same strong and unique association, fit will be relatively high.
While specific antecedents of different kinds of fit are discussed below, the
implications for sponsorship of the Broniarczyk and Alba (1994) study can be formally
stated as:
P3S: Perception of generic fit by consumers will be high when the event
and brand share unique, salient attributes.
Within the context of composite branding alliances, Park, Jun and Shocker
(1996) propose structural determinants of fit between two brands. Specifically, they
propose that when the header brand (first brand in composite name) and the modifier
brand (the brand which follows by in the name) have complementary attributes, and are

48
not seen as logically incompatible, the resultant composite brand will have a good
likelihood of success. Key to their definition of good fit are complementary attributes
which are not logically inconsistent. In other words, sharing attributes is not necessary.
Here they argue that so long as the attributes are logically consistent, non-overlapping
attributes help the combination. Implications of this idea are discussed in detail in the
attribute fit section.
Fit has also received attention in the celebrity endorsement literature. While several
empirical outcomes of fit have been documented, little attention has been devoted within
this donum to defining or measuring things which lead to fit. Kamins and Gupta (1994),
for example, found that increased congruence between the spokesperson and the product
resulted in a more favorable product attitude. Speck, Schumann and Thompson (1988)
found that when the endorsers relationship to the product is more obvious, information
can be processed in a shorter time.
Ryssel and Stamminger (1988) report that Adidas has engaged in some interesting
measurements of the perceived personalities of various tennis stars, and how these
measures fit with the Adidas personality. They do not report in this paper whether or not
they measure brand personality along similar scales to those used to evaluate the celebrity,
but the article gives the impression that the company uses some sort of mechanism for
comparing celebrity personality with brand personality for fit. This points out the
interesting possibility that fit can include more abstract attributes, like personality
characteristics.
Despite its frequency of mention in the above literatures, there has been almost no
serious attempt to systematically define or measure the fit construct. There is little

49
detailed consideration of its potential antecedents, properties and consequences. There
are few details on how one might measure it. The rest of this chapter represents a first step
toward remedying this situation.
There are at least three kinds of (or approaches to) fit. While the three are likely
to be highly correlated in many situations, there are reasons to believe that they have
different impacts on the seven mechanisms discussed in the previous chapter. The three
types are attribute fit, audience fit, and positioning fit. This chapter first describes
attribute fit in terms of various psychological constructs, such as similarity (Tversky, 1977;
Tversky Gati, 1982) and categorization (Mervis and Rosh, 1982; Rosch, 1976) which give
insight into possible antecedents, properties and consequences of attribute fit, as well as
suggestions for its measurement. Audience fit and positioning fit are then presented.
Issues relevant to their special impact on several of the seven mechanisms are discussed,
as are issues related to their measurement. Specific predictions on how these types of fit
impact the seven mechanisms, however, are left for a future chapter.
Attribute Fit
Although similarity is recognized as fundamental to marketing and marketing
strategy, little attention has been paid to the way similarity judgments are
produced Johnson (1986, p. 59)
Attribute fit is the most general type of fit. It is not specific to sponsorship in any
way. All of the antecedents, properties and consequences are, in theory, as applicable to
any other form of commercial association as they are to sponsorship. Attribute fit is
defined as the degree to which two associates share common attributes, and do not

50
possess incongruous attributes. By attributes are meant associations with the brand or
event which are commonly elicited in the minds of consumers when thinking about the
event or brand. These associations can vary along a concrete-abstract continuum. An
example of a concrete attribute shared by sponsor and event is automotive. Speed is
a more abstract association, and macho is more abstract still.
This definition closely resembles Tverskys (1977) definition of similarity, which
increases with the addition of common features or the deletion of distinctive features.
More will be said of the relationship between similarity and attribute fit. First,
however, possible antecedents of attribute fit are discussed.
One thing which should influence attribute fit is the frequency and recency with
which two things are experienced together. Barsalou (1982), for example, investigated
the ways in which context-dependent properties of an object might be made to be more
context-independent. In other words, he explored how situation-dependent associates of
an object can be generated more spontaneously. He found that repeated simultaneous
exposure to the context-dependent property of an object can make the property more
context-independent.
An example of a context dependent property which can be made to be more
context independent is the ability of a basketball to float (Barsalou 1982). Most people,
when asked to lists things associated with basketball would not list its ability to float.
People who had frequently and or recently seen a basketball floating in a pool, however,
would be much more likely to come up with this association. This leads to the following
proposition:

51
F36: When attributes which overlap between event and brand are context
dependent, exposing consumers to the relevant context increases attribute fit.
Sponsorship associations are likely to be a context-dependent property initially,
but may, over time become context-independent. Initially, for example, Miller beer may
be a context-dependent property of NASCAR. At this initial stage, one would expect
Miller to cue NASCAR only when the sponsor is also cued, perhaps while one is watching
a race. As the association becomes stronger (as might be the case for a serious fan who is
spontaneously aware that Miller sponsors his favorite driver), Miller might become a
context independent property of NASCAR. In this latter case, one of the spontaneous
associations that comes to mind when this person free-associates about Miller beer would
be NASCAR.
It therefore seems possible that
P37: High involvement with an event leads to greater spontaneous awareness
for the brand-event association.
P38: Frequent attendance at an event leads to greater spontaneous awareness
for the brand-event association.
P39: Recent attendance at an event leads to greater spontaneous awareness
for the brand-event association.
In another study, Barsalou (1985) found that central tendency did not predict
graded structure in goal-derived categories. Goal-derived categories are categories which
are formed for a specific purpose, like assembling things to sell at a garage sale. From this
example, it is clear that goal derived categories are often categories which are formed in

52
order to make sense out of a group of objects which could not easily be separated by
preexisting, taxonomic categories. Barsalou found that frequency of instantiation predicted
graded structure in both goal derived and taxonomic categories to a large and
approximately equal extent. This provides evidence that, to the extent to which seeing a
brand paired with an event leads one to create an ad hoc category (Barsalou 1983), in
order to make sense out of what is being seen, frequently and recently seeing this brand-
event pairing should enhance attribute fit. Noting the continuous nature of this
relationship, Barsalou explains that the strength of association is free to vary
continuously as a function of how frequently and recently an association has been active in
working memory (p. 212).
P40: Frequent attendance at an event leads to greater salience of common
attributes.
F41: Recent attendance at an event leads to greater salience of common
attributes.
Markman and Gentner (1993) argue that mental representations (which could
include degree of fit between two objects) are formed through structural alignment
which is derived from analogical reasoning. They find support for the proposition that
the likelihood of a structural alignment between two stimuli is greater after a similarity
comparison than before (p. 435). It seems, then, that consciously considering the
similarity between two objects can shift attention away from the obvious, surface
dimensions of fit, to the more relational dimensions of fit. Since conscious consideration
of the similarity between the sponsor and the event is something which may be facilitated

53
by communications, the details of this finding will be discussed in the association
enhancing communications chapter. If, however, one supposes that frequency of
instantiation of the brand/event pairing might spontaneously lead people to consider the
similarity, then it may be possible that over several encounters with the pairing, not only
is fit facilitated, but the more abstract, relational dimensions of fit might become more
prominent. These ideas are formally stated:
F42: Frequent attendance at an event leads to greater increase in the salience
of abstract relational common attributes than in the salience of more
concrete common attributes.
P43: Recent attendance at an event leads to greater increase in the salience of
abstract relational common attributes than in the salience of more concrete
common attributes.
Within the broader research stream involving how people fit things into categories,
Cohen and Basu (1987) suggest that there is a real possibility... that the consumer
environment favors category definition in terms of specific exemplars rather than category
defining features.. . (p. 470). This is in reference to a perceived uncertainty as to when
categories will be defined by exemplars as opposed to being defined in a probabilistic
fashion (see also Smith and Medin, 1981). The critical issue here is that because brands
are not incredibly important to most people, the mechanism associated with categorizing
brands is not likely to involve capacity-depleting probabilistic rules. This seems to suggest
that when it comes to the fit between a brand and an event, the large bulk of the fit will be
determined by a perceived match between exemplary attributes, with less of a role for
attribute matches which do not immediately or spontaneously come to mind.

54
F44: Given peoples low motivation to form associated brands and events
into categories, attribute fit is determined primarily by overlap in chronically
available attributes of the two associates.
Having explored several potential antecedents of attribute fit, possible properties
of attribute fit are now considered. As mentioned above, the attribute fit construct is quite
close in definition to Tverskys similarity. It therefore seems appropriate to consider
other properties of similarity which make sense in the context of attribute fit. One
seemingly important fact is that Tversky (1977) generally argues that similarity should not
be treated as symmetrical. He proposes and demonstrates, for example, that North Korea
is perceived to be more similar to China than China is to North Korea. This is an example
of a more general idea, that the variant is more similar to the prototype than the
prototype is to the variant, because the prototype is generally more salient than the
variant (Tversky, 1977, p.333). This has several implications for the perceived fit
between sponsor and event.
P45: When the brand is larger (more salient) than the event, fit is perceived
to be greater when the association is stated in the form of the event fits with
the brand than the brand fits with the event
P46: When the event is larger (more salient) than the brand, fit is perceived
to be greater when the association is stated in the form of the brand fits
with the event than the event fits with the brand.
It should be noted that similar results have been obtained in a marketing context.
Johnson (1981) finds Shasta was judged to be more similar to Coke than Coke was to
Shasta. Here again, asymmetries are demonstrated such that an object with many

55
distinctive features is judged as less similar to an object with fewer distinct features than is
an object with few distinct features to an object with many. Put another way, a small (less
salient) brand is perceived to be more similar to a big (more salient) brand than is the big
brand to the small one.
Tversky (1977) also claims and demonstrates that at similar levels of salience,
similarity asymmetry is reduced, or nonexistent. An example of this case is a comparison
of the similarity between Hungary and Czechoslovakia. Since neither country represented
a prototype, the perceived similarity of Hungary to Czechoslovakia was about the same
as was the perceived similarity of Czechoslovakia to Hungary. The implications for the
perceived fit between sponsor and event is:
P47: For sponsors and events of equivalent perceived size (salience) there is
little or no fit asymmetry.
Another property Tversky (1977) finds for similarity is that it is not transitive. For
example, while Cuba is similar to Jamaica (both Caribbean islands), and Cuba is similar to
Russia (both Communist countries at the time) Jamaica is not perceived to be similar to
Russia at all. While this seems incredibly intuitive in the context of countries, it may be
less so in the context of the fit between sponsor and event.
For example, it seems entirely possible that a marketing manager might develop
rough heuristics by which to evaluate potential sponsorship opportunities. One possible
strategy might be to consider looking for other sponsors of events which his company
currently sponsors, then examine which other events those companies are sponsoring. In
fact, databases for sale to companies involved in sponsoring are set up to, among other

56
things, do this very thing (e g., IEG Sponsordex software). While such a database
might prove useful as a decision aid, it seems quite reasonable (and not unimportant) to
propose that
P48: Attribute fit between sponsors and events is not transitive.
A final property of similarity proposed and demonstrated by Tversky is its
relationship with dissimilarity. Similarity has generally been found to be strongly
negatively correlated with dissimilarity, but people attend more to the common features
in judgment of similarity than in judgment of difference (Tversky, 1977, p.340). As a
result, prominent pairs (e.g., East Germany and West Germany) are found to be more
similar than non-prominent pairs (e g., Ceylon and Nepal) in a similarity condition and
more dissimilar than non-prominent pairs in a dissimilarity condition. This effect has
also been demonstrated in a marketing context. Johnson (1981) finds that Coke and Pepsi
are very similar in the context of similarity, and very dissimilar in a dissimilarity context.
This suggests that
P49: When an event has many salient associations which are similar to
associations of the brand as well as many salient associations which are
incongruente the magnitude of similarity depends on the frame (fit vs. lack-
of-fit) such that the sponsorship association is perceived as having either high
fit or high lack-of-fit.
Before considering the consequences of attribute fit, it should be noted that
similarity has received some criticism for representing at best an intermediate measure for
predicting category membership, and at worst a midpoint in a circular argument. Murphy
and Medin (1985) point out that any two entities can be arbitrarily similar or dissimilar by

57
changing the criterion of what counts as a relevant attribute. They go on to present the
use of similarity to explain categorization as analogous to citing scoring more points than
the opposing team as a reason for victory in sports. The authors posit that rather than
being a true determinant of conceptual coherence, similarity may be a mere by-product
(p. 291). While the merits of this conclusion in the context of understanding category
structure is debatable, the criticism seems less relevant to the issue at hand. As stated
above, the process of evaluating fit between a sponsor and an event is likely to be only
tangentially related to taxonomic category structure. For this reason, the usefulness of the
similarity construct in the exploration of the properties of attribute fit seems to be
reasonably whole.
In order to understand how attribute fit might moderate the seven mechanisms, it
is first important to consider the direct consequences of attribute fit. As mentioned above,
to the extent that sponsorship associations elicit an attempt on the part of a consumer to
make sense of the association through a process of categorization, the categories formed
are likely to be what Barsalou (1983) refers to as ad hoc categories. He goes on to
describe and demonstrate several outcomes of such categories. Specifically, it is
demonstrated that ad hoc categories tend to be less established in memory, and therefore
less available in memory. This means that category concepts, concept-to-category
associations and category-to-concept associations will generally be less established in
memory compared to those of common categories (Barsalou, 1983).
This seems to indicate that the weaker the attribute fit, the less the sponsorship
associations will be available in memory. Therefore, it is possible to predict that

58
P50: The greater the attribute fit, the greater the recall for the association.
P51: The greater the attribute fit, the greater the recognition for the
association.
P52: The greater the attribute fit, the greater the recall for the event when
the brand is cued.
P53: The greater the attribute fit, the greater the recall for the brand when
the event is cued.
P54: The greater the attribute fit, the greater the recognition for the event
when the brand is cued.
PS5: The greater the attribute fit, the greater the recognition for the brand
when the event is cued.2
Murphy and Medin (1985) propose another interesting consequence of attribute
fit (or conceptual coherence). They propose that such fit can facilitate inference-making
about the relationship. Specifically, they present the possibility that coherent or useful
categories are the ones that allow the most inferences to be made. They continue by
proposing that a vague category... enables few if any inferences to be made( p. 293).
While they do not present conclusive evidence to support these assertions, it is certainly
possible to imagine how strong attribute fit might facilitate the kind of inferences
necessary for the functioning of mechanisms such as implied size and implied
endorsement. Therefore, it is tentatively proposed that
P56: Strength of attribute fit is positively correlated with the number of
inferences made about the association.
2 It should be noted that the linear form assumed is only proposed, and that other research
indicates other possible functional forms. The functional form of the relationship between
attribute fit and memory, therefore, is an empirical question.

59
FS7: Strength of association is positively correlated with the confidence in
ones inferences about an association.
Cohen and Basu (1987) reach some similar conclusions about categorization and
inferences. They elaborate on the potential of affect to serve as an intermediate variable in
the inference-making process. They propose that since categorization produces a
reduction in uncertainty, positive affect may result from a successful fit and negative affect
from an inability to categorize an item particularly if the resulting judgments or
inferences are important (p. 470). They go on to say that by relating the assignment of
affect to more fundamental categorization processes we are also more likely to anticipate
the type of inferences that consumers use to fill in information gaps when direct product
knowledge is absent or when motivation to seek such information is low (Cohen and
Basu 1987, p. 471).
These observations are relevant to sponsorship because they point out that the
affect generated by successfully fitting things into a category may moderate the number
and type of inferences generated. It can therefore be proposed that
P58: The affect generated by the association of a brand with an event
moderates the relationship between attribute fit and inference making
proposed in P56 and P57.
Ortony (1979) also provides some insight into the consequences of attribute fit.
This work attempts to go beyond literal similarity to include what he refers to as
metaphorical similarity. He defines many properties of metaphors, and relates these
properties to Tverskys (1977) concept of similarity. Key to this discussion is the idea that

60
the more metaphorical is a comparison, the more asymmetry there will be (e.g., cigarettes
are like bullets, but bullets are not like cigarettes). This is the idea which led to the quote
..the degree of symmetry is inversely related to the degree of metaphoricity( p. 172).
Ortony suggests that when no attribute of the second term in a comparison seems
to apply to the first, that a speaker (or listener, reader or writer) may reorder the salience
of the attributes of (especially) the second term in the comparison. (Ortony, 1979,
p. 173). This may have relevance to sponsorship. If one equates the metaphoricity
construct with the levels of abstraction of the association, then it is possible to make some
interesting propositions regarding levels of abstraction and attribute salience.
For example, when Porsche cars are compared to race cars, the comparison is
quite literal, and the salient attribute of a race car (speed) can be easily linked to the brand
Porsche. When McDonalds sponsors a race, however, the literal similarity is quite low,
and the desirable fast attribute is less likely to transfer. When a full-size pickup truck
sponsors a race car, however, the most salient attribute of the race car, speed, is not
relevant, and a reordering may take place until a match, such as can take abuse is found.
If this takes place, not only should the toughness of the truck be made salient, but the
toughness of the race cars should be made salient as well. Based upon Ortonys (1979)
results, the following propositions can be made:
P59: When attribute fit between event and brand is the result of concrete
attribute overlap, metaphorical elaboration does not enhance attribute fit.
P60: When attribute fit between event and brand is the result of abstract
attribute overlap, metaphorical elaboration enhances attribute fit.

61
Another consequence of attribute similarity may be a tendency to infer that one
associate has the attributes of the other associate. In other words, attribute fit between
sponsor and event along several attributes may lead to inferences that a brand possesses an
additional attribute of the event, which the brand is not known to possess. There is
evidence to support the idea that two things which are strongly associated can lead to
inferences about shared attributes which are not known to be shared. Sedikides, Olsen
and Reis (1993) investigate whether providing information about the relationship between
two individuals (that they are married) would lead to greater spontaneous transference of
attributes (e.g., is vegetarian) during recall than for pairs of non-married individuals.
They do, in fact, find that for married couples, more attributes of one spouse are
mistakenly attributed to another spouse than is the case for non-married pairs of people
(p<05). The implications of these findings for sponsorship might be
P61: The tendency to transfer attributes from event to sponsor occurs
more frequently under high attribute fit than under low attribute fit.
In addition to investigating possible consequence of fit, it is necessary to consider
possible consequence of lack-of-fit. Within the domain of investigating complex social
conjunction categories, Hastie, Schroeder and Weber (1991) cite evidence which suggests
likely char acteristics of attributions made about compound categories which are seemingly
incongruent. In this paradigm, the authors compare conjunctive categories formed by
pairing either congruent or incongruent attributes of a person. They find that attributions
about the incongruent compounds are more extreme along trait adjective scales than are
either the original components of the pairing or the congruent compounds. They also find

62
that freely elicited attributes of the incongruent compound include many attributes which
are not elicited by either simple category. They suggest that evaluation of incongruent
combinations may be based upon a more complex judgment process than evaluation of
simple categories.
Put another way, their results seem to indicate that when incongruent things are
paired, evaluation of the pair is extreme, and attribute elicitation about the pair includes
attributes which are not elicited by either of the paired items alone. Assuming the pairing
of event with brand engages similar mechanisms,
P62: Lack of attribute fit between sponsor and event leads to more extreme
evaluation of the association along trait adjective scales.
P63: Lack of attribute fit between sponsor and event leads to the elicitation
of attributes which are not elicited for either the event or sponsor alone.
P64: Evaluation of high lack-of-attribute-fit sponsorships requires consumers
to engage in more complex judgment processes than do sponsorships which
have better attribute fit
Finally, it is important to qualify all of the assumptions of a linear relationship
between attribute fit and consequences in light of findings which suggest more complex
relationships. Meyers-Levy and Tybout (1989), for example, demonstrate that products
that are somewhat congruent with their associated category schemas will stimulate
processing that leads to a more favorable evaluation relative to products that are either
congruent or extremely incongruent. In addition to the preexisting notion that very
incongruent pairings are unnatural, weird, and/or poorly liked, this research suggests the
possibility that extremely similar pairings are seen as mundane or boring, and therefore,

63
poorly evaluated. While the result obtained by these authors is not entirely intuitive, its
existence suggests some possibilities for the impact of attribute fit in sponsorship.
P65: In some situations, extremely high attribute fit results in lower
evaluation of the association than does moderate fit3
No discussion of a construct like attribute fit would be complete without an
examination of possible methods of operational measurement. Several interesting
possibilities are presented in the literature. Others are developed below.
Johnson (1986) performs a direct test of the contrast model (Tversky 1977) in a
consumer product domain. Specifically, the study finds that similarity judgments are
predicted by a linear combination or contrast of the average number of common and
distinctive features of the products being judged (p. 56). This suggests that attribute fit
might be measured by free elicitation of attributes (abstract and concrete) from
independent groups of subjects for an event and a brand. The attribute fit measure could
then be calculated as a function of the number of common attributes and unique attributes
of each associate. Given the possibility that events may generally elicit more abstract, and
fewer concrete attributes than brands, it may be necessary to normalize elicited attributes
along an abstract-concrete continuum before calculating the attribute fit measure.
Similar techniques have been suggested within the domain of brand equity
evaluation Farquhar and Herr (1992) point out the distinction between strong category to
brand associations (category dominance) and brand to category associations (instance
3 When this is the case remains an empirical question.

64
dominance). The authors also point out the implications of different levels and types of
dominance in the domain of brand extension. They point out that many researchers either
ignore directionality or look at the wrong associative direction. The authors also hint at
other brand aissociates, such as usage situations, benefits and attributes, but do not discuss
the implications of these associations in any detail.
They then discuss three methods for assessing category dominance. They first
describe naming methods, which involve a subject naming as many brands in a category as
possible in some short time, or naming some pre-specified number of brands. Brands
named, and especially named early-on are said to have high category dominance. Such
methods could be applied to sponsorship by asking subjects to list events sponsored by
X or brands sponsoring Y, using unaided recall as a surrogate for fit.
Farquhar and Herr (1992) also describe latency methods, which use reaction time
measures to determine category dominance. According to the authors, brands which are
quickly and correctly identified as members of the category have high category dominance.
Here a slight modification of the measurement method might be appropriate for
sponsorship. Cueing a consumer with an event, and measuring his/her reaction time and
accuracy to whether or not a brand is a sponsor would be a fairly good method of
measuring the spontaneous availability component of attribute fit. Reaction time
measures of this type have been demonstrated to discriminate between spontaneously
available properties and context dependent properties (Barsalou, 1982, experiment one).
Finally, the authors discuss facilitation methods, which involve providing
subjects with a category, then presenting them with a brand covered up with dots. The
authors claim that brands which are recognized earlier are category dominant. In the

65
context of sponsorship, an event rather than a category could be cued. Brands for which
the event cue resulted in facilitation of recognition could be said to have better attribute
fit.
Although all these measures could be used as surrogates for attribute fit, it seems
that all three techniques would be poor direct operationalizations of the attribute fit
construct. Such measures would really describe how successful a previous association had
been. This might be strongly influenced by things like duration of the sponsorship, and
amount of money spent on promoting awareness of the association. While these things
have been proposed to impact attribute fit, they are not truly components of attribute fit.
As mentioned at the beginning of the chapter, one measure against which others
could be compared would be some sort of respondent-defined, generic fit measure. Direct
measures of generic fit could be measured along continuous scales. Correlations between
these respondent-defined measures of generic fit and other measures mentioned above
would certainly help to illuminate the properties of attribute fit.
Audience Fit
One type of fit which is more relevant to sponsorship than other commercial
associations is audience fit. Audience fit refers to the nature of the overlap between the
audience for the event and target market for the brand. As mentioned earlier, audience fit,
in many circumstances, will be highly correlated with other types of fit. It does, however,
have some interesting properties with unique implications for several of the seven
mechanisms. Before describing these properties in detail, a brief review of the notion of
audience fit in the sponsorship literature is presented.

66
Many authors have espoused the commonsense idea that having overlapping
audiences is a good idea. Varadarajan and Menon (1988) suggest that any cause-related
marketing effort should share the demographics of the target market for the product.
Nicholls, Laskey and Roslow (1992) suggest that understanding the differences in
audience from one Hallmark Event to another may allow marketers to develop optimal
messages. Otker (1988) states that the chance of an effective sponsorship is maximized
when there is a perfect marriage of target groups.. .of sponsor and the sponsored body
(p. 85). Marshall and Cook (1992) found that 49% of respondents listed reaching a
specific target audience as the most important determinant in selecting an event to
sponsor. Thomas (1985) states that the first concern of a company with sponsorship in
mind should be to select the sport, the art or the kind of music which is.. .most appealing
to its consumer/audience (p. 322).
Other authors have written about the ability of sponsorship to reach different
and/or prized audiences. Crowley (1991) points out that sponsorship reaches many
audiences, including current and potential customers, suppliers, employees, the general
public and local community, the business community, and shareholders. He presents the
results of a survey which shows the relative importance of each of these audiences by firm
type, and orientation. Withcer et al. (1991) found that companies used sponsorship of
cultural and art institutions to reach opinion leaders and to achieve community relations
goals, whereas sports sponsorship was used primarily to reach the general public directly.
Wolton (1988) states that arts sponsorship is often employed in order to reach decision
makers, government leaders, and opinion leaders (p. 89).

67
Finally, there has been some elaboration on the notion that there are several
distinct event audiences which may respond differently to sponsorship. Hastings (1984)
makes a distinction between audience members who are directly involved in the
sponsored events as participants, those who are directly involved as personal
spectators, and those who follow the event through the media. The author notes that
event sponsorship differs from advertising in the clear delineation between various
audiences for an event. He points out that these differences must be taken into account
when creating potential objectives for a sponsorship. These different event audiences
could probably serve as a surrogate for an event-involvement construct which may have
an impact on the functioning of the seven mechanisms. Specifically, it can be proposed
that
P66: Event-involvement is highest for participants, lower for those who
attend the event live, and lower still for those who view the event through the
amass media.
Apart from these quotes and ideas, there is no extant elaboration on the audience
fit construct. There has been no discussion of its nature, properties, or impacts on
sponsorship success. This shortcoming is addressed below.
Audience fit involves two distinct dimensions. The first, which is widely
recognized in other contexts, is intersection, which simply refers to the size of the
overlap in the audience of the event and the target market for the sponsor. Intersection is
analogous to the media planning concept of reach (see Sissors and Surmanek, 1986).
As far as the intersection component of audience fit is concerned, an overlap of the target
market for brand X with event Y of 100,000 is equivalent in value to an overlap of the

68
target market for a brand X with event Z of 100,000, regardless of the absolute size of Y
and Z, or the relative size of event audience and brand market (see Figure 2 below). For
several of the seven mechanisms, intersection is probably the only component of audience
fit that matters.
The second component of audience fit is subsumption, or the degree to which
one audience is subsumed by the other. Here the relative audience size of the two events
does make a difference, as does the percentage of each audience which is represented in
the intersection. A few examples of this construct are given in Figure 2 for clarification.
/ GMC (
\ indy \ Equivalent
l Trucks l
/ 500 J Subsumption
/ GMC f
Y\-MudBog Event
1 Trueles
jj~'Truck luce Subsumption
Brand
Holly
Carburetors V\_
) ^ j Subsumption
FIGURE 2
TYPES AND AMOUNTS OF SUBSUMPTION
As can be seen in the above figures, the type and degree of subsumption depends
upon the area of intersection between the brand and the event, the size of the event, and
the size of the brand. Equivalent subsumption is defined by the intersection representing
the same proportion of the event audience as the brand audience. Event subsumption is

69
defined by the intersection of brand and event audiences representing a larger percentage
of the event audience than the brand audience. Brand subsumption is defined by the
intersection representing a higher percentage of the brand audience than the event
audience. Percent subsumption is simply the percentage of the subsumed associate
represented by the intersection.
Looking briefly at the right side of the above figure, it should be clear that there
are several differences in the nature of the 100,000 person intersections presented. In the
case of equivalent, low subsumption, the intersection of Indy Car People and GMC
People is not particularly predictive of the attributes of either group. In the case of Event
subsumption, the relatively high percentage of people watching the Mud Bog race who
are also in the pickup truck target market may cause people in the intersection to
increasingly believe that GMC Trucks are for people like me. Likewise, in the case of
brand subsumption, the high percentage of people in the high performance carburetor
market, who also watch the Indy 500, may lead people in the intersection to increasingly
believe that the Indy 500 is for people like me. In other words, when there is sufficient
percent of subsumption, affiliation with the smaller, more individuating associate should
flow to the larger, less individuating associate. From these ideas, the following
propositions can be made:
P67: Sponsorship under event subsumption leads to greater affiliation with
the brand than sponsorship under equivalent or brand subsumption
P(58: Sponsorship under brand subsumption leads to greater affiliation with
the event than sponsorship under equivalent or event subsumption

70
Positioning Fit
Positioning fit refers to the fit between the message of the commercial association
and the message of the rest of the integrated marketing communications campaign. To the
extent that this is explained by the similarity between the audience for an event and the
audience for traditional media vehicles employed by the brand, positioning fit will be
highly correlated with audience fit. But positioning fit also has a temporal component, and
a sub-segmentation component that make it worth discussing as a unique type of fit. The
temporal component refers to the fact that the marketing campaign for some products may
reflect certain seasonal characteristics. This makes the timing of events a determinant of
positioning fit. The sub-segment component refers to the fact that marketers often give
slightly different messages about their product to different sub-segments. This later
component is related to the unique ability of sponsorship to engage in partial
positioning, which is defined below. First, however, a brief summary of the literature
relevant to the positioning fit concept is presented.
Very little has been said about positioning fit as a determinant of success in
commercial associations. One exception to this is an article by Crimmins and Horn (1996)
who discuss the role of sponsorship within the overall marketing campaign. They stress
how important it is to check the fit between the sponsored property and your marketing
objectives (p.20). This is perhaps the closest thing to a direct reference to positioning fit
in the existing literature.
Within the related domain of corporate image management, Gray and Smeltzer
(1985) discuss the importance of having all of a companys communications have a

71
coherent message. Specifically, they state that the various sources that communicate
corporate image must be coordinated so that a congruent and effective image is projected
(p. 77). While sponsorship associations are not directly cited, their argument clearly
could include sponsorship as an image-communicating device.
Finally, within the domain of brand extensions, Park, Milberg and Lawson (1991)
point out the value of having all of a companys brands share a coherent position. They
state that the degree of perceived fit is a function of among other things, brand-
concept-consistency perceptions (p. 186). They also find that extensions work best when
both core brand and extension are either prestige or functional in nature. Therefore, at a
very high level of abstraction, two associates having a similar positioning seem to be an
indicator of success in some circumstances.
As mentioned in the introduction to this section, one reason why positioning fit is
particularly relevant to sponsorship is the temporal nature of events. An example of a
brand which takes on a somewhat seasonal positioning strategy is Coors Light In the
Summer, Coors Light is positioned as fun time beer to take with you to the beach so you
wont be too full to play volleyball. In the winter, the brand position seems to move
toward being a cozy, by the fire, after-ski beverage. In short, the changing seasonal
consumption situations for the product are reflected in the advertising of the product. By
sponsoring beach volleyball in the summer and downhill ski racing in the winter, the
seasonal positioning is reflected in the sponsorships chosen.
A related, and perhaps more interesting facet of the positioning fit between event
and brand is the sub-segment component. Much like some brands have seasonal
positioning strategies, others try to position their brands as different things to different

72
sub-segments. Clearly, it is not possible for a brand to be all things to all people, but the
degree to which brands modify their messages to fit different segments is somewhat
remarkable.
Budweiser beer is a case in point. While Bud positions itself as the working
mans beer on weekend sports shows, the message it presents on David Letterman and
MTV is quite different. This takes place for two reasons. One reason for this is that the
MTV audience may not be interested in a beer that is just for blue collar types. Secondly,
and more importantly is the fact that the working men do not stay up on Sunday night to
watch 120 Minutes on MTV. Hence, a partial positioning strategy is possible because
the more avant-garde message is never seen by the core market who would find it
confusing at best, and disturbing at worst.
Sponsorship, where the association is the message, represents a nearly no
downside way for a firm to try partial positioning. An example of this is AT&Ts major
sponsorship of the Rock and Roll Hall of Fame. While AT&T is not generally considered
to be a rock and roll kind of company, anyone who would find this association to be a
negative for AT&T never sees it. The key to the ability of sponsorship to engage in partial
positioning in this case is the fact that the message is limited to those who go to the
museum.
Conclusion
This chapter has examined how the various types of fit relate to sponsorship. A
brief review of how the fit construct has been presented in sponsorship and other
commercial association domains was reviewed. Other, more technically defined types of

73
fit, such as attribute fit, audience fit and positioning fit were then discussed and
defined. The chapter concluded with a discussion of measurement issues. While this
chapter did address the idea of lack-of-fit, it did not explore the potential outcomes of
associations which are viewed by consumers as inappropriate. This issue is explored in the
next chapter.

CHAPTER SIX
SACRILEGE.
Tottenham Hotspur, Liverpool, Manchester United: these names have magical
associations for those -who -worship the great football god Singing, swaying
masses on Wembley terraces. Muddy mayhem in the twilight gloom. The glory
and the gory. Red vs. Blue. Catholic vs. Protestant. North vs. South. And now...
AVCO vs. Crown Paints. JVC vs. Sharp. Holsten vs. Home Ales. Wright, (1988,
p. 104)
No discussion of sponsorship would be complete without acknowledging the fact
that some consumers may think that it is going too far. In an age of increasingly bold
associations, the problem should only continue to grow. One acknowledgment that
people may be getting fed up is evidenced by Taco Bells April Fools Day prank in
which the company placed an ad in the April 1, 1996 New York Times proclaiming:
TACO BELL BUYS THE LIBERTY BELL
In an effort to help the national debt, Taco Bell is pleased to announce that we
have agreed to purchase the Liberty Bell, one of our countrys most historic
treasures. It will now be called the Taco Liberty Bell and will still be accessible
to the American public for viewing ....
Clearly, at least among the advertising creatives at Taco Bell, there may be a problem.
The idea that maybe there are some things which are sacred, and therefore
should not be involved in commercial association is an intriguing one. The idea that some
products may be so profane that any association with a non commercial entity would
74

75
result in poor evaluation and even outrage is equally attractive. When combined, the two
notions form a complex and largely unexplored area in marketing. This area is referred to
here as sacrilege.
The notion that consumption objects can take on dimensions of being sacred or
profane has received some attention in the consumer behavior literature. Belk,
Wallendorf and Sherry (1990) describe how the processes of the secularization of religion
and the sacralization of the secular are at work in contemporary society. They find that
consumers maintain the sacred in consumption through separation of sacred and profane,
and through the performance of rituals. Belk and Wallendorf (1990) explore the sacred
and profane dimensions of money itself. They find that the sacred and profane aspects of
money depend upon its sources and its uses.
Hirschman (1991a) explores the way advertisers utilize sacred and profane
consumpti on imagery to persuade. In an exploration of process of the secularization of the
sacred, Hirschman (1991b), describes how the commercial transactions involved with
surrogate motherhood and in-vitro fertilization can be made less profane. She finds that
this is achieved by stressing the sacred aspects of the acts of childbirth and family creation.
While the sacred and profane nature of consumption objects has received attention,
there has been no examination of these issues in the context of commercial associations.
Some might argue that words like sacred, profane and sacrilege are too strong to be
used in the context of commercial associations. There are several recent examples,
however, which, seem to meet the more traditional notion of sacrilege. DeCoursey (1995)
reports on the widespread feeling of over-commercialization associated with the Popes
1995 visit to the United States. Brewer (1993) describes attempts to associate the Pilot

76
Precise Pen with Middle East peace4. Brewer (1993) also reports on an attempt by K-
Swiss shoes to capitalize on the fact that the Pope bought a pair of their sneakers while in
Denver. Finally, there is the case of the McDonalds World Cup Soccer promotion in
which sacred Koran scripture was featured on two million throw-away hamburger bags.
This promotion greatly offended Muslims in Great Britain (Alonzo, 1994).
The above examples seem to provide evidence of sacred events or objects which
are difficult to successfully exploit through commercial association. Similar positive
examples of associations which failed because the product was too profane, are not as
readily available. While Tambrands was among the corporate sponsors of the Womens
National Basketball Team (Rosner, 1995) and Angel Soft bathroom tissue is involved with
a cares for kids program (Weisz, 1995), there are no known cases of either feminine
hygiene products or bathroom tissues having an event named after them. The same is true
for other product classes found to be causes of irritation (Aaker and Bruzzone, 1985),
such as laxatives and breath fresheners.
Ail interesting and complicated possibility is that sacrilege may be perceived for
brand-event associations in which the brand is not strictly profane and the event is not
strictly sacred. Some combinations of associates, for reasons which have not yet been
examined, may lead consumers to consciously elaborate on the "inappropriateness" of the
association. This may sometimes go beyond mere elaboration in ones own mind, leading
to actual verbalization. Discussion among consumers about the crass and commercial
nature of the association can then ensue. The impact of such discussion among consumers
about associations is also currently unexplored.
4 The pen was apparently used in signing some of the documents.

77
An example of this might be the case of the Tostitos Fiesta Bowl, which was
played to determine the national college football champion of the 1995 season. While the
Fiesta Bowl may not have been strictly sacred and com chips may not have been strictly
profane, discussions which were seemingly the result of perceived sacrilege could be
heard among Florida Gator fans around the time of the game. Comments like Thats not
right.. .the taco chip national championship and Good grief, the Tostitos logo on the
sweatshirts is bigger than the words Florida Gators were common.
Using a small, convenience sampling of comments, it also seemed that the people
who were bigger fans were more likely to express disdain for the association. While there
is absolutely no empirical work in this area, it is possible to propose the following:
P69: The more mundane the brand, the more likely the association to result
in perceived sacrilege.
P70: The more sacred the event, the more likely the association to result in
perceived sacrilege.
P71: For associations which run the risk of sacrilege, greater consumer
involvement with the event results in a greater likelihood of perceived
sacrilege.
P72: For associations which run the risk of sacrilege, lower levels of attribute
fit lead to a greater likelihood of perceived sacrilege.
P73: The more blatant the perceived commercial intent of the association, the
greater the likelihood of perceived sacrilege.
The most frequent response to perceived sacrilege may well be for the consumer to
elaborate that those two things dont belong together. This elaboration may provide the
necessary condition for a dissimilarity frame. Given that people attend more to the

78
common features in judgment of similarity than in judgment of difference (Tversky, 1977,
p.340), it is possible to propose:
P74: Under perceived sacrilege, attributes common to event and brand
become less salient.
P7S: Under perceived sacrilege, attributes not shared by event and brand
become more salient
Therefore,
P75: Under perceived sacrilege, attribute fit is reduced.
The elaborations generated by perceived sacrilege may well be the commercial
association equivalent of counter-arguing. This is serious, because one of the benefits of
commercial associations in general is the inability of consumers to counter-argue in the
classic sense, when the association is the message. It is therefore possible to propose
that:
P77: Sacrilege-inducing associations will lead to more counterarguing than
will non-sacrilege-inducing associations.
Conclusion
While consumer outrage at associations considered to be sacrilegious is understudied, the
growing number and forms of commercial association make it an especially salient avenue
for further research. As the boundaries of what can be sponsored get pushed farther and
farther, the importance of this area of investigation should only increase. One way to
minimize the chance of negative consumer response to an association is through additional

79
promotional effort which explains the association to the public. The potential benefit of
such additional communication about the association is presented in the following chapter.

CHAPTER SEVEN
ASSOCIATION-ENHANCING COMMUNICATIONS (AEC)
As alluded to earlier, sponsorship associations do not exist in a vacuum. Firms can
engage in a variety of communication activities in order to maximize the impact of the
association for which they have paid. These communications are referred to here as
association-enhancing communications, or AEC. The expenditures above and beyond
the fee paid to be the sponsor are genetically referred to by practitioners as
leverage(Aiidrews 1996). Leverage can attempt to enhance awareness of the
sponsorship among consumers through advertising. It can also, however, take the form of
tie-in promotions, sales promotions, case lot discounts, and other perks to the channel
members in conjunction with the event.
Leverage has received some attention in the sponsorship literature. Meenaghan
(1991a) defines leveraging as the additional effort, largely promotional, which must be
invested by the sponsor in order to properly exploit the opportunity provided as a result of
securing particular sponsorship rights (p. 43). He goes on to say that it is generally
agreed that an amount at least equal to the sponsorship fee should be allocated to
leveraging the sponsorship. Otker (1988 ) presents a somewhat higher estimate, stating
that for each dollar spent on buying, another three dollars is spent on exploitation (p.
77).
80

81
Gardner and Shuman (1987) empirically investigate the amount spent on leverage.
They survey 300 Fortune 500 companies about their sponsorship activities. Eighty-nine
percent of responding companies engaged in sponsorship report some form of project
support, such as advertising tie-ins. The authors find that within their sample, an average
of forty percent of sponsorship cost was spent on leveraging the sponsorship (including
advertising, promotion, and employee tie-ins). They also find that less than ten percent of
companies spend more money leveraging the sponsorship than the sponsorship actually
costs.
Although the optimal level of spending on leverage seems to be a matter of some
dispute, the facts seem to support the idea that most sponsorships involve some additional
spending on communications about the association. The interesting question then
becomes what should be said and to whom? Crimmins and Horn (1996) propose that
when the link between an event and a brand is clear, as with basketball and athletic shoes,
consumers can make the link themselves. When the link is harder to forge, however, the
link must be made for them, as with Visa driving home the message that the Olympics
dont take American Express. While this advice seems to be face valid, it does not answer
the biggest question. When does the meaning of the association need to be driven
home? When does this help? When might it hurt? When might it be merely a waste of
money?
In the related area of brand extensions, Aaker and Keller (1990) find that
elaborating on brand extension attributes improves evaluations of the extension. They
propose that this occurs by alleviating negative inference making. Explaining which

82
attributes the extension shared with the core brand is found to be superior to elaborating
upon the quality of the original brand in generating positive evaluations of the extension.
An example of how this is relevant to sponsorship may make the point clear.
Suppose a dry cleaner is sponsoring a stock car. While attributes of a race car like
speed are positive for a dry cleaner, attributes like dusty or oily are not. Given the
Aaker and Keller (1990) findings, and the implication of the above, it can be stated:
P78: Communications stressing the positive attributes of an event for a
brand have greater impact on brand evaluation when the event possesses
both brand-positive and brand-negative attributes than when the event
contains only brand-positive attributes.
Previously the findings of Markman and Gentner (1993) were discussed. These
authors found that the act of merely considering an association had an impact on subjects
structural alignment. Subjects who were told to make a similarity comparison tended to
associate two scenes along more abstract, relational dimensions. Assuming that
association enhancing communications (AEC) can successfully induce people to make a
similarity comparison, then it can be proposed:
P79: AEC efforts which explicitly tell a consumer to see the fit between
event and sponsor will lead to greater perceived fit
P80: The effects predicted in P79 are greater when the event and brand share
attributes which are abstract and relational than when the event and brand
share attributes which are concrete.
Quite a bit of research has been done which focuses on whether explicitly stating
conclusions is helpful. Kardes (1988) found that more positive brand perceptions are
derived from allowing the consumer to draw his/her own conclusions in advertising.

83
Sawyer arid Howard (1991) demonstrated that for high involvement individuals, open-
ended advertisements were superior to closed-ended ads in terms of attitudes, purchase
intention, and choice. The authors found no significant difference, however, for the low
involvement groups.
These studies seem to contradict early work which found that explicitly drawing
the conclusion for the audience was superior to letting them draw the conclusion
themselves. Hovland and Mandel (1952), for example, found that over twice as many
subjects agreed with a speakers argument when the conclusion was explicitly drawn by
the speaker as opposed to being left up to the audience (neither intelligence nor
personality type had a measurable impact on this effect). The theoretical explanation for
this seeming inconsistency is that the impact of omitting the conclusion depends upon how
difficult the conclusion is to draw. If it is very difficult, then few people will draw the
appropriate inference, and effectiveness is reduced. If people are capable of getting it
on their own, however, omitting the conclusion should be more effective.
Several authors have hypothesized as to why this might be so. Linder and Worchel
(1970) demonstrate how mechanisms like cognitive dissonance can lead to difficult self
drawn conclusions being believed more than conclusions which were easier to draw. One
explanation for this finding is that people want (or need) to find value in accomplishing a
more difficult task, and that discovering something with which one agrees is more valuable
than discovering something with which one does not agree.
Similar ideas have been put forth in the domain of categorization, which is more
closely related to the pairing of commercial associates. Cohen and Basu (1987), for
example, propose a direct impact on affect of discerning what two things have to do with

84
one another. Specifically they state that since categorization produces a reduction in
uncertainty, positive affect may result from a successful fit and negative affect from an
inability to categorize an item, particularly if the resulting judgments or inferences are
important (p. 470). Combining this idea with the findings of Linder and Worchel (1970),
it can be proposed that
Pill: For those who determine how a sponsor is like an event on their own,
the more difficult the association is to understand, the more positive affect
will be generated for the association.
P82: The effect predicted in P81 is greater for consumers who are highly
involved with the event than for those who are less involved with the event
The above propositions must be qualified, however, by the generally accepted idea
that for most people, figuring out what a sponsor has to do with an event is not an activity
likely to receive much cognitive effort. Therefore, any association that requires more
than the simplest consideration to understand will likely be ignored. This fact does not,
however, make the above discussion irrelevant. On the contrary, it allows some specific
predictions about the effectiveness of certain strategies for association-enhancing
communications. Some specific AEC strategic proposals are as follows:
P83: Communications which explain obvious associations harm the
association.
P84: Communications which all but explain the associations that are not
oibvious are superior to communications which actually do explain the
association.

85
Conclusion
Given the industry rule of thumb of spending three times the sponsorship fees on
leverage, the study of how and when these communications about the association will be
most effective is an interesting area of investigation. While the extant literature has done a
thorough job of looking at the impact of omitted conclusions in the context of traditional
advertising, the unique mechanisms involved in sponsorship warrant further research into
how to best spend these billions.
This chapter has considered how firms can maximize the impact of engaging in
sponsorship through leverage, or additional spending above the sponsorship fee. Such
additional spending on association enhancing communications was posited to have a
measurable impact on several sponsorship variables. Consumer difficulty in determining
what a brand and event have in common was proposed as a key variable in deciding how
much should be said about an association. Formal propositions about how AEC will
impact these variables were also presented.
The last several chapters have explored broad aspects of the relationship between
sponsorship variables at a somewhat abstract level. The following three chapters focus
more narrowly on a subset of the issues raised thus far. These chapters develop and test a
two-stage model of building brand equity through sponsorship. After an explanation of
the models conceptual underpinnings, eight empirical studies are presented which test
various components of this model. Chapter 11 then summarizes the findings from the
eight studies, and discusses their theoretical and managerial implications. The dissertation
concludes with suggestions for future research.

CHAPTER EIGHT
THE IMPACT OF EVENT AND ASSOCIATIONAL FACTORS ON INFERENCES
ABOUT THE BRAND
The preceding chapters have provided an overview of what is currently known
about sponsorship. They then went on to map out the somewhat vast unexplored
territory of how sponsorship may work. While the wide array of unexplored issues
makes the general area of sponsorship an exciting domain in which to work, it also makes
it necessary to limit to a subset the number of issues which can reasonably be addressed
within the scope of a dissertation. This first empirical chapter, therefore, begins with a
brief discussion of which issues were selected for empirical investigation and why they
were chosen.
The core unexplored issues in sponsorship are how and when does associating a
brand with an event build equity for the brand, and, ultimately, increase choice probability
for a segment of consumers. Given that the mechanisms at the lower end of the
necessary elaboration continuum (i.e. awareness, affect transfer) are not expected to
differ functionally from their counterparts in the domain of advertising, it seems most
fruitful to pursue those mechanisms at the upper end of the continuum. Very little is
known about mechanisms like implied size, implied endorsement and reciprocity. These
mechanisms are also likely to be more greatly impacted by sponsorship than by other
forms of promotional activity.
86

87
These mechanisms are so far up the necessary cognitive elaboration continuum
that they all strongly rely upon consumer inference making in order to function. For a
sponsorship association to impact the perceived size of a company or brand, a consumer
must make an inference that this company must be big in order to be able to sponsor such
an event. In order for the implied endorsement mechanism to be engaged, a consumer
must make the inference that this brand must be OK, or else it wouldnt be allowed to be
the sponsor of this event. In order for reciprocity to be engaged, a consumer must make
the inference that this brand is doing something to make possible and/or better an event I
care about.
The realization that inference making is key to so many important, uniquely
sponsorship-related mechanisms has led to the development and testing of the inference
based model of building brand equity through sponsorship. This model yields many
specific hypotheses about how audience, event, and associational factors impact the
inferences consumers draw about sponsoring brands, and how these inferences impact
choice probabilities for various classes of brands. The inference-based model is depicted
graphically in Figure 3.
As the figure makes clear, the inference-based model consists of two stages. In
stage one the consumer who is exposed to the sponsorship association makes inferences
about the brand. The number and kind of inferences made are determined by event and
associational factors, as well as by characteristics of the audience members relationships
with the event (audience factors). Stage two consists of converting these inferences into
changes in choice probability for a particular brand on the basis of the sponsorship-based
inferences. Here, categories which differ in experience ambiguity, and brands which differ

88
in their goodwill position, are expected to be differentially impacted by each type of
sponsorship-based inference.
FIGURE 3
THE INFERENCE-BASED MODEL OF BUILDING BRAND EQUITY THROUGH
SPONSORSHIP
The remainder of this chapter details a series of three studies which investigate the
impact of event size, brand/event fit, duration of association, and share of presence
(defined below) on consumer inference-making about a sponsoring brand. The following

89
chapter describes two experiments which explore the roie of audience factors on consumer
inference making. Chapter ten explores how sponsorship-based inferences differentially
impact different brands and categories.
Conceptual overview and construct definition for studies la, lb and lc
Studies la,b and c investigate how event and associational factors impact inference
making. Specifically, they explore how event size, brand/event fit, share of presence, and
duration of association impact inferences about brand size, brand legitimacy and event
facilitation. These constructs are defined below.
Event Factors
Event Factors, or attributes of the event being sponsored, are expected to
directly impact consumer inference making about the sponsoring brand. Many aspects of
an event, such as perceived size and prestige, might influence inferences about sponsoring
brands. Pretesting was undertaken to determine the relatedness of these constructs. One
hundred eleven actual sponsored events were evaluated by independent groups of subjects.
Perceived prestige was evaluated along a seven point scale anchored by very prestigious
and not prestigious at all (n=21). Perceived size was evaluated along a seven point scale
anchored by very large and very small (n=18). Mean scores for each construct were
calculated. The means for both constructs, for all 111 events, were then submitted to
simple correlational analysis. The correlation coefficient between the two constructs was
found to be .785.

90
While this correlation was quite high, further investigation of the data attempted to
determine if the two constructs could be orthogonally manipulated experimentally. While
it was possible to find events which differed in prestige and size, these events fell into very
specific event categories. That is, nearly all of the high prestige, low size events were arts
related, while nearly all the high size, low prestige events were dirt-track motor sports or
rodeos. Rather than limit the initial test of the inference-based model to these rather
unique categories of event, where size and prestige were poorly correlated, a decision was
made to include only one of the two constructs. This was done with the knowledge that
the chosen construct is most likely correlated with the other. Given the interesting
theoretical predictions about the impact of event size on the perceived size mechanism, the
perceived event size construct was chosen for exploration.
Perceived Size
Perceived size of the event is expected to impact inferences about the size of the
sponsoring firm. Here, consumers are expected to infer that only a large brand could
afford to sponsor a large event. Perceived size of the event is also expected to affect
inferences about the commitment of the brand (or firm) to supporting the class of activity.
For example, an athletic shoe sponsoring a large tennis match might be perceived as more
committed to tennis than an athletic shoe which sponsors a smaller, less important tennis
match. Perceived size of the event is also expected to impact consumer inference making
about the necessity of the sponsor to the event taking place (event facilitation). Here,
sponsoring a small event is more likely to lead consumers to infer that without the support
of the brand, the event would not take place.

91
TABLE 1
MEAN PERCEIVED PRESTIGE AND SIZE OF 111 EVENTS AND
ORGANIZATIONS
Eywit/OnuntaBon
fiat
Pristine
1
1
c
1
fia
Pristine
m2
Praatiaa
AMA Proffering
4.00
3.33
Florida State Fat
2.88
2.62
Pittsburgh Three Rives Regatta
289
4.X
AMC Cancar Research Center
5.36
6.00
Heed of tvs Chales Regsito
3.00
3.43
Poiarts Amphitheater
267
3.33
American Cancer Satisfy
6.11
6.76
Henry Ford Mutaun
3.44
4.10
Poxfcys Rating Te^n
200
1.X
American Diabetes Assoczcn
5.94
6.57
Houston Uvestodc Show
2.72
Z29
Power Team
267
2.67
American Hsert Ctitf
5.17
6.00
IMAX Corporation
4.50
3.95
Raserarch Hospital /ALSAC
4.S6
5.14
American Horae Shows Acsce.
2.44
3.62
IndyCar
5.00
4.14
Ringing Bra. Grajo
6.56
4.10
American LumbaJeck
1.86
229
inf^ Open LA (Tennis)
4.06
5.14
Rockdeilar Canter
5.61
6.14
American PooJpiaycra
222
2.81
Intemstocd Fcoilvd of lighto
4.22
4.19
Rocky Mountain Bk Fcund^cn
3.11
3.52
American Quarter Hcm>
2.36
3.43
Jeffrey Ballet
Z67
4.71
Roys! Academy of Arts
3.44
4.48
American Racing Series
3.61
3.33
Junior Adhlov'smcnt Inc.
3.78
3.81
San Diego Street Scene
2.X
2.76
Arthritis Foundation
5.36
5.52
Kennedy Center for Perf. Arts
4.56
5.48
San Diego Zoo
3.83
4.81
Baltimore Zoo
4.11
4.43
Kennedy Krieger institute
3.33
4.43
Sat Fmtdsco Jazz Festive
3.X
4.43
Big Apple Cira*. Ltd.
3.36
3.33
KO Drugs Boring
3.36
3.14
Science Museum of Minnesota
2.78
4.X
Brookfield ZooObcs^o
3.86
420
Lttdcnd Prowlers Hockey Teem
2.56
Z43
Shea's Performing Arts Caite-
3.17
4.33
Builders Squsr Alamo 0c .71
4.28
3.57
Lbarty Soane* Carter
2.86
3.X
Shed Aquarium
2.94
3.67
Calgary Stenpato
2.83
2.62
nctri Pssrk Zoo Chicago, IL
3.22
4.62
SL Marys Food Bank
4.28
4.71
Calgary Zoo
3.50
4.24
Martin of tornos
5.61
6.X
lEsiecfCckrado
239
2.81
Canadian Figure Skating Aaaoc.
322
4.62
Mary Bridge Children's Hospital
4.11
5.29
TadeGattoy
3.17
4.10
Canadian Olympic Aaaoddtan
4.33
5.67
Ete&te! Museum of Rna Arts
2.86
4.76
Taam Danmark DENMARK
Z44
3.81
Capital City RtariTaat
3.00
3.62
Multiple Sclerosis Society
4.78
5.X
The ChMerms Thtoter
3.44
4.X
Carnegie Hal
5.86
6.48
Museum of Science & Industy
4.X
4.X
Tho Nsiur Conafflvuncy
4.X
4.X
Chertctt Motor Speedway
4.44
3.00
Mystic Seaport Museum
3.11
3.71
Tcdttbe Tannis Ctasaic
3.72
5.X
Cherry Creak Arte Festival
2.94
3.62
Names
3.17
3.86
US. 10K Ctasaic
ZX
3.76
CMdrVs Home Sodoty
3.50
4.57
Nd Assoc, of tntsrcoQ. Athletics
4.78
4.81
U.S. Luge Association
Z94
3.33
Childrens Hospital o Alabama
4.11
4.71
Nal AsBsoc.Seccndsry Schools
322
3.20
U.S. Taekwondo Union
Z78
3.38
Children's Hospital of Pittsburgh
4.67
5.06
National Captioning Institute, Inc.
ZX
3.43
United States Sailing
3.17
4.10
Children's Miracle Network
6.17
6.00
NCAA
6.X
5.71
Untied States Termte Association
4.61
5.29
Children's Museum of todtonopoHs
3.67
4.14
National Geographic Society
5.44
5.38
Untied Way
6.39
5.X
Child's Play Touring Trieste
3.33
4.24
National Wiidife Federation
5.17
5.24
University Musical Society
3.61
4.52
Cirque duSoia
3.22
4.57
Nsw Jaraay Bdtoon Fes&v^
Z18
Z57
USA shooting
2.94
3.24
City of Hope
3.39
4.62
NMMA Bod Tour
Z22
3.X
USA Wresting
4.X
3.19
Caches v. Cancer
5.00
5.38
Nynet
4.X
4.14
Vancouver Art GaNery
2.78
429
Cranbroek Institu of ScteK
3.00
4.81
Optyland
2.67
3.29
Venetian Festival
289
4.X
Downtown District Houston. TX
3.50
3.43
QrotgeBowfl
6.22
5.X
Walt Dianeys World on ice
522
4.57
Bvis Preey^ Gracaiand
3.44
3.52
Qr^on hGJtespsi? Fodivd
267
3.71
Washington DC Paries
3.00
4.29
Epilepsy Fcundslon o America
4.72
5.81
Osfem America
239
3.14
Western Idaho Fair
272
Z24
Feed the Children
5.17
5.76
PedAc Nrtional Exhtotiicn
283
3.24
Wildlife Conservation Society
4.00
4.52
The idea that sponsoring an event can influence the perceived size of the sponsor
has been examined and has received some empirical support. McDonald (1991) did not
find evidence of such an effect. This may have been due to the fact that perceived size of
companies studied was already very high (e.g., Phillips and Cadbury). Evidence that
sponsorship can influence perceived size was found by Rajaretnam (1994) who reports
that a tire maker that spent almost its entire marketing budget on sponsorship
dramatically increased its perceived size.

92
The marketing literature provides other evidence of the benefits of perceived size.
Kirmani and Wright (1989) propose three possible mechanisms through which perceived
expenditure on advertising might lead to inferred quality. One is the notion that
confidence in quality leads companies to spend more on promoting a product. Another is a
perceived relationship between promotional spending and quality in some markets. Finally,
they cite the idea that perceived advertising expense is a signal of a firms financial
strength, probable social acceptance or some other factor that defines quality in some
markets (p. 344). Kirmani (1990) proposes that perceived expenditure will be used as
cues to quality when they are more diagnostic than other cues.
Associational Factors
Associational factors, or attributes of the association between the event and the
brand, are also expected to impact inference making about the brand. Associational
factors include fit between the event and the brand, the duration of association between
the event and the brand, and the share of presence a particular brand has relative to all the
brands at the event. Brand/event fit was discussed at length in chapter five. Duration of
association refers to the event-brand associational history (number of years the brand
has sponsored the event). Share of presence is a measure of consumer perception of how
large a presence a brand has at an event relative to the other sponsors.
The fit between a brand and an event is expected to be a strong determinant of the
number and kinds of inferences a consumer makes about a sponsoring brand. As
previously discussed, fit can be directly measured as a unidimensional construct which is
perceived by consumers. Operationally, it has been defined in several ways by researchers

93
in other areas of commercial associations. In the domain of brand extensions, Aaker and
Keller (1993) operationally defined fit to be the degree to which the brand extension is a
substitute or a complement for the core brand. Broniarczyk and Alba (1994) seek out
brands with strong, unique associates. If the core brand and the extension possess the
same strong and unique association, they are defined to have relatively high fit. This
suggests that consumer perception of fit is high when the event and brand share unique,
salient attributes.
Many examples of the impact of fit on the success of endorsers have been found.
Friedman and Friedman (1976) found that fit between the type of product being endorsed
and the type of endorser significantly predicted ad evaluation, such that ads for jewelry
were best evaluated when done by a celebrity, whereas ads for cookies were better
received when done by a homemaker. Kamins and Gupta (1994) found that increased
congruence between the spokesperson and the product resulted in a more favorable
product attitude. Speck, Schumann and Thompson (1988) found that when the
endorsers relationship to the product was more obvious, information could be processed
in a shorter time.
Fit has been defined at even higher levels of abstraction, and has been found to
impact marketing success. Park, Milberg and Lawson (1991), for example, define high fit
as occurring when both the core brand and its extension are either prestige or
functional. They found that prestige extensions worked best for prestige core brands
and functional extensions worked best for functional brands.
While definitions of fit vary widely in the extant commercial association literature,
certain commonalties can be found. In most of the examples cited above, fit is defined as

94
the perceived similarity or congruence between two associates. This is what was defined
in chapter five as generic fit. While other, more complex conceptualizations of the fit
construct were provided in that chapter, the simplicity of perceived fit, and its frequent
usage in the extant commercial association literature make it an appropriate construct for
investigation in the emerging domain of sponsorship.
Duration of Association
Duration of association is also expected to impact inference making about a
sponsoring brand. Duration of association is defined as the length of time that the brand
has been associated with the event. A new or first-time sponsor represents extremely
low duration of association. For most events, duration of association is measured in years.
Consumers are more likely to be aware of long term associations. Since awareness
is a necessary condition for inference making, long term associations should lead to more
consumer inference making than short term associations. In addition, long term
associations may imply greater commitment on the part of the brand to the event. It is
expected, then, that long term associations should lead to greater consumer inference
making about the commitment of the company to the event than do short term
associations. It is also expected that long term associations will lead to more event
facilitation inferences than short term associations.
Share of Presence
The share of presence a brand has at an event relative to all other sponsors is also
expected to impact inference making. Share of presence is a measure of how much the

95
brand is perceived to be sponsoring the event. Things like the amount and prominence of
brand signage at the event and on promotional materials for the event are expected to
moderately impact perceived share of presence. Naming the event after the brand is
expected to have a strong, positive impact on perceived share of presence.
Sole sponsorship of an event would represent 100% share. As the number of co
sponsors at the same level of sponsorship goes up, an individual brands share of presence
at the event goes down. Like duration of association, share of presence is expected to
imply commitment of the brand to the event. Therefore, consumers are expected to make
more event facilitation inferences about brands with a high share of presence than about
brands with low share of presence.
Hypotheses About Specific Inferences
Review of the extant literature on sponsorship and other commercial associations
led to the conceptualization of three specific types of inference that might be generated
from an exposure to a sponsorship association and might impact brand choice. These
inferences are: brand size, brand legitimacy, and event facilitation. These three
types of inference, and their a priori relationship to event and associational factors are
described below. Other consumer inferences which emerged from the data are described
in the results section.
Brand Size
Inferences about brand size occur when a consumer changes (or forms) his/her
perception of the size of a firm or brand based on a sponsorship association. As

96
previously mentioned, there is some evidence from both the sponsorship literature
(Rajaretnam 1994), and other literatures (Kirmani 1990), that perceived expenditure can
influence perceived firm size. Within the sponsorship context, perceived brand/firm size
inferences are hypothesized to be driven by perceived financial ability to sponsor. The
idea is that the association cues some response that the brand must have substantial
resources to engage in such an association. It can therefore be predicted:
HI: Sponsoring larger events leads to more inferences on the part of consumers
that the brand is large.
H2: Longer duration of association between brand and event leads to more
in ferences on the part of consumers that the brand is large.
H3: Engaging in a sponsorship with a large, as opposed to small, share of
presence leads to more inferences on the part of consumers that the brand is
large.
Given that inferences about brand/company size are expected to be driven strictly by
perceived ability to engage in sponsorship, the congruence between brand and event
should not impact perceived size. Therefore:
H4: Brand/event fit does not moderate the relationships in H1-H3
Because of the prediction of no effect inH4, a liberal significance value is
necessary to test it credibly (i.e. p>. 15).

97
Legitimacy
Inference about the legitimacy of a company or brand occur when a consumer changes (or
forms) his/her perception of whether this brand is reasonable or of some minimum
quality because of its association with an event. Sponsorship associations are expected to
lead to such legitimacy inferences when a consumer perceives some gatekeeping function
on the part of the event, so as to not allow shoddy, or substandard products or companies
to be sponsors. Assuming large events have more to lose from an association with a
shoddy product:
H5: Sponsoring larger events leads to more brand legitimacy inferences on the
part of consumers
Event gatekeeping should be tighter in the long run than in the short run. As a
result, long term associations should imply greater endorsement of the brand by the event.
Therefore:
H6: Longer duration of association between brand and event leads to more brand
legitimacy inferences on the part of consumers
Events should be more scrutinizing gatekeepers for major, as opposed to minor sponsors.
Therefore:
H7: Engaging in a sponsorship with a large, as opposed to small, share of
presence leads to more brand legitimacy inferences on the part of consumers
Event gatekeeping should be higher for more related products than for less related
products. That is, an event should be more careful not to allow sponsorship by a shoddy
event-related product than by a shoddy, non-event-related product. For example, a tennis

98
tournament should be more concerned about sponsorship by a poor quality brand of tennis
balls than about sponsorship by a poor quality bank. Therefore:
H8: Brand/event fit moderates the relationships in H5-B7, such that high fit
amplifies these effects
Event Facilitation
Event facilitation inferences occur when a consumer comes to the conclusion that a
company or brand is helping to make possible, or improve the quality of an event. Given
that small events are more likely to be completely dependent on the goodwill of the
sponsor:
H9: Sponsoring smaller events leads to more event facilitation inferences on the
part of consumers
Since long term associations should imply more commitment on the part of the sponsor:
H10: Longer duration of association between brand and event leads to more event
facilitation inferences on the part of consumers
Since major sponsors should be viewed as doing more to insure the success of the event:
Hll: Engaging in a sponsorship with a large, as opposed to small, share of
presence leads to more event facilitation inferences on the part of consumers
Since sponsors may contribute technical knowledge and expertise to an event:
H12: Brand/event fit moderates the relationships in H9-H11, such that high fit
amplifies these effects

99
Study la
Overview
This study tests the impact of event size, brand/event fit, duration of association
and share of presence on inferences about a sponsoring brand. A completely between
subjects 2x2x2x2 factorial design exposed subjects to an unfamiliar brand in the context of
a poster for an event where the brand was a sponsor (see Appendix H). Unaided and cued
open ended inference measures were taken, followed by closed ended inference scale
measures.
Design
The study consisted of a 2(big/small event size) by 2(high/low brand-event fit) by
2(long/short duration of association) by 2(high/low share of presence) between subjects
design. This design is depicted graphically below.
Big Event
Small Event
high
share
low
share
high
share
low
share
Lowt
long
duration
Fit
short
duration
High
Fit
long
duration
short
duration
FIGURE 4
16 CELL DESIGN FOR STUDY 1A (N=89)

100
Subjects
Eighty-nine undergraduates in an introductory marketing class at a major
southeastern university took part. Each received course extra credit for participation.
Independent Variables:
Four independent variables were orthogonally manipulated. They were event size,
brand/event fit, duration of association and share of presence.
Event size
The event class chosen was a running event. Pretesting was undertaken to find
two events, which while largely unfamiliar (to make sure subjects would not be aware of
the actual sponsors), yet varied substantially in perceived size within the population to be
studied. Both actual and fictitious foreign running events were tested for perceived size
by twenty subjects. Results of this pretest are shown in Table 4 below. From this data, it
was determined that a clean manipulation of perceived size could be accomplished by
using two fictitious running events which ostensibly take place in Belgium. These two
events, The Euro-Capitol International Marathon (mean perceived size 5.35) and The
Leuven Student Fun Run (mean perceived size 2.50) differed significantly (F=43.14,
p<001). The use of fictitious, unfamiliar events had the additional advantage of allowing
maximum flexibility in manipulating the sponsorship association (see study lc).

101
TABLE 2
AVERAGE PRECE1VED SIZE OF RUNNING EVENTS* (N=20)
Runnina Event
Mean
St Dev
AUCKLAND MARATHON
2.90
1.59
BERLIN MARATHON
5.15
1.46
COEUR D' ALEE MARATHON
3.80
1.15
COLOGNE 10K
3.80
2.14
COMRADES MARATHON
3.45
1.90
DUBUN MARATHON
4.75
1.52
DUSSELDORF MINI-MARATHON
2.75
1.16
EURO-CAPITOL INTNL MARATHON
5.35
1.53
FIRENZE MARATHON
3.85
1.39
FLETCHER CHALLENGE MARATHON
3.95
1.57
HELSINKI CITY MARATHON
4.30
1.59
JUNGFRAU-MARATHON
2.90
1.07
LEUVEN STUDENT FUN-RUN
2.50
1.19
LIVERPOOL HALF MARATHON
3.20
1.06
LONDON MARATHON
6.20
0.95
LUXEMBORG 10K
4.00
1.89
MARATHON DES SABLES
2.05
1.05
NEWCASTLE 10K
3.45
1.54
PARIS MARATHON
6.10
1.25
REYKJAVIK MARATHON
3.55
1.47
ROMA CITY MARATHON
5.65
0.93
ROTTERDAM MARATHON
3.60
1.27
SHANNON CITY FUN-RUN
2.35
1.23
STOCKHOLM MARATHON
4.95
1.05
TILLBERG CITY 10K
2.95
1.36
VIGARANO MARATONA
3.30
1.53
* 1=VERY SMALL; 7=VERY LARGE
Brand/event fit
Fit was manipulated by selecting two brands which were equally unfamiliar to the
student population and varied substantially in degree of perceived fit with a running event.
Ideally, these would be from categories which actually engage in running event
sponsorship. Pretesting was performed on thirty-nine brand categories currently involved

102
in event sponsorship. Twenty subjects rated perceived fit between a running event and
each category. Means and standard deviations are provided in Table 3. The two brands
chosen are shown in italics. They are coffee (mean perceived fit=2.75), and sports drink
(mean perceived fit=6.9). This difference in perceived fit is statistically significant
(F=157.5, p<001). Coffee and sports drink are both in the beverage category, and both
categories have brands which sponsor running events.
Having selected the categories, it was necessary to select the actual brands.
Preference was given to actual foreign brands, provided they were very unfamiliar to the
subject population. To find such brands of coffee and sports drink, three European sports
drink brands and twelve European coffee brands were pretested for familiarity. These
brands were embedded with other European brands being pretested for an other study. In
all, twenty subjects rated 40 European brands for familiarity. Results of this pretest are
shown in Table 4. On a seven point scale, anchored by never heard of it and very
familiar, Rombouts Coffee had a mean rating of 1.05, and Extran Sports Drink had a
mean rating of 1.35. These differences are not statistically significant, even at a
conservative p=.10.

103
TABLE 3
PERCIEVED LEVEL OF FIT BETWEEN A RUNNING EVENT AND SELECTED
PRODUCT CATEGORIES
Mean
St Dev
Liquor Store
1.65
0.99
Oil Company
2.00
0.79
Mexican Food
2.20
0.89
Realtor
2.20
0.95
Fast Food Restaurant
2.25
1.12
Wine
2.25
1.02
Ice Cream Bar
2.60
1.05
Candy Bar
2.70
1.98
Software Company
2.70
0.98
Coffee
275
1.45
Brewery
2.95
2.24
Snack Food
3.00
1.78
Soft Drink
3.00
1.92
Steak house
3.00
1.08
Insurance Company
3.30
1.03
Music Store
3.45
1.43
Automobile Manufacturer
3.55
1.47
Drug Store
3.85
1.90
Airline
4.10
1.52
Newspaper
4.10
1.62
Cellular Phone Company
4.20
1.40
Hotel
4.40
1.57
Bank
4.45
1.15
Grocery store
4.55
1.54
Radio Station
4.70
1.42
Film
4.75
1.65
Hospital
4.95
1.70
Low Fat Frozen Entre
4.95
1.67
Pasta
5.00
2.00
Herbal Tea
5.25
1.33
Fruit juice
6.30
0.98
Health Care System
6.35
0.88
Mineral Water
6.55
0.76
Sporting Goods Store
6.70
0.47
Health Club
6.75
0.91
Performance Drink Concentrate
6.90
0.31
Sports Drink
6.90
0.31
Athletic Shoe
7.00
0.00
Energy Bar
7.00
0.00
* l=very low fit; 7=very high fit

104
TABLE 4
MEAN FAMILIARITY WITH VARIOUS EUROPEAN BRANDS
Mean*
St Dev
Max Ice Cream
1.05
0.22
Feast Ice Cream
1.05
0.22
Vescovi Caffe'
1.05
0.22
Rombouts Coffee
1.05
0.22
Jacobs Kaffee
1.05
0.22
Rader Candy Bar
1.05
0.22
Verkade Candy Bar
1.05
0.22
Dextro Energy Bar
1.10
0.31
Brao Caffe'
1.10
0.31
Douwe Egberts Koffie
1.10
0.31
Kanis & Gunnink Koffie
1.10
0.31
Simba Snacks
1.15
0.37
Coffee Rio Candy
1.15
0.37
Cup O Gold Candy
1.20
0.52
Callebaut Candy Bar
1.20
0.52
Droste Candy Bar
1.20
0.70
Extran Energy Bar
1.25
0.72
Torrefazione Italia (coffee)
1.25
0.72
Max Havelanr Koffie
1.25
0.72
Milka Candy Bar
1.25
0.55
Extran Sports Drink
1.35
0.81
Glycomax Sports Drink
1.35
0.75
Magnum Ice Cream
1.35
0.75
Leo Candy Bar
1.35
0.93
Dextro Energy Drink
1.40
1.35
Cometto Ice Cream
1.40
1.35
P-Nuttles Candy
1.50
1.19
agio Ice Cream
1.55
1.28
Whirty pop Candy
1.55
1.10
Zip Candy Bar
1.65
1.18
B Molino Coffee Roasters
1.70
1.38
Fairtime Taffy
1.70
1.17
Bella Caf
1.80
1.58
Solero Ice Cream
1.85
1.76
Baker Street Snacks
2.00
1.62
Calippo Ice Cream
2.05
2.52
Amanti Coffee
2.10
1.77
Lindt Candy Bar
2.15
2.18
B Presidente Coffee
2.50
1.88
Bounty Candy Bar
2.60
2.16
*l=never heard of it; 7=very familiar

105
Duration of association
The duration of association variable was manipulated directly. In the short
duration of association conditions, the target sponsor (Rombouts or Extran) was described
as Proud to be a New Sponsor. In the long duration of association condition, the target
sponsor was described as Proud to be a sponsor for 5 Years.
Share of presence
Share of presence was also directly manipulated. In the low share-of-presence
condition, the target sponsor was simply listed as one of three sponsors of the event. In
the high share-of-presence condition, the target (Extran or Rombouts) was labeled as a
Key Sponsor, whereas the other two sponsors (always Adidas and Generate Bank) were
listed as Associate Sponsors. Also, in the high share of presence condition, the target
sponsors logo was approximately twice as large as the associate sponsors logos. In the
low share-of-presence condition, the logo of the target brand was the same size as the
logos of the other two sponsors.
Experimental Procedure
Subjects entered the lab in groups of four to sixteen. On each desk was a packet
which contained one of sixteen posters which exposed the subject to one of sixteen
combinations of size, fit, duration and share. Each subject saw only one poster. Subjects
were instructed to open the packet to the page containing the poster. They were given 60
seconds to look at the poster as you would if you were to see it while waiting for a bus
or train.

106
Subjects were then instructed to turn to the next page which instructed them as to
how to complete the unaided thought listing task. The instructions read:
On the following page, we would like you to give us your thoughts about the
poster you just saw, by writing one thought in each box. Please list in these boxes
the thoughts that went through your mind as you looked at the poster. Please list
only one thought in a box. You might have positive, negative or neutral thoughts.
ALL ARE FINE.
You may ignore spelling, grammar and punctuation, and you may use phrases.
Simply write down the first thought that comes to your mind in the first box, the
second thought in the second box and so on. Dont worry if you dont fill in every
box. Just write down as many thoughts, feelings and reactions as occur to you.
Subjects then turned to the following page which contained seven thought listing
boxes. Each box was 1.8 high by 13 cm. long. All boxes were centered horizontally, and
placed one below the other on the page (see Appendix D). Subjects were given 90
seconds to list their thoughts. At the end of 90 seconds, they were instructed to finish
working on your current thought, but do not go on to an additional box. When all
subjects had recorded their last thought, all were instructed to turn to the following page,
which contained the instructions for the brand cued thought listing task.
The instructions for the brand cued thought listing task read :
Now, on the following page, we would like you to give us any thoughts you may
have had about the sponsor ROMBOUTS COFFEE (or EXTRAN SPORTS
DRINK). As before, please list in these boxes any thoughts that went through
your mind about ROMBOUTS COFFEE (or EXTRAN SPORTS DRINK)as you
looked at the poster. Please list only one thought in a box. You might have
positive, negative or neutral thoughts. ALL ARE FINE.
Again, you may ignore spelling, grammar and punctuation, and you may use
phrases. Simply write down the first thought that comes to your mind in the first
box, the second thought on the second box and so on. Dont worry if you dont
fill in every box. Just write down as many thoughts, feelings and reactions as
occur to you.

107
Immediately afterward subjects were instructed to turn to the following page
which contained seven thought listing boxes, identical to those on the first thought listing
page. Subjects were given 90 seconds to list their thoughts, this time, specifically about
the target brand. At the end of 90 seconds, they were instructed to finish working on
your current thought, but do not go onto an additional box. When all subjects had
finished their last thought, they were instructed to turn to the following pages which
contained the closed ended inference measures.
At their own pace, subjects evaluated the target brand as to their perception of its
size, legitimacy, quality and its facilitation of the event. Manipulation checks for event size
and brand/event fit were then taken. Finally, subjects were asked if they had ever: been to
Europe, been to Belgium, heard of Rombouts coffee (Extran Sports Drink), run in a
marathon, or run in a 1 OK race.
After completing these closed ended questions, subjects were instructed to turn
back to the thought listing pages, and to self code the thoughts they had listed in the
boxes. They were instructed to place a mark on the lines to the left of each thought box.
They were asked to place a + if the thought was positive, a if the thought was
negative, and am n if the thought was neutral. These self coded valence marks were to
used to help the coders understand what the subjects might have meant in ambiguous
situations. After completion of the self coding task, the subjects filled out paperwork
associated with the extra credit and were allowed to go.

108
Results
Manipulation Checks
Manipulation checks on perceived size and perceived brand/event fit were taken
after all other measures, except for demographic information. The check on the event size
manipulation was measured by asking subjects Overall, how large do you perceive the
running event advertised in the poster to be Responses were given on a single seven
point scale, anchored by a very small event and a very large event Mean response to
this scale were 5.13 for subjects in the large event condition and 4.32 for those in the small
event condition. This difference is significant (F=3.62, P<05), and in the appropriate
direction.
Perceived brand/event fit was measured by asking subjects Overall, how would
you rate the fit between the running event advertised in the poster and EXTRAN SPORTS
DRINK (ROMBOUTS COFFEE)? Responses were given along a seven point scale
anchored by very poor fit and very good fit. Mean response for the high fit groups
was 5.39, and for the low fit groups it was 3.32. This difference is statistically
significant(F=19.55, p<001), and in the expected direction.
Closed Ended Inference Measures
The three types of inferences which were predicted, a priori, to be influenced by
the event and associational factors, are brand size, brand legitimacy and event facilitation.
Because each of these dependent variables is associated with three main effects predictions
(impact of event size, duration of association and share of presence) as well as three
interaction prediction (fit will or will not moderate each), each dependent variable was

109
regressed on fit, size, duration and share, as well as the interactions of fit with size, fit with
duration and fit with share. Unless otherwise specified, the statistics reported for these
predictors on the dependent variables are based on the parameter estimates from a seven
degree of freedom, dummy coded regression model.
Impact on Perceived Size of the Brand
Two measures were taken of the perceived size of the brand. Absolute size was
measured by a single seven point scale anchored by one of the smallest sports drinks
(coffees) in Belgium and one of the largest sports drinks (coffees) in Belgium.
Perceived market share was measured by a single seven point scale which asked how large
they thought the brands market share was compared to other category brands in the
country. It was anchored by very small share and very large share. Analysis revealed
no significant impact of the independent variables on the absolute size measure.
Regressing on the perceived market share variable revealed two significant main
effects. First, the size of the event sponsored significantly predicted perceived market
share (F=5.29, p<03). As predicted, brands which sponsored large events were perceived
to have larger market shares. Thus, HI is supported.
The duration of association also predicted perceived market share (F=4.87, p<05).
As predicted, brands associated with an event for five years were perceived as having
larger market shares than were new sponsors. Therefore, H2 is supported as well. No
support, however, was found for H3. Share of presence at an event had no measurable
impact on perceived market share of the brand (F<1).

110
H4, which predicted no moderating effect of brand/event fit was not fully
supported. The moderating role of fit on the impact of duration on perceived brand share
(as measured by the interaction of fit and duration) was non-significant (F<1). This is as
predicted. It appears, however, that fit may have some moderating role in the impact of
event size on perceived market share. While the interaction of fit with event size did not
reach traditional levels of significance, it was close (F=3.52, p<07, 2=.02). While this is
not strong support for a moderating role of fit, neither is it strong support of H4 which
predicted no interaction. The effect size associated with this interaction falls within what
Cohen (1977) refers to as a small effect. The power of this experiment to detect an effect
of this size is only .37. In order to obtain a power of .80 to detect such an effect, the
sample size for this experiment would need to more than double from 89 to just over 200.
To further explore this interaction, the data was split into a high fit and a low fit
group. Event size was then regressed on perceived share for each of the two fit groups.
In the high fit group, event size had a significant impact on perceived market share
(F=8.13, p<01). In the low fit group, event size had no measurable impact on perceived
brand share. Caution is necessary in interpreting these results, however, remembering that
the original interaction term was not significant.
Brand Legitimacy
An attempt was made to measure subjects perception of the legitimacy of the
sponsoring brand using two seven point scales. The first asked How legitimate of a
sports drink (coffee) do you perceive Extran (Rombouts) to be?. The scale was
anchored by A fly-by-night brand and a legitimate brand. The second asked How

Ill
real of a brand do you perceive Extran (Rombouts) to be?. The scale for this question
was anchored by not very real and very real. None of the independent variables had
any significant impact on either of these two variables.
The simple logic that levels of gatekeeping by the event could translate into
legitimacy for a sponsoring brand made exploring legitimacy inferences, as opposed to
overall quality inferences, appealing Because it was anticipated, however, that the
concept of legitimacy would be difficult to convey to subjects in survey form, a direct
measure of perceived quality was also taken. This question asked what do you think the
quality level of Extran (Rombouts) is likely to be? The scale was anchored by very poor
quality and very high quality.
Regression of the independent variables on perceived quality yielded one
significant main effect. Event size significantly predicted perceived quality (F=5.5, p<03).
This was in the predicted direction. Brands associated with large events were perceived to
be of higher quality than were brands associated with small events. This supports H5.
Support for H8 was not found, however, in that fit did not moderate this effect. The
Interaction of event size with fit had no impact on perceived quality (F=l. 1, P>.25). No
support was found for H6 or H7. Neither duration nor share had any measurable influence
on perceived quality (F<1 for both).
Event Facilitation Inferences
Two measures of event facilitation inferences were taken. The first, called event
enabling, attempted to measure inferred importance of the brand to the event taking place

112
at all. The other, called event helping, attempted to measure inferred importance of the
brand to the event being a success.
Event enabling was measured by asking how important do you think the support
of EXTRN SPORTS DRINK (ROMBOUTS COFFEE) is to the sponsored event taking
place. The scale was anchored by not important at all and very important. The
independent variables had no significant impact on brand enabling.
Event helping was measured by asking How important do you think the support
of EXTRAN SPORTS DRINK (ROMBOUTS COFFEE) is to the success of the
sponsored event. The scale was anchored by not important at all and very important.
The independent variables were regressed on brand helping. Weak support was found for
H10. A long duration of association with an event seems to have some effect on event
helping inferences. This effect was not statistically significant at traditional levels (F=3.02,
p<09). Counter to prediction, this effect was not moderated by fit (F=1.22, p>25)
Therefore, HI2 is not supported. No support was found for H9 or HI 1. Neither event
size nor share of presence had any measurable impact on event facilitation inferences.
Open Ended Inference Measures (Thought Listings)
The thoughts in the thought listing boxes were coded by two independent raters
who were naive to both experimental condition and hypothesized relationships. Initially,
the thoughts were to be coded only for brand size, brand legitimacy and event facilitation
inferences. An iterative process of looking at subsets of the data and updating led to an
additional class of inference-making which was included in the final coding scheme. The
additional inference class included was termed commercial motive. Thoughts coded as

113
commercial motive inferences included negative inferences about the brand because of the
association, such as this is a waste of money or probably raise price to pay for event.
Each rater independently coded all of the thoughts. Initial agreement between
raters was quite high. Out of a total of 826 thoughts, the coders matched on 784 (94.9%
of the time). The remaining discrepancies were reconciled through discussion and mutual
agreement. Of the 89 subjects, 29 made a total of 46 inferences about the sponsoring
brand. The other 60 subjects listed no thoughts which could be considered inferences
about the brand.
The fact that about one third of subjects who saw a poster for an event listed an
inference about a sponsoring brand is encouraging. With 46 total inferences made,
however, it was unlikely that enough power would be available to test the impact of the
four independent variables on each of the four kinds of inferences. Given that number of
inferences is count data, a series of log-linear models were run to determine if any impact
of fit, size, duration, share and the three fit interactions could be detected on each of the
four inferences. As expected, no significant results were found.
In order to test whether any impact of the independent variables on inference
making could be detected, the data was re-coded to create a new variable total
inferences. This variable was a simple summation of the four individual types of
inference. Individual loglinear models tested the impact of fit, size duration and share on
the total number of inferences made. This analysis revealed a strong main effect of fit on
overall inference-making (Chi Sq.=7.02, p<01). The direction of this effect was that
subjects exposed to high fit associations made more inferences (Mean =.72) than did

114
subjects exposed to low fit associations(mean=.31) Some caution must be used, however,
before attempting to interpret this result (see study lb).
Discussion
Results of study la are quite encouraging. Support was found for HI (large event
leads to large brand inference), and H2 (long duration leads to large brand inferences).
The finding that aspects of a sponsorship association can significantly impact perceived
size of the brand is totally new in an experimental setting. If replicable, this finding should
be quite interesting for brand categories in which large size has a positive impact on choice
probability (see chapter 10).
H5 (large event leads to brand legitimacy(quality) inferences) was also supported.
Somewhat surprisingly, this effect was not moderated by brand/event fit. It made no
difference whether there was good or poor fit with the event. For both sports drink and
coffee, sponsoring a large running event led subjects to infer greater quality than
sponsoring a smaller one.
H10 (long duration leads to event facilitation inferences) received weak support.
It certainly makes sense that associating ones brand with an event for a long period of
time would lead people to infer greater facilitation on the part of the sponsor. Here again,
and counter to prediction, fit did not moderate the effect. For both sports drinks and
coffee, sponsoring a running event for a long time led to more facilitation inferences than
did being a new sponsor.
The open ended data yielded considerable evidence of spontaneous inference
making about a sponsoring brand. The number of each of the four specific inferences,

115
however, was not great enough to detect effects on specific effects. The open ended data
did reveal that fit between brand and event might facilitate overall inference making about
the brand. That is, sponsorship association with high fit (Extran) led to more inferences
about the brand than did the association with low fit (Rombouts).
Generalizability of these findings beyond a single type of event is, of course,
unwarranted. By choosing to manipulate fit by holding event constant and varying the
sponsor, fit is essentially confounded with brand. One way to increase the generalizability
of the findings would be to choose both another event and a set of two new brands. Here
again fit would be confounded with brand. Another approach is to find another class of
event which has a high fit with coffee, and a low fit with sports drinks. If the results
replicate in this case, both the overall generalizability, and the internal validity of the fit
construct, will be enhanced. Study lb, then, attempts to replicate the findings of study la
with an event for which coffee has a high fit and sports drinks have a low fit. Since no
evidence of the share of presence manipulation was found, this factor is eliminated from
the design of study lb. A stronger manipulation of share is attempted in study lc.
Study lb
Overview
This study attempts to replicate the findings of study la using a different event
category for which fit is reversed (see above). It again tests the impact of event size,
brand/event fit and duration of association on inferences about a sponsoring brand. A
completely between subjects 2x2x2 factorial design exposed subjects to an unfamiliar
brand in the context of a poster of a similar format to that used in study la. As before,

116
unaided and cued open ended inference measures were taken, followed by closed ended
inference scale measures, manipulation checks and demographic information.
Design
The study consisted of a 2(big/small event size) by 2(high/low brand-event fit) by
2(long/short duration of association) between subjects design. This design is depicted
graphically below:
Big Event
Small Event
Low
Fit
duration
short
duration
High
long
duration
Fit
short
duration
FIGURE 5
8 CELL DESIGN FOR STUDY IB (N=80)
Subjects
Eighty undergraduates in an introductory marketing class at a major southeastern
university took part. Each received course extra credit for participation.
Independent variables:
Three independent variables were orthogonally manipulated. They were
brand/event fit, event size, and duration of association.

117
Brandy event fit
Pretesting was conducted to find a category of event which had significantly higher
fit with coffee than with sports drinks. Brainstorming led to a list of potential events.
Fourteen subjects rated perceived fit between each product category and both coffee and
sports drink. Literature festivals met this criterion (F=71.7, P<001). Mean perceived fit
with coffee was 5.71, whereas mean perceived fit with sports drinks was 1.64.
Event size
Manipulation of event size was accomplished as similarly to that used in study la
as possible. Posters for two fictitious events were created. They were pretested for
perceived size by separate groups of subjects, drawn from the same population as the main
experiment. Eighteen subjects rated their perception of the size of The European Union
International Literature Festival and Poetry Competition, along a seven point scale
anchored by a very small event and a very large event (mean=5.1) Eighteen others
rated the Leuven City Library Literature Festival and Poetry Competition along the
same scale (mean=4.6). The difference between these two was not statistically significant.
In order to increase the strength of the manipulation, additional information about
the events was included in the posters. The poster for the small event listed the speakers as
Wilfried Vanhonacker(Head Librarian), Luk Warlop (Poetry Club President) and
Christophe Van Den Bulte (Literature Professor). The poster for the large event listed
Her Royal Highness Queen Fabiola, Kurt Vonnegut, Hugo Claus (Poet laureate of
Belgium) and Seamus Heaney (Winner of the 1995 Nobel Prize in Literature).

118
Thirty-six additional subjects were exposed to one of the two revised posters and
rated their perception of event size as described above. With the strengthened
manipulation The European Union International Literature Festival and Poetry
Competition had a mean rating of 5.7 whereas the Leuven City Library Literature
Festival and Poetry Competition had a mean rating of 3.2. This difference is significant
(F=12.78, p<01). These two posters were then used for the main experiment.
Duration of association
Duration of association was manipulated as described previously.
Brand logo size and placement
Brand logos were the same size as in the low share-of-presence conditions in study
la. Due to the nature of the event, one of the filler sponsors(Adidas) was replaced by a
more logical associate (Coolkat Books).
Results
Manipulation Checks
As before, manipulation checks on perceived size and perceived brand/event fit
were taken after all other measures, except for demographic information. Using the same
scales as before, the difference between large size (mean=5.35) and small size
(mean=3.42) was significant (F=29.5, p<001). The difference between high fit (5.25) and
low fit (mean=2.55) was also significant (F=59, pc.001).

119
Closed Ended Inference Measures
Impact on perceived size of the brand
As in study la, the size of the event had a measurable impact on one of the
perceived size variables Event size significantly predicted absolute perceived size
(F=4.74, p< 05). Once again, brands associated with large events were perceived to be
bigger than brands associated with small events. As in study la and in contradiction to
original prediction, brand/event fit appeared to have a weak, moderating role (F=2.72,
p< 11, (Di=.02). The power of this experiment to detect this small effect was 0.18. In
order to obtain a power level of .80 to detect this effect, sample size would need to
increase from 80 to 224. Unlike in study la, duration of association had no impact on
perceived brand size.
Impact on brand legitimacy and event facilitation
Unlike in study la, event size had no measurable impact on perceived quality or
legitimacy. Duration of association also had no impact on brand legitimacy inferences.
The impact of duration of association on event facilitation did not replicate. Event size
also had no measurable impact on event facilitation inferences.
Open Ended Measures
Once again two raters independently coded all of the thoughts. Initial agreement
between raters was again quite high. Out of a total of 728 thoughts, the coders matched
on 701 (96.3%). The remaining discrepancies were reconciled through discussion and

120
mutual agreement. Of the 80 subjects, 20 made a total of 26 inferences about the
sponsoring brand. The other 60 subjects listed no thoughts which could be considered
inferences about the brand.
Once again the small number of overall inferences made it unlikely to be able to
detect effects of event size and duration of association on specific inferences. As
expected, none was detected. As before, a series of loglinear models were used to look
for effects of the predictors on total inferences.
Just as in study la, fit seemed to have an impact on overall inference making (Chi
Sq=3.64, p<06). The low fit group made an average of .45 inferences, whereas subjects
in the high fit group made an average of .20 inferences, meaning the effect of fit found in
la had reversed! Therefore, it should be immediately obvious that the increased inference
making in la was not necessarily the result of fit, but some other aspect of Extran Sports
Drink.
Discussion
Study lb provided a good bit of additional insight into the consumer inference
making process, and how it can be influenced by event and associational factors. By
replicating the significant impact of event size on perceived brand size, greater confidence
in the robustness of this effect is obtained. As mentioned previously, such an effect should
be of great interest for categories in which perceived brand size strongly impacts choice.
Secondly, and of considerable importance, is the complete reversal of the impact of
fit on the total amount of inference making. This relationship is depicted below in Figure
6.

121
Low Fit
High Fit
Study la
Running
Rombouts <
£ Extran
Event
Coffee
Sports Drink
Study lb
Extran
> Rombouts
Literature
Sports Drink
Coffee
FIGURE 6
DIFFERENCES IN TOTAL INFERENCE MAKING BETWEEN HIGH AND LOW FIT
GROUPS FOR STUDIES 1A AND IB
Study la seems to support a positive impact of fit on inference making. Study lb
seems to support a negative impact of fit on inference making. The parsimonious
explanation for this pattern of data is that some aspect of Extran Sports drink, and not
brand/event fit, is driving the differences in inference making. Without the second study,
researchers in the domain of sponsorship might have begun to look for boundary
conditions (i.e. when fit facilitates inferences and when it does not). Fortunately, the
intentionally conservative design employed in these first two studies has eliminated any
chance of following such a false path.
There are any number of reasons why one might expect more inferences to be
made about a sports drink than a coffee. Perhaps there are more attributes to think about.
Perhaps encountering an unfamiliar sports drink is more unusual than encountering an
unfamiliar coffee. Whatever the reason, it seems clear that the particular brand, not fit
with the event, is not driving the number of inferences made.

122
Thus far, a good deal has been learned about fit, duration and event size. Study
la, however, found no measurable impact of share of presence. Since the manipulation of
share in study la represented a somewhat limited range of the share of presence construct,
study lc explores the impact of a much more intrusive (and managerially pertinent) level
of share of presence: title sponsorship, or putting the brand name into the name of the
event.
Study lc
Overview
This study attempts to extend study la by putting the brand name into the name of
the event during exposure. A completely between subjects 2x2 factorial design exposed
subjects to an the same unfamiliar brands as in study la. Once again, unaided and cued
open ended inference measures were taken, followed by closed ended inference scale
measures.
Subjects
Sixty-two undergraduates in an introductory marketing class at a major
southeastern university took part. Each received course extra credit for participation.
Design
The study consisted of a 2(big/small event size) by 2(high/low brand-event fit)
between subjects design. This design is depicted graphically below:

123
Big Event
Small Event
Low
Fit
High
Fit
FIGURE 7
4 CELL DESIGN FOR STUDY 1C
Independent Variables
Two independent variables, event size and brand/event fit, were orthogonally
manipulated. Fit was manipulated exactly as in study la Associating a running event with
a coffee represented low fit and associating a running event with a sports drink
represented high fit. The small size event was again a student ftm run, but this time it
was either the Rombouts student fim-run or the Extran student fun-run The large size
events were the Rombouts Euro-Capital International Marathon or the Extran Euro-
Capital International Marathon.
Results
Manipulation Checks
Using the same scales as before, the high fit mean was 5.56 and the low fit mean
was 3.13 on the fit check measure. This difference is significant (F=43, p<001). The

124
large event size mean was 5.2, and the small event size mean was 3.7 on the size check
measure. This difference is statistically significant as well (F=13.43, p<001).
Closed Ended Measures
Once again, sponsoring a large event impacted the perceived absolute size of the
brand (F=3.05, p<086). As before, associating a brand with a large event led subjects to
infer that the brand is large. This effect may again have been moderated by fit (F=2.8,
p< 10, 2= 02). The power of the experiment to detect this effect was 0.47. To obtain
power of .80, sample size would need to increase from 62 to 172. Event size once again
also had a strong positive impact on perceived quality (F=12.09, p<001) While this effect
was found in study la, the impact appears to be much stronger under title sponsorship
(this possibility is tested below). No other significant effects were detected.
To test whether title sponsorship impacted inference-making, the two data sets
from la and lc were combined. This allowed a comparison of the same brands associated
with the same events. The only difference was whether or not the target sponsor was a
title sponsor.
This analysis revealed that title sponsorship can aid event facilitation inferences.
The difference in response to the event enabling measure was significantly higher in study
lc than in study la (F=5.54, p<03). This effect was not moderated by event size (F<1).
Therefore, it appears that being a title sponsor leads to more event facilitation inferences
regardless of whether the event sponsored is large or small. No other significant effects of
title sponsorship on inference making were found.

125
Since the effect of event size on perceived quality appeared to be bigger under title
sponsorship, the interaction of title sponsorship with event size was tested to see if this
difference was significant. A model which regressed size, title sponsorship and their
interaction on quality was run. In this combined model, event size still has a significant
impact (F=5.54, p<02), but neither the main effect of title sponsorship (F=1.86, p>. 15)
nor its interaction with event size (F<1) was significant. Therefore no statistical support is
found for an amplification of the impact of event size on perceived quality under title
sponsorship.
Open Ended Measures
As in la and lb, two raters independently coded all of the thoughts. Initial
agreement between raters was high. Out of a total of 561 thoughts, the coders matched
on 519 (92.5%). Discrepancies were again reconciled through discussion and mutual
agreement. Of the 62 subjects, 24 made a total of 43 inferences about the sponsoring
brand. The other 38 subjects listed no thoughts which could be considered inferences
about the brand.
Once again the small number of overall inferences made it unlikely to be able to
detect effects of event size and duration of association on specific inferences. As
expected, none was detected. As before a series of loglinear models were used to look for
effects of the predictors on total inferences. No impact of either fit or size was detected
on total inference making.

126
Discussion
This study further supports the impact of sponsoring large events on perceived size
of the brand. It also replicates the significant impact of event size on perceived quality.
While this effect initially looked to be bigger under title sponsorship, the formal statistical
test did not support this conclusion. Nonetheless, these replications do add support to the
previous finding that the size of a sponsored event can influence important inferences
about the sponsor.
Comparing the data in this study with that in la also revealed a strong impact of
title sponsorship on event facilitation inferences. Being a title sponsor led to a much
greater perception that the event was being made possible by the sponsor. Given the
costs associated with title sponsorship, this should be a most interesting finding to those
engaged in tactical sponsorship decision making (i.e. being the title sponsor at one event
vs. being an associate sponsor at many).
Once again, while over one third of the subjects generated spontaneous inferences
about the brand, the inadequate number of inferences prevented the detection of any
effects of the independent variables on specific inferences. One way to further explore the
data with greater statistical power was to combine the open ended data from the three
studies. Since subjects did not have equal probability of inclusion in each of the three
studies, results of this analysis are not based on a true experiment. The potential for
additional insight into the spontaneous inference-making process, however, makes such
an analysis worthy of consideration.

127
Additional Cross-Study Analysis
Because event size and fit were manipulated across all three experiments, effects of
these two variables are tested on the pooled data from all three studies. Impact of
duration, which was manipulated in la and lb, but held constant in lc, is tested on pooled
data from studies la and lb only. Since share was manipulated only in la, no additional
analysis is possible. The impact of title sponsorship is also tested further. Since title
sponsorship was explored only in the context of a running event, it is confounded with
event type. Therefore, title sponsorship was tested on the pooled data from la and lc (the
two running event studies) only.
Total inferences and the four classes of spontaneous inferences were regressed on
fit and size. This was done using the pooled data from all three studies. No additional
significant effects were detected.
Total inferences and the four classes of spontaneous inferences were regressed on
duration of association using the pooled data from la and lb. One significant effect was
found. Here duration of association significantly predicted the number of brand size
inferences made (Chi Sq.=3.9, p<05). Therefore, the support for H2 found with the
closed ended data in study one is replicated with the spontaneous inference data.
Finally, the four classes of inference, as well as total inference making, were
regressed on title sponsorship using pooled data from la and lc. This analysis also
revealed one significant effect. Title sponsorship significantly predicted the number of
commercial association inferences made (Chi Sq.=4.3, p<05). It appears, therefore, that
while naming an event after a brand has positive effects (facilitation inferences) it also may
have a downside (more inferences about the commercial intent of the sponsor).

128
The additional across-studies analysis further supported one effect and uncovered
a new one. Replicating the positive impart of duration of association on making the
inference that the brand is large is important. Finding such confirmation from data
gathered before any mention had been made to the subjects about the size of the brand is
indeed exciting. It not only replicates a previous finding, it demonstrates that attributes of
a sponsorship associations can influence inference making about brand size without ever
mentioning the size of the brand to the subjects. Subjects in the long duration of
association condition made inferences about brand size at more than three times the rate of
those in the short duration-of -association condition.
The new finding from this analysis is also quite interesting. The fart that being a
title sponsor may also have a downside is an intriguing possibility which has implications
for tactical sponsorship decisions. For example, if a brand is attempting to engage
reciprocity through a sponsorship which it hopes will lead to event facilitation inferences,
it may want to be a title sponsor. Being a title sponsor, however, might also lead to more
negative inferences, such as they are just trying to sell me something. When, and for
whom, each of these inferences is more likely is certainly fertile ground for further
research
Overall Discussion and Limitations of Studies la-lc.
The series of three studies presented in this chapter have yielded quite a bit of
information about the impart of event and associational factors on consumer inference
making about sponsoring brands. A summary of findings from these three studies is
presented below in Table 5.

129
TABLE 5
SUMMARY OF SUPPORT FOR THE MAIN EFFECTS HYPOTHESES IN STUDIES
1A, IB, AND 1C
Study
la
Study
lb
Study
lc
open
ended
Any
Support
HI Large event leads to large brand inference
**
**
*
YES
H2 Long duration leads to large brand inference
**

**
YES
H3 Large share leads to large brand inference

NO
H5 Large event leads to legitimate brand
inference
**
***
YES
H6 Long duration leads to legitimate brand
inference
NO
H7 Large share leads to legitimate brand
inference

NO
H9 Small event leads to event facilitation
inference
NO
H10 Long duration leads to event facilitation
inference
*

YES
HI 1 Large share leads to event facilitation
inference

**
YES
*=p< 10 **=p< 05 ***=p<01 = Not Tested
As the table makes clear, the hypothesis receiving the most robust support is HI.
In all three studies, evidence was found that associating a brand with a larger event leads
people to believe that it is a larger brand. The other hypothesis to receive support in
multiple studies is H5. Associating a brand with a larger event appears also to improve
perceptions of quality. These results seem to indicate that event size does matter. Being
associated with large events can lead to at least two positive inferences about the brand.
Duration of association also was found to impact inference making. Using pooled,
open ended data, a main effect of duration on perceived brand size was found. This result
is especially interesting because it was obtained without cueing subjects about any brand
attribute. The inferences were spontaneously generated. The impact of duration on

130
perceived size was also obtained from the close-ended data in study la. This lends
additional support for the robustness of the effect.
In the comparison of Study lc to la, title sponsorship was found to impact event
facilitation inferences. The same analysis revealed, however, that being a title sponsor can
also lead to a greater number of negative, commercial motive inferences. As mentioned
previously, an important issue for future research will be to explore when consumers make
each type of inference.
This series of studies has examined how event and associational factors impact
consumer inference making about brands. The first stage of the inference based model
(see Figure 3) also includes the potential for audience factors to impact the inference
making process. This issue is addressed in the next chapter.

CHAPTER NINE
THE IMPACT OF AUDIENCE FACTORS AND FIT ON BRAND INFERENCES
To explore only the effects of the event and the association on consumer inference
making would ignore one crucial term in the equation: the consumer. This chapter looks
at how audience involvement with an event impacts the kinds of inferences that are drawn
from an association. Specifically, it examines how the intrinsic personal relevance of the
event, felt involvement with the event, and event domain knowledge impact inference
making about a sponsor of the event.
The first set of studies (la, lb and lc) explored inference making about unfamiliar
brands associated with unfamiliar events. While experimental realism (i.e. task
involvement) was high, subject involvement with the events and the brands was quite low.
The brands used were unfamiliar to the subjects and the events used were fictitious. For
these reasons, it would be quite difficult to examine the role of audience involvement with
the event using one of the events from the previous studies.
When studying the impact of audience involvement on inference making, two
possible approaches are available. One option is to attempt to manipulate involvement
with a real or an imaginary event. The other option is to select a real event, and measure
subjects extant involvement and knowledge. Because exploring how sponsorship works
is a largely substantive endeavor, external validity is a key consideration. For this reason
131

132
option two, measuring involvement with an extant activity, was chosen as the method of
choice.
Well known, popular events were chosen so as to ensure that some subjects would
exhibit measurable levels of involvement with the events. Because real events are used, it
was thought that associating fictitious or unfamiliar brands with them would lead to
unacceptably low levels of mundane realism. Rather, real brands were chosen for
association with the event. In the key study (2a), the brands chosen are actual sponsors of
the event.
Because real brands are used, it is unlikely that a single experimental exposure will
influence perceived brand size, legitimacy or quality. Event facilitation and commercial
motive inferences, however, might still be influenced by associational variables. It is these
inferences, therefore, that are the focus of this pair of studies.
Extant commercial association research, such as brand extension and celebrity
endorsement has not addressed the issue of involvement with the provider. This makes
the current study of how involvement with an event impacts inferences about a sponsoring
brand largely exploratory. For this reason, no formal hypotheses are provided.
This set of studies looked at the impact of audience involvement on inference
making. Study 2a is the main study. Study 2b is used to check for a possible confound of
fit with brand. The manipulation of fit in study 2a is reversed in study 2b (as was done in
la and lb). This allows any impact of fit in the key study (2a) to be checked in 2b, to
make sure the effect is the result of fit, and not some other aspect of the two sponsoring
brands. The flipped fit is depicted in figure 8 below.

133
Low Fit
High Fit
Study 2a
U.S Open
(Tennis)
Texaco
Evian
Study 2b
Indy 500
Evian
Texaco
FIGURE 8
DESIGN FOR STUDIES 2A AND 2B
Study 2a
Overview
Subjects are exposed to an actual high or low fit sponsorship association by
viewing a poster for the US Open Tennis Tournament. The poster featured either a high
fit or a low fit sponsoring brand. They then responded to open and closed ended measures
of inference making about the sponsoring brand (similar to studies la-lc). They then rate
their involvement with, and subjective knowledge of tennis. Finally, they are quizzed on
their actual knowledge of the rules of tennis.
Subjects
Thirty-eight undergraduates in an introductory marketing class at a major
southeastern university took part. Each received course extra credit for participation.

134
Design
The simple design consists of a single, two level variable (high/ low brand-event
fit), which is manipulated between subjects. Covariates relating to involvement with and
knowledge about tennis are included for a planned analysis using ANCOVA.
Independent Variable
The manipulated independent variable is brand/event fit. The event chosen was the
US Open tennis tournament. A list of the 19 actual sponsors of the US Open were
pretested for perceived fit with the event. Forty-six subjects rated the fit between these
brands and the US Open (see Table 8). The high fit brand selected was Evian. Along the
same seven point scale used previously, Evian had a mean fit rating of 6.8. The low fit
brand chosen was Texaco, which had a mean fit rating of 2.8. This difference in perceived
fit is significant (F=329, p<0001).
Several audience factors, or attributes of the relationship between the audience and
the event, were measured. The intrinsic personal relevance of the event, and felt
involvement with tennis were measured. Domain knowledge about tennis (both objective
and subjective) was then assessed.
Felt involvement is a point-in-time measure of a consumers involvement with an
event. It includes both enduring and situational components, and represents an
instantaneous measure of the level of meaning of an event in a persons life. It was
measured immediately after exposure to the poster using two seven point scales anchored
by strongly agree and strongly disagree Item one was The message I just saw was

135
important to me. Item two was The poster didnt have anything to do with me or my
needs.
TABLE 8
PERCEIVED FIT BETWEEN THE US OPEN AND 19 ACTUAL SPONSORS
(N=46)
Brand
mean
stdev
Tyco
2.46
1.38
Texaco
2.80
1.42
Mass Mutual
3.02
1.53
Tiffany
3.13
1.92
Heineken
3.57
2.08
Prudential
3.59
1.94
Chase
3.80
1.64
NY Times
4.11
1.72
Infiniti
4.33
1.94
IBM
4.41
1.94
Citizen
4.87
1.73
AmEx
5.35
1.59
Pepsi
5.80
1.24
Canon
5.91
1.26
Fuji
5.98
1.58
Fila
6.57
0.93
Evian
6.80
0.45
Tennis Mag
6.91
0.35
Wilson
6.93
0.33
Subjects then performed a set of thought listing tasks identical to those in la-lc.
Following this, they filled in the closed ended inference measures. Because event
facilitation inferences were key to this study, a three item scale was developed to measure
perceived event facilitation. All three items were anchored by not important at all and
very important. The three items were how important do you think the support of
Evian (Texaco) is to the US Open Taking Place, how important do you think the
support of Evian (Texaco) is to the success of the US Open, how important do you
think the support of Evian (Texaco) is to the quality of the US Open. The Cronbach

136
alpha for these three items was over .85, so the three items were combined into a single
measure of perceived event facilitation.
Because commercial motive inferences emerged in previous open ended data, a
closed ended measure of commercial motive inference-making was added. A seven point
scale measured agreement with the statement Evian (Texaco) is just trying to get a lot of
people to see their name. This scale was anchored by strongly disagree and strongly
agree. A manipulation check was then taken for fit between the brand and the US Open,
as was a measure of the perceived size of the event.
Subjects then rated their involvement level with tennis using a twenty item form of
the Personal Involvement Inventory, or PII (Zaichkowsky, 1985). Instructions for
completion of the PH were as follows:
Now, we would like you to please take a moment (approximately 5-10 seconds) to
think about tennis and what it means to you. Then complete these scales with
regards to tennis.
The scales were anchored by items like important/unimportant, trivial/fundamental,
and baring/fascinating.
Subjects then rated their subjective knowledge about tennis. They expressed their
level of agreement with the statement I know a great deal about the game of tennis. The
scale was anchored by strongly agree and strongly disagree.
Subjects then completed the intrinsic source of personal relevance, or ISPR scale
(Celsi, Chow, Olson, and Walker, 1992), adapted to tennis. Agreement was measured
with five items along a seven point strongly agree/strongly disagree scale. The five
items were playing tennis lets people see me as I would ideally like to be seen, tennis
helps me attain the type of life I strive for, I can make connections or associations

137
between tennis and experiences in my life., tennis is of high personal relevance to me,
and tennis helps me express who I am
Subjects then listed the number of times they had played and watched tennis in the
last week, month, and six months. Finally, subjects took a 12 question quiz about the
rules of playing and scoring tennis. This represented the objective knowledge measure.
Results
Manipulation Check
On the same scale used in previous studies, the high fit brand (mean=5.65) was
rated as having better fit with the event than the low fit brand (mean=2.89) This
difference is statistically significant (F=49.69, p<0001)
Close Ended Data
The overall analysis strategy was to determine if fit impacted inference making,
and if so, if any of the audience factors moderated the relationship. Fit had no measurable
impact on inferences about the commercial motive of the sponsor (F=1.82, p>. 15). It did
however, have a significant, positive impact on event facilitation inferences (F=6.64,
p<02). The high fit brand (Evian) was perceived as doing more to help the event than
was the low fit brand (Texaco).
A series of ANCOVAs were then run to test which, if any, of the audience factors
might moderate this relationship. These models revealed no moderation by frequency of
watching, frequency of playing, ISPR, PH, or felt involvement. Significant moderation
was found, however, for both subjective (F=4.15, p<05), and objective (F=12.89,

138
p< 001) domain knowledge (as measured by the interaction term in a model containing fit,
the knowledge variable, and the interaction of fit and the knowledge variable).
To see the nature of the fit by subjective knowledge interaction, the data was split
at the mean of subjective knowledge. Separate follow up analysis then tested the effect of
fit on event: facilitation for each of the two subjective knowledge groups. This analysis
revealed that for the high subjective knowledge subjects, fit had no impact on event
facilitation inferences (F<1). For low subjective knowledge subjects, however, Fit had a
strong, positive impact on event facilitation inferences (F=l 1.99, p< 01)
In the same way, the data was split at the mean of objective knowledge. Here
again, for high knowledge subjects, fit had no impact on event facilitation inferences
(F<1). For the low knowledge subjects, fit had a strong and significant effect on
perceived event facilitation (F=25.66, p<0001).
Open Ended Bata
Open ended inference measures were coded as before. Overall initial agreement
between the two raters was quite high. Out of 316 total thoughts, they agreed on 293
(92.7%). Discrepancies were resolved as before. Inferences about the sponsor were made
by 12 out of 38 subjects. A total of 12 brand inferences were made. No impact of fit or
any audience factors was detected on any of the four specific inferences. Neither was any
impact of fit or the audience factors detected on total inference making. Given the small
number of inferences, however, the lack of significant results may be due to low power.
It appears that the main effect of fit on event facilitation inferences was being
driven by subjects with low knowledge levels about tennis. Before discussing the

139
implications of this finding, however, it is necessary to make sure that event facilitation is
being driven by fit, and not some other aspect of Evian and Texaco. To this end, study 2b
checks to see if the main effect result replicates when Texaco is high fit and Evian is low
fit.
Study 2b
Overview
This study replicated the method in 2a by using an event with high perceived fit
with Texaco and low perceived fit with Evian Open and closed ended measures of event
facilitation and commercial motive were taken. Since the event which best flipped the
fit was exclusively a spectator sport, it did not make sense to measure the audience
factors used in 2a. This study simply checks for the confound of fit with brand in fits
impact in inference making.
Subjects
Forty-one undergraduates in an introductory marketing class at a major
southeastern university took part. Each received course extra credit for participation.
Design
The design is identical to 2a, except the event was changed so as to make Evian
the high fit event and Texaco the low fit event.

140
Independent Variable
In order to find an event for which Evian had high fit and Texaco low fit, a pretest
was conducted in which 13 subjects were asked to rate perceived fit between both brands
and several large events. The Indianapolis 500 was chosen. This event had a mean
perceived fit with Evian of 5.07, and a mean perceived fit with Texaco of 6.38. These
differences are statistically significant (F=7.41, P<02)
Results
Manipulation Check
Using the same manipulation check scales as in previous studies, mean perceived
fit was 5.9 for the high fit brand and 2.62 for the low fit brand. These differences are
statistically significant (F=56.03, p<0001).
Closed End Inference Measures
Mean perceived event facilitation, as measured by the summation of the three,
seven point scales, is 7.52 for the low fit group and 10.45 for the high fit group. This
difference is significant (F=5.28, p<.03), and in the same direction as in 2a. This supports
the conclusion that fit, and not extraneous brand factors, had amplified event facilitation
inferences. No impact of fit on commercial motive inferences were found (F<1).
Open Ended Inference Measures
Open ended inference measures were coded as before. Overall initial agreement
between raters was reasonably high. Out of 276 total thoughts, they agreed on 232

141
(91.3%). Discrepancies were resolved as before. Inferences about the sponsor were made
by 14 out of 41 subjects. A total if 19 brand inferences were made. No impact of fit or
any audience factors was detected on any of the four specific inferences. Neither was any
impact of fit or the audience factors detected on total inference making.
Replicating the impact of brand/event fit on event facilitation inferences lends
credibility to the assumption that this effect is truly being driven by fit, and not anomalous
factors associated with the two brands. Instead of revealing a confound as was the case in
the comparison between la and lb, the possibility that this effect is being driven by brand
and not fit is strongly refuted.
Unfortunately the spectator-only nature of Indy Car racing made replicatory
measures of many audience factors seem unrealistic. Asking subjects, for example, their
level of agreement with Indy Car racing helps me attain the type of life I strive for
seemed unreasonable. In hindsight, it is clear that measures of subjective, and objective
knowledge of Indy Car racing could have been taken. While the objective measures
would have been somewhat difficult to calibrate, the objective measure alone would have
been valuable. This shortcoming, however, should not diminish the interesting finding of
study 2a.
Overall Discussion
The purpose of this chapter was to explore the role of audience factors on
inference making about sponsoring brands, and to examine the inference making process
which occurs when familiar brands are associated with large, well known events. Two
main findings were discovered. First, evidence that brand/event fit enhances event
facilitation inference making was found in 2a and replicated in 2b. The fact that this effect

142
replicated when fit was flipped provides strong evidence that fit, and not some other
attribute of the brand is driving the effect.
This effect was not detected in study 1. It is possible that the effect was somehow
enhanced through the use of real brands and events. Perhaps the fact that subjects already
had beliefs about the size and quality of Evian and Texaco allowed more opportunity for
event facilitation inferences to be created. Perhaps the lack of other sponsors on the
poster increased event facilitation inferences in a way which allowed the impact of fit to be
detected. A reasonably parsimonious explanation for not detecting this effect in previous
studies, however, is that there are some anomalous attributes of Rombouts and Extran
which made them unique.
While it appears that fit between brand and event can lead to event facilitation
inferences, it is not clear when this would be expected. Future research should attempt to
find boundary conditions on this effect. Possible studies could examine the role of
familiarity of the brand, and the familiarity of the event, as well as the number of other
sponsors as possible moderators of this effect.
The second major finding of this set of studies is that for the US Open, knowledge
about tennis diminished the impact of fit on event facilitation inferences. That is, groups
of subjects high in tennis knowledge showed no significant difference in event facilitation
inferences for Evian vs. Texaco. Subject groups low in tennis knowledge, however, made
much greater event facilitation inferences about the high fit sponsor (Evian) than the low
fit sponsor (Texaco).
One interesting possibility is that low knowledge subjects are unaware of the large
size of the event. If low knowledge subjects think the event is not as large, they may be

143
more likely to attribute a larger role to the supply of bottled water (Evian) to the success
of the event. That is, a small event is helped more by this form of event support than a
large one.
To test this, subjective and objective knowledge were regressed on the perceived
size of the event. The impact of subjective knowledge on perceived size was significant
(F=5.51. p<03). The impact of objective knowledge was nearly significant (F=3.64,
p<065). In both cases, as knowledge decreases, the perceived size of the event decreases
as well.
Therefore, it appears that the somewhat mysterious moderating role of knowledge
on facilitation inferences may have a possible explanation. In addition, this finding can be
seen as supporting H9, which states that as the perceived size of an event decreases,
perceived facilitation by a sponsoring brand increases. While indirect, this finding is of
some importance to the development of overall sponsorship strategy. This topic is
covered in depth in chapter eleven.
This chapter completes the initial exploration of the first stage of the inference
based model. A good deal has been learned about how event, associational and audience
factors impact inference making about a sponsoring brand. How these inference impact
choice is covered in the following chapter.

CHAPTER TEN
THE IMPACT OF SPONSORSHIP-BASED INFERENCES ON BRAND CHOICE
The preceding chapters have examined how and when sponsorship associations
lead to inference making about the brand. While interesting from a theoretical
perspective, this knowledge is managerially relevant only if these inferences impact brand
choice. The purpose of this chapter is, therefore, two-fold. The first goal is to
demonstrate that sponsorship-based inferences can impact brand choice. The second goal
is to demonstrate that this impact can vary from category to category and from brand to
brand.
All four types of sponsorship-based inferences could easily be investigated with the
method used in this chapter. In order to keep the number of attributes in the choice task
manageable for the subjects, however, the number of sponsorship-based inferences was
limited to two. Because the previous five studies have provided the strongest evidence of
sponsorships ability to impact brand size and event facilitation inferences, the impact of
these inference on choice is investigated in this series of studies.
In order to demonstrate differential impact of the inferences, it was necessary to
select two product classes which varied along some dimension. One such dimension
which might lead to differential impact of each inference is the ambiguity of product
144

145
experience (Hoch and Deighton, 1989). Experience ambiguity was defined as the degree
to which a consumer could know quality immediately after purchase.
Pretest
Forty-three products were pretested for experience ambiguity along a nine point
scale anchored by you can perfectly know quality immediately after trial and you can
never fully know quality. Thirty-three pretest subjects evaluated the 43 brands for
perceived experience ambiguity. These categories and the mean experience ambiguity
ratings are provided in Table 7. The high experience ambiguity product class chosen was
auto insurance (mean=4.0). The low experience ambiguity category chosen was yogurt
(mean=2.33). This difference is statistically significant (F=8.59, p<.01).
Because of the radically different attributes in the two product categories to be
examined, it is not recommended that these two categories be jointly modeled. Study 3a,
therefore, investigates the impact of perceived size and perceived event facilitation on auto
insurance choice. Study 3b investigates the impact of these variables on yogurt choice.
Discussion of these two studies is left until after the results section of study 3b.
Study 3a
Overview
Subjects made a series of eight choices between two brands of auto insurance (952
total choices). Each choice profile consisted of five brand attributes (including size and
event facilitation) for each of two brands. These attributes were manipulated orthogonally
through a fractional factorial design. Each subject completed half (8 cells)of this 16 cell

146
design. While the fractional design employed does not allow testing of all higher order
interactions, it is extremely efficient. The full factorial design associated with this study
would have 65,536 cells!
Subjects
One hundred nineteen undergraduates in an introductory marketing class at a
major southeastern university took part. Each received course extra credit for
participation.
Independent Variables
Five independent attribute variables were manipulated orthogonally. They were
price (four levels), deductible (two levels), company size (two levels), event facilitation
(two levels), and hours of customer service (two levels). Each is described below (see
Appendix E).
Price level was determined by measuring perceived auto insurance cost. Pretesting
was used to find a range of perceived prices for six months auto insurance coverage.
Twenty-two subjects estimated cost for six months insurance coverage. The mean
estimate was $570.45, and the standard deviation was $89.52. The four manipulated price
levels were set at the mean minus one half standard deviation ($525), the mean ($570), the
mean plus half a standard deviation ($615) and the mean plus one standard deviation
($660). The amount of the deductible was manipulated at two levels. The low deductible
level was set at $500. The high deductible level was set at $1000.

147
TABLE 7
PERCIEVED EXPERIENCE AMBIGUITY OF SELECTED PRODUCT
CATEGORIES
Product Category mean stdev
razors 2.03 1.79
icecream 2.09 2.08
frozen dinner 2.18 2.19
beer 2.33 1.98
deodorant 2.33 1.90
yogurt 2.33 2.26
athletic shoes 2.52 2.08
encyclopedia 2.64 2.36
carwax 2.70 1.90
cellular phone 2.70 2.24
film 2.70 1.94
shampoo 2.79 1.96
airline 2.88 1.98
hand tools 2.91 2.08
paint 2.94 1.90
cielingfan 3.00 2.19
exercise equip. 3.03 2.32
personal watercraft 3.03 1.85
DVD 3.06 2.01
computer 3.30 2.31
charcoal 3.42 2.60
car 3.48 2.44
Internet service 3.48 2.02
long distance 3.52 1.99
music club 3.52 2.15
credit card 3.58 2.25
condom 3.61 2.41
cable 3.70 1.91
cell phone service 3.73 2.23
bank 3.82 2.04
battery 3.91 2.31
medical insurance 3.97 2.39
car insurance 4.00 2.36
motor oil 4.24 2.19
gasoline 4.30 2.19
investment company 4.30 1.83
dogfood 4.33 2.37
mutual find 4.36 2.29
renters insurance 4.45 2.24
car battery 4.52 2.40
catfood 4.55 2.20
sparkplugs 4.61 1.97
life insurance 5.06 2.38

148
The key attributes of interest were company size and event facilitation.
Perceptions of both were shown in previous studies to be impacted by sponsorship
associations. Company size was manipulated at two levels. At the low level, the profile
read small company. At the high level, it read large company. Event facilitation was
also manipulated at two levels. The attribute was given in the profile as company
involvement with favorite local events. At the low level of event facilitation, the profile
read not involved. At the high level of event facilitation the profile read helps make
them possible.
Finally, customer service hours were manipulated at two levels. At the low level,
the profile read 8AM to 6PM, Mon.-Sat. (60 hours per week). At the high level, the
profile read 24 hours/day, 7 days/week (168 hours per week).
Results
A logistic regression was used to test the impact of the five independent variables
on auto insurance choice. The logit model regressed price, deductible, company size,
event facilitation, and service hours on choice. Parameter estimates for each variable are
provided in Table 8 below.

149
TABLE 8
IMPACT OF FIVE MANIPULATED VARIABLES ON BINARY LOGIT MODEL OF
AUTO INSURANCE CHOICE
Parameter
Estimate
SE of
Parameter
Asymptotic
t-Stat
Pr(Z>|t|)
intercept
.21532644
. 86447026E-01
2.491
.0127
price
-.23668224E-01
. 21119801E-02
-11.207
.0000
deductible
-. 27158952E-02
.2593348IE-03
-10.473
.0000
facilitation
.68801263E-01
.57565562E-01
1.195
.2320
company size
.50404886
.12435944
4.053
.0001
service hours
.10788367E-01
.11847266E-02
9.106
. 0000
As can be seen in the table, four of the attributes had a significant impact on
choice. The sign of the parameter estimate indicates direction of this impact. Therefore,
in this data set, as price and deductible increase, choice decreases. As service hours, and
company size increase, choice increases. Event facilitation, however, had no detectable
impact on choice for auto insurance.
Study 3b
Overview
Subjects made a series of eight choices between two brands of yogurt (1,248 total
choices). Each choice profile consisted of five brand attribute (including size and event
facilitation) for two yogurt brands. As in 3a these attributes were manipulated
orthogonally. Each subject completed half of a 16 cell, radical fractional factorial design.
Subjects
One hundred fifty-six undergraduates in an introductory marketing class at a major
southeastern university took part. Each received course extra credit for participation.

150
Independent Variables
Five independent attribute variables were manipulated orthogonally. They were
price (four levels), shelf life (two levels), company size (two levels), event facilitation (two
levels), and calories (two levels). Each is described below (see Appendix F).
Price was set by surveying actual prices for single 6-8 oz. cups of yogurt at a local
supermarket. Four equally spaced prices within the observed price-range were selected.
These prices were $0.47, $0.59, $0.71, and $0.83. Shelf life was manipulated at two
levels. The high level of shelf life was given as 15 days. The low level was 8 days.
Company size and event facilitation were manipulated exactly as in 3 a. Calories were
manipulated at two levels. The low level was 110, and the high level was 160. These are
both within the range observed in the supermarket.
Results
As in 3 a, logistic regression was used to test the impact of the five independent
variables on choice. The logit model regressed price, calories, company size, event
facilitation, and shelf life on choice. Parameter estimates for each variable are provided
below in Table 9.

151
TABLE 9
IMPACT OF FIVE MANIPULATED VARIABLES ON BINARY LOGIT MODEL OF
YOGURT CHOICE
Parameter
Estimate
SE of
Parameter
Asymptotic
t-Stat
Pr(Z>|t|)
intercept
.10824559
. 72024387E-01
1.503
.1329
price
85104661E-01
.61957660E-02
-13.736
.0000
calories
-.21349918E-01
.20470950E-02
-10.429
. 0000
facilitation
.16591370
. 48397839E-01
3.428
.0006
company size
.31744675
.10012614
3.170
.0015
shelf life
.10856916
. 14422567E-01
7.528
.0000
As can be seen in the table, all five of the attributes had a significant impact on
choice. As before, the sign of the parameter estimate indicates direction of impact on
choice. Therefore, in this data set, as price and calories increase, choice probability
decreases. As event facilitation, company size, and shelf life increase, choice probability
increases. In the yogurt category, both event facilitation and company size had a
significant, positive impact on choice. Large companies and those that facilitated events
were preferred.
This pair of studies has shown that the impact of sponsorship-based inferences can
differ by brand category. Both studies manipulated company size and event facilitation.
Perceptions of both of these attributes have been shown in previous studies to be
influenced by sponsorship. In the auto insurance category (study la), company size had a
measurable impact on choice. Event facilitation had no measurable impact.
For the yogurt category, however, both event facilitation and company size had a
significant impact on choice. These findings lend support to the idea that specific
sponsorship-based inferences can have different effects on choice in different brand

152
categories. Therefore, managers need to determine the attribute perceptions which impact
choice in the category before setting sponsorship strategy.
It is also possible that brand specific factors might affect the impact sponsorship-
based inferences have on choice. That is, within a category, certain inferences might
impact some brands more than others. To investigate this possibility, a third study was
run, where the two brands in the same category differed along a relevant dimension.
Study 3c
Overview
Subjects made a series of eight choices between two brands of auto insurance (984
total choices). Each choice profile consisted of the same five brand attributes at the same
levels as in 3a. Once again these choices were made between two brands. The two brands
used in this study differed, however, in that one was a domestic auto insurance company,
and the other was a Japanese auto insurance company. The five attributes were
manipulated orthogonally, using the same fractional design as in 3 a. As before, each
subject completed half (8 cells) of the 16 cell design.
Subjects
One hundred twenty-three undergraduates in an introductory marketing class at a
major southeastern university took part. Each received course extra credit for
participation.

153
Independent Variables
Five independent attribute variables were manipulated orthogonally in exactly the
same manner as in 3 a. Again, these attributes were price (four levels), deductible (two
levels), company size (two levels), event facilitation (two levels), and hours of customer
service (two levels). These attribute levels were identical to those in 3a.
The additional manipulated variable in this study is country-of-origin. In this
study, subjects were choosing between the Wilson auto insurance company of Atlanta,
GA, and the Kobayama auto insurance company of Japan. Pretesting revealed that auto
insurance is a product class where domestic brands were strongly preferred. The same 43
categories as in Table 7 were tested for domestic vs. foreign preference. The same 33
subjects rated these 43 categories along a 9-point scale anchored by Always prefer an
American brand and Always prefer a foreign brand. Results of this pretest are shown
below.

154
TABLE 10
PREFERENCE FOR DOMESTIC VS. FOREIGN BRANDS
Cateaorv
mean*
stdev
credit card
1.94
1.30
car insururance
2.03
1.47
medical insurance
2.24
1.84
long distance
2.27
1.51
life insurance
2.30
1.49
bank
2.36
1.88
cable
2.48
1.92
renters insurance
2.48
1.62
cell phone service
2.61
1.69
airline
2.70
1.81
investment company
2.73
1.82
encyclopedia
2.79
1.90
mutual fund
2.79
1.82
frozen dinner
3.18
1.83
internet service
3.30
1.83
yogurt
3.30
1.79
music club
3.42
1.75
condom
3.48
1.91
deodorant
3.58
1.84
ice cream
3.70
2.02
athletic shoes
3.76
2.05
gassoline
3.82
1.96
car battery
4.06
1.90
battery
4.09
2.11
dogfood
4.18
1.53
catfood
4.24
1.70
film
4.27
2.11
shampoo
4.27
1.51
razors
4.36
1.54
tools
4.45
1.80
spark plugs
4.64
1.45
motor oil
4.67
0.99
carwax
4.76
1.32
charcoal
4.76
1.09
exercise equip.
4.85
1.72
ceiling fan
4.88
1.14
paint
4.91
1.13
computer
5.03
2.32
personal watercraft
5.27
1.75
cellular phone
5.67
2.15
beer
6.24
2.22
DVD
6.33
2.10
car
6.48
2.50
1= Always prefer American; 9=Always prefer foreign
As can be seen in the table, auto insurance is a category in which American brands
are almost always preferred. This means that, relative to a domestic company, a foreign

155
auto insurance company faces additional obstacles to acceptance. This fact may mean that
sponsorship-based inferences may have a different impact on the foreign brand
Kobayama, than on the domestic brand Wilson. This possibility is tested formally
with a logit model that includes brand-specific parameters for both company size and
event facilitation
Results
Once again, a logistic regression was used to test the impact of the independent
variables on choice. To test whether the impact of brand size and event facilitation was
different for the foreign and domestic brands, brand-specific parameters were estimated
for these two variables. This resulted in seven total parameter estimates of interest: the
impact of deductible, price, and service hours on overall choice, as well as the impact of
company size on foreign and domestic brands (two parameters), and the impact of event
facilitation on foreign and domestic brands (two parameters). The parameter estimates for
this model are presented in Table 11 below.

156
TABLE 11
ALTERNATIVE SPECIFIC IMPACT OF MANIPULATED VARIABLES ON AUTO
INSURANCE CHOICE: FOREIGN VS. DOMESTIC
Parameter SE of Asymptotic
Estimate
Parameter
t-Stat
Pr(Z>|t|)
intercept
.27265667
.17080180
1.596
.1104
price
-.17256471E-01
.15446752E-02
-11.172
. 0000
deductible
-.15683230E-02
.21380779E-03
-7.335
. 0000
domestic
facilitation
57844041E-01
.79390851E-01
-.729
.4662
foreign
facilitation
. 17589399
.80882673E-01
2.175
.0297
domestic
brand size
.94754983E-01
.16374539
.579
.5628
foreign
brand size
.96620186
.17070479
5.660
. 0000
service hours
.92624914E-02
.10521120E-02
8.804
. 0000
As was the case in 3a, choice decreased as price and deductible increased. Also as
in 3a, as service hours increased, choice increased. All of these effects are as would be
expected.
Turning to the brand specific parameter estimates, however, the results are quite
different from those found in 3 a. Event facilitation, which previously had no impact on
auto insurance choice, now has a significant impact (p<03) on choice for the foreign
brand. For the domestic brand, no influence of event facilitation on choice is found. It
appears, therefore, that event facilitation can differentially impact different brands within
the same category.
Brand size as well shows differential impact on the foreign and domestic brands.
In this study, brand size has a significant positive impact on choice for the foreign brand,
but no measurable impact on choice for the domestic brand. Once again, evidence is
found that even within the same product category an attribute of the brand has a different

157
impact on one brand vs. another. Given that sponsorship has been shown to influence
perceived brand size and perceived event facilitation, the findings of this study
demonstrate a potential difference in the value of a sponsorship between two brands in the
same category.
Overall Discussion of Studies 3a-3c.
The goal of this chapter was to determine whether attribute perceptions, which had
been shown to be impacted by sponsorship, would have different impacts on choice for
different brands. The three studies were set up so as to detect differential impact in
different categories, as well as on different brands within a single category. A third, and
unexpected differentiating factor, market structure, may also have an impact on the
importance of various attribute perceptions.
A comparison of studies 3 a and 3b demonstrates that an attribute which can have
its perceptions shaped by sponsorship (i.e. event facilitation) can have an impact on one
product category (yogurt), but not on another (auto insurance). This finding supports the
notion that managers need to consider the overall needs of their category when setting
sponsorship strategy.
Study 3b demonstrates that an attribute which can have its perceptions shaped by
sponsorship (i.e. company size) can have an impact on one brand (Kobayama), but not on
another (Wilson). This finding supports the notion that managers must also consider the
specific situation of their brand when developing a sponsorship strategy.
A comparison of studies 3 a and 3 c demonstrates that an attribute which can have
its perceptions shaped by sponsorship (i.e. company size) can have an impact under one

158
market structure (domestic competitor), but not on another (foreign competitor). This
finding supports the notion that managers need to consider the competitive environment
when setting sponsorship strategy.
Taken together, these three findings suggest that in order to evaluate the value of a
sponsorship opportunity, a manager must consider the needs of the specific brand and the
category within the current competitive environment. It is not necessary, however, to test
multiple scenarios as has been done in this chapter. An individual decision maker need
only calculate the relative value of each sponsorship-based inference for that brand. Once
these are known, selecting from among various sponsorship opportunities becomes a
matter of considering the likely inferences generated by each available association, and
picking sponsorships that can generate the inferences most desirable for the particular
brand
Basic strategies for developing sponsorship strategies are covered in the following
chapter, which discusses the combined findings of the eight studies, and how they can be
used by managers to make sponsorship decisions. The next chapter also discusses future
research which will eventually lead to more rigorous mathematical screening criteria for
selecting sponsorship associations for brands.

CHAPTER ELEVEN
CONCLUSION, IMPLICATIONS, AND FUTURE RESEARCH
The way in which sponsorship impacts consumers is not at all obvious. If you
were to ask a consumer who attended an event, what impact did sponsorship have on
you the likely answer would be none. Should the conclusion be drawn, therefore, that
that the six billion dollars spent per year on sponsorship is wasted money? The results of
the eight studies presented in this dissertation would argue definitely not.
It is easy for a consumer to say that an advertising campaign, a sponsorship, or
even a persuasive salesperson had no influence on his/her decision. Because everyone
has experience as a consumer, it is easy to extrapolate a feeling of no influence on me to
a theory of no influence on anyone. For both personal selling and advertising, there is,
however, a long history of a measurable impact on sales. This body of evidence makes it
easy to prevent the no impact on me feeling from guiding beliefs about aggregate
effects. Such a long history does not exist for sponsorship. It is, therefore, harder to
refute the assumption that the impact of sponsorship on behavior is limited.
The companies pouring the billions into sponsorship must believe that they are
getting something in return. Reading the trade press and talking to practitioners has made
it clear, however, that there is no consensus about the mechanisms through which
159

160
sponsorship impacts behavior. This dissertation is the result of an attempt to address these
issues.
More specifically, the purpose of this dissertation has been to explore the variety
of ways in which sponsorship works to enhance the equity of the sponsoring brand. In
particular, it has investigated how engaging in sponsorship influences various inferences
about the brand, and how these inferences impact brand choice. It began by presenting a
review of extant sponsorship research, which was found to be largely descriptive and
speculative in nature. A summary of relevant research from related commercial
association domains, like celebrity endorsement and brand and line extension followed.
A conceptual framework for commercial associations in general was then
developed. The similarities in findings from celebrity endorsement, brand extension and
other commercial association literatures were presented. Likely implications for
sponsorship of the common findings of these literatures were developed into formal
propositions. The unique nature of sponsorship, and its likely impact on brand equity and
positioning, was then provided. The marketing variables which may be affected by
sponsorship were examined, and specific proposals about the impact of sponsorship on
brand equity and positioning were presented.
The specific mechanisms through which sponsorship works were then considered.
Exploratory research and managerial intuition as to how sponsorship works were
developed into seven testable mechanisms through which sponsorship may function in
various situations. In approximate order of necessary cognitive elaboration, these
mechanisms were posited to be simple awareness, affect transfer, image transfer,
affiliation, implied size, implied endorsement, and reciprocity. The psychological

161
underpinnings of these seven mechanisms were proposed and discussed. This allowed
many formal propositions to be developed.
Constructs hypothesized to have an impact upon the functioning of these
mechanisms, such as fit and sacrilege were then developed. The impact of various types
of brand/event fit were discussed at length. A review of how the fit construct has been
presented in the study of sponsorship and other commercial association domains was
presented. Other, more technically defined types of fit, such as attribute fit, audience
fit and positioning fit were then defined. The discussion of fit concluded with
measurement issues.
Sacrilege arising from associations perceived to be inappropriate by consumers
was then discussed. Formal propositions about when it may arise and its likely impact on
sponsorship success were also presented. A theoretical discussion of how other
communications can be deployed to maximize the impact of the sponsorship association
concluded the background and theory chapters.
A two stage inference-based model of building brand equity through sponsorship
was then developed to explain the functioning of several of the seven mechanisms. The
first stage in the model considered how audience, event and associational factors of a
sponsorship impact consumer perceptions about brands. Five studies were presented
which explored these issues. The second stage of the model considered how these
inferences impact choice for different brands and product categories. Three studies
explored these issues using stated-preference choice experiments.

162
Implications
Early on, it was proposed that sponsorship was the only means of communicating
with customers without the benefit of a verbal message, and that any message received
must be derived from the association in the mind of a consumer. Some authors, however,
have suggested that the primary impact of sponsorship on consumers is simple top-of-
mind awareness (e.g. Aaker, 1991). If this were true, the only pertinent issues in
sponsorship research would be the size and demographics of each event audience.
Brands such as Sony, however, claim to use sponsorship to reach the previously
unattainable goal of making their product a bigger part of the Gen-X lifestyle (Williams
1996). LA Gear claims to have successfully used sponsorship associations to reposition
its product line (Proctor 1996). Tanqueray gin says it used sponsorship to portray a less
stodgy image (Elliot 1995).
It appears, therefore, that the practitioners making sponsorship decisions believe
that associating a brand with an event can do more than simply remind people about the
brand name. This view has been shared by a variety of authors who have used anecdotal
evidence to posit mechanisms through which sponsorship might influence brand
perceptions (Hastings, 1984; Meenaghan, 1991b ; Thomas, 1985; Crowley, 1991; Parker,
1991; McDonald, 1991; Otker, 1988; Ryssel and Stamminger,1988; Schoch, 1994;
Walshe and Wilkinson, 1994). The research presented in this dissertation, however, is the
first to find empirical support for such mechanisms. These findings, and their implication
for sponsorship strategy, are revisited below.

163
The most robust finding of sponsorships impact on brand perception is that
associating a brand with a larger event leads to perceptions that the brand is large. This
result is found across three studies (la,lb and lc), for both sports (running) and cultural
(literature) events. These results replicate experimentally what Rajaretnam (1994) found
with a pre-test/post-test field study. Taken together, these findings seem to give strong
support to the notion that associating a brand with large events can lead to the perception
that brand itself is large. This impact of sponsorship on brand perceptions is likely to be
more helpful to some brands than others (see below).
Another result to replicate across multiple studies (la and lc) is that associating a
brand with a larger event leads to an increase in perceived product quality. While this
effect initially appeared to be stronger under title sponsorship, follow-up analysis showed
no significant enhancement from the event being named after the brand. This impact of
event size on perceived brand quality obtained across both brand categories (coffee and
sports drinks) and both event categories (running and literature).
Perceived brand size appears to also be affected by duration of association
between the brand and the event. In study la, duration had a significant, positive impact
perceived brand size. A comparison of the pooled, open ended data from studies la and
lb replicated this effect. This is important, because it implies that the benefit to the brand
of sponsoring an event is enhanced over time. While media departments in advertising
agencies are constantly looking for new, more cost effective means to reach target
audiences, these results suggest that for sponsorship, a longer term approach is more
beneficial.

164
Event facilitation inference making was found to be enhanced by three factors.
Analysis of the pooled data between la and lc showed a significant impact of title
sponsorship on event facilitation inferences. Study la showed a positive impact of
duration of association on event facilitation inferences. Follow-up analysis in 2a, which
found event knowledge to be correlated with the perceived size of the event, supports the
notion that sponsoring smaller events leads to greater event facilitation inferences.
Therefore, if a brand is trying to improve its perception as a good corporate citizen, these
results suggest a sponsorship strategy of being the title sponsor of many, smaller events
for a long time. With this type of strategy as well, it appears that an additional positive
impact of sponsorship accrues over time.
If one were to assume, however, that each type of inference was equally valuable
to all brands, the results of the first five studies could be said to provide limited assistance
to a manager trying to assess the value of different sponsorship opportunities. Assuming
such an equal value for all brands would only support a strategy of being the title sponsor
of high fit events for a long time. Assuming different effects on different brands, however,
allows more specific, and perhaps less obvious recommendations to be made.
The results of studies 3 a, 3 b, and 3 c, in fact, support differential impact of
inferences on different brands. These studies show that the importance of individual
inferences can vary between product categories, between brands within a category, and
even between different competitive situations for a single brand. Finding that each
inference need not impact all brands uniformly makes the findings of the first five studies
much more useful. An example of how this information can be used to select between two
sponsorship opportunities is presented below in Table 12 .

165
TABLE 12
TWO HYPOTHETICAL SPONSORSHIP PROPOSALS
ProDosal
Total # exDosed
Cost
#1-1 event perceived to be large
10,000,000 people
$110,000
#2 20 events perceived to be small
10,000,000 people
$100,000
Given the two proposals above, a brand in a similar situation to one of the domestic auto
insurance companies in study 3 a should probably choose proposal #1, since the brand size
variable had a significant impact on choice while the event facilitation had no measurable
impact in that market. Given the lower cost, however, a brand in the situation of the
yogurt brands in study 3b, should probably choose proposal #2. A brand which is strongly
impacted by event facilitation inferences, but not by brand size, would clearly choose #2.
This example should make two things clear. The first is that a good deal has been
learned about how sponsorship impacts brand choice. The second, however, is that there
is still much work to be done on quantifying this impact. Potential avenues for improving
quantitative rigor are presented below.
Future Research
The research presented in this dissertation represents a first step toward
understanding how to optimally include sponsorship in an overall marketing campaign. It
has been shown that sponsorship, the promotional medium that conveys no overt
message, is in fact capable of generating inferred messages in the minds of consumers. It
has also been demonstrated that these inferences can impact brand choice, and that this

166
impact varies from brand to brand. Some general guidelines for generating brand size
vs. event facilitation inferences were presented above. Such general guidelines are only
a first step towards more rigorous analysis which should eventually make more specific
recommendations possible. The additional information which is needed before such
analysis will be possible is presented below.
Absolute Measures of Inference Making
Studies la, lb, lc,2a, and 2b provided some absolute measures of inference making.
These studies support the notion that sponsorship associations can lead to inferences
about the brand. It is not known, however, how the level of inference making exhibited in
this particular experimental setting compares with inference making from more intrusive
exposures to associations such as event promotional messages, mass media coverage and
actual event attendance.
The absolute amount of inference making in the five studies is easy to calculate.
Across the five studies which measured inference making (la,lb,lc,2a,and 2b) 99 out of
310 subjects (31.93%) made at least one inference about the size, legitimacy, event
facilitation or commercial motive of the sponsoring brand. This level of inference making
is fairly uniform across the five experiments, from a low of 25% in lb, to a high of 38% in
lc. It could be argued, however, that this number is inflated due to the task demand
involved with asking subjects to list thoughts. That is, consumers might not make as many
inferences when they are not asked to list their thoughts.
It is quite possible, however, that the amount of spontaneous inference making in
these experiments is much lower than would occur with actual sponsorship. It should be

167
pointed out that the inferences recorded in the thought listing task were the result of a
single, 60 second exposure to a poster for an event. If compared to multiple pre-event
promotional exposures, and/or several hours of actual event viewing or attendance, the
percentage of people drawing inferences about a sponsoring brand in these experiments
could easily be underrepresented.
Determining the actual amount of inference making generated by sponsorship is
key to using the inference-based model to assess the actual value of a sponsorship for a
particular brand. Techniques similar to those employed in 3a,3b,and 3c should be able to
evaluate the change in choice probability once a certain inference is made. Being able to
calculate what percentage of consumers exposed to an association will make each
inference would allow precise calculations of the value of a sponsorship to be made. Some
combination of experimental evidence (to establish causality) and correlational field
research (to establish magnitude) must be brought together in order to make reasonable
estimates of the number of consumers who will draw certain inferences when exposed to
an association.
Once methods for making such estimations exist, calculating the value of a
sponsorship to a brand becomes a straightforward endeavor. The percentage of people
drawing an inference can be multiplied by an estimate of change in choice probability given
the inference. One possible form for such an evaluative function is provided in Appendix
I.

168
Impact of Audience Involvement and Knowledge on Inference Making
One area which has been left somewhat unresolved is the issue of how involvement
and knowledge impact inference making. Given that absolute levels of inference making
need to be calculated, the role of differential knowledge and involvement cannot be
overlooked. Studies 2a and 2b measured both constructs at an individual level. Field
research, however, will probably have to rely upon average audience involvement and
knowledge across an event. This more aggregate measure, while less valuable for
investigating specific mechanisms, should prove quite adequate as an input to predicting
event-wide rates of inference making. This level of abstraction will be appropriate to the
task of calculating absolute rates of inference making across each event.
One interesting finding from the individual level data must be considered in the
aggregate. Study 2a revealed that within the sample, low knowledge was highly correlated
with lower perceived event size. The possibility that people low in knowledge about an
event generally perceive it to be smaller than high knowledge people is intriguing. This
finding provides a new insight into the relationship between knowledge and event
facilitation inferences. Future research should investigate when event knowledge is and is
not correlated with size. It is quite possible that people low in knowledge about events
generally perceive them to be smaller than high knowledge people. Determining the
frequency with which this is true will allow a deeper understanding of the role of domain
knowledge in inference making.

169
Conclusion
The research reported here provides support for an impact of sponsorship on brand
perceptions, and ultimately on brand choice. Such findings have not been demonstrated
previously. Of special interest, evidence was found that the perceived size of a sponsoring
brand can be influenced by the size of the sponsored event and by the duration of the
association. Perceptions that the brand is helping the event, however, are shown to be
enhanced by title sponsorship, sponsoring small events, and being a sponsor for a long
time. These findings are especially important because it also was found that perceptions of
brand size and perceptions of helping the event are not equally important to all brands.
Taken together, this set of findings suggests the sponsorship association which will
be most beneficial can only be determined once the brand situation is known. Once the
value of changing different brand perceptions has been determined, the research presented
in this dissertation can provide some general guidelines for optimal sponsorship choice.
Future research in this area should make these recommendations much more specific.
Some academic researchers and managers still perceive sponsored events as little
more than billboards for brand logos. They consider sponsorship to be exclusively an
exposure vehicle which is only capable of building top-of-mind awareness. The
conceptual framework and empirical evidence presented in this dissertation, however,
suggest much richer possibilities for the promotional medium of sponsorship.

APPENDIX A
OPEN ENDED INFERENCE MEASURES AND MANIPULATION CHECKS FOR
EXPERIMENT 1A,1B AND 1C
Now we would like to get your impressions about one of the sponsoring brands:
Extran Sports Drink (or Rombouts coffee)
Do you think Extran Sports Drink (Rombouts Coffee) is:
One of the smallest sports One of the largest sports
drinks (coffees) in Belgium | | | | | | | | drinks (coffees) in Belgium
How large do you think Extrans (Rombouts) market share is, compared to other sports
drinks in Belgium?
Very small share | | | | | 1 | | very large share
How legitimate of a sports drink (coffee) do you perceive Extran Sports Drink
(Rombouts Coffee) to be?
A fly-by-night brand | | | | | | | | A legitimate brand
How real of a brand do you perceive Extran Sports Drink (Rombouts Coffee) to be?
Not very real | | | | | | | | Very real
170

171
What do you think the quality level of Extran Sports Drink (Rombouts Coffee) is likely
to be?
Very poor quality | ]_ | | j | | | Very high quality
How important do you think the support of Extran Sports Drink (Rombouts Coffee) is to
the sponsored event taking place?
Not important at all | | | | | | | | Very important
How important do you think the support of Extran Sports Drink (Rombouts Coffee) is to
the success of the sponsored event?
Not important at all | | | | | | | | Very important
How many events do you think Extran Sports Drink (Rombouts Coffee) sponsors each
year?
Very few events | | | | | | | | Lots of events
Compared with other brands of sports drink (coffee) how many events do you think
Extran (Rombouts) sponsors
Far fewer than others | | ¡ |_ | | | | Many more than others
Overall, how large do you perceive the running (literature) event advertised to be?
A very small event | _| J | | | | | A very large event
Overall, how would you rate the fit between the running (literature) event advertised in
the poster and Extran Sports Drink (Rombouts Coffee)
Very poor fit | | | | | | | | Very good fit

172
Have you ever been to Europe? Y N
Have you ever been to Belgium? Y N
Have you ever heard of Extran
Sports Drink (Rombouts Coffee)
Before today? Y N
Have you ever run in a marathon
(published a poem)? Y N
Have you ever run in a 1 OK
Race(gone to a poetry reading)? Y N

APPENDIX B
OPEN ENDED INFERENCE MEASURES AND MANIPULATION CHECKS FOR
EXPERIMENT 2A AND 2B
Now we would like to get your impression about EVIAN (TEXACO) (TEXACO) and the US Open
(Indy 500) (Indy 500)
How important do you think the support of EVIAN (TEXACO) (TEXACO) is to the US Open (Indy 500)
(Indy 500) taking place?
Not at all important I I I 1 | | 1 | Very important
How important do you think the support of EVIAN (TEXACO) is to the success of the US Open?
Not at all important I I I 1 | | | | Very important
How important do you think the support of EVIAN (TEXACO) is to the quality of the US Open?
Not at ail important 1 I I I I 1 I I Very important
Please indicate your level of agreement with these statements:
EVIAN (TEXACO) must be a good water, or it wouldnt be able to be a sponsor of this event
strongly disapee I I I I I I I | strongly apee
EVIAN (TEXACO) is demonstrating good corporate citizenship through its support for this event
strongly disapee I 1 i 1 I I 1 | strongly agree
EVIAN (TEXACO) is just sponsoring this event to get a lot of people to see their name,
strongly disagree I I 1 1 I I I I strongly agree
173

174
Please answer the following:
How many events do you think EVIAN (TEXACO) sponsors each year?
Very few events I 1 I | 1 1 1 [ Lots of events
Compare to other waters, how many events do you think EVIAN (TEXACO) sponsors
Far fewer than others | | 1 | | | 1 | Many more than others
Overall, how large do you perceive the US Open (Indy 500) to be
a very small event I I I I I 1 I I a very large event
Overall, how would you rate the fit between EVIAN (TEXACO) and the US Open
Very poor fit I I I I 1 1 I I very good fit

APPENDIX C
INVOLVEMENT AND KNOWLEDGE MEASURES USED IN EXPERIMENT 2A
Please indicate your level of agreement with the following statements by placing an X in the appropriate
box.
The message in the poster I just saw was important to me
strongly agree | | | | | | | | strongly disagree
The poster didnt have anything to do with me or my needs
strongly agree | | | | | | | | strongly disagree
Please take a moment (approximately 5-10 seconds) to think about tennis and what it means to you. Then
complete these scales.
important 1
I
1
1
1
!
1
J unimportant
of no concern (_
1
1
1
1
1
1
_J of great concern
irrelevant |_
|
J
1
1
1
1
_| relevant
means a lot to me |_
I
I
I
1
1
1
I means nothing to me
useless |_
|
1
I
I
1
1
_J useful
valuable |_
|
1
I
I
1
I
_| worthless
trivial 1_
|
1
1
1
1
1
J fundamental
beneficial |_
|
1
I
I
1
I
_J not beneficial
matters to me [_
1
1
1
I
1
I
_J doesnt matter
uninterested |_
|
1
I
1
1
1
_] interested
significant 1
1
1
I
!
1
1
J insignificant
vital |_
1
1
1
1
1
1
I superfluous
boring |
1
I
1
1
1
1
1 interesting
unexciting |
|
I
I
1
1
1
I exciting
appealing |
I
1
1
1
1
1
1 unappealing
mundane (_
I
1
I
!
1
1
1 fascinating
essential |
_L_
I
I
1
_l_
1
J nonessential
175

176
undesirable I I I I 1 [ I I desirable
wanted 1111 111! unwanted
not needed 1 I 1 j ¡ ¡ | | needed
Please rate your level of agreement with the following statements:
know a great deal about the game of tennis
strongly agree | j | | | | | |strongly disagree
I know more about the game of tennis than most of my friends
strongly agree | | | | | | | |strongly disagree
I watch most of the major tennis events that are on TV
strongly agree | | | | | | | |strongly disagree
I watch more tennis on TV than most of my friends
strongly agree | | | | | | j ¡strongly disagree
I play tennis quite frequently
strongly agree | | | | | ¡ | ¡strongly disagree
I play tennis more often than most of my friends
strongly agree | | | | ¡ | | ¡strongly disagree
I am a good tennis player
strongly agree | | | ¡ | ¡ | ¡strongly disagree
I am a better tennis player than most of my friends
strongly agree | | ¡ | | | ] ¡strongly disagree

177
Playing tennis lets people see me as I ideally would like them to see me
strongly agree | | | | | | | |strongly disagree
Tennis helps me to attain the type of life I strive for
strongly agree | | | | | | | |strongly disagree
I can make many connections or associations between tennis and experiences in my life
strongly agree | | | | | | | |strongly disagree
Tennis is of high personal importance to me
strongly agree | | | | | | | |strongly disagree
Tennis helps me express who I am
strongly agree | | | | | | | |strongly disagree
How many tim es did you play tennis last week?
How many times did you play tennis last month?
How many times did you play since last May?
How many times did you watch tennis last week?
How many times did you watch tennis last month?
How many times did you watch tennis since last May?

178
Quick Tennis Quiz
(If not sure, make your best guess)
1 When a player hits a ball into the net, or outside the court, or fails to hit the ball before
it bounces twice, his/her opponent .
2 Returning the ball over the net before it bounces is called a .
3 For what percentage of the game does the starting server serve %
4 To serve, must the server stand behind the base line or the service line (circle one)?
5 When serving from the right, must the server hit the ball into the left or the right
service court (circle one)?
6 A serve that is not good is called a ?
7 A serve that touches the top of the net and lands in the proper service court is called a
?
8 Is it OK for the receiver to return a serve before it bounces once? Yes No (circle one)
9 A score of zero or nothing is called .
10-A score of 30-30 is called Deuce. True False (circle one)
11 You must win at least games to win a set.
12 In championship matches of big tournaments, you must win sets to win the
match.

APPENDIX D
OPEN ENDED THOUGHT LISTING BOXES
179

APPENDIX E
SAMPLE CHOICE PROFILE USED IN EXPERIMENT 3 A
Davis Auto
insurance
Wilson Auto
Insurance
Customer service
Hours
8AM-6PM
Mon-Sat
24 Hours 7
days/week
Company Size
Large
Company
Small
Company
Company involvement
with favorite local events
Not involved
Helps make
them possible
Deductible
$1000
$500
Price for 6 months
coverage
$570.00
$615.00
U
u
Please indicate your
choice (check only one)


180

APPENDIX F
SAMPLE CHOICE PROFILE USED IN EXPERIMENT 3B
Brand G
Yogurt
Brand K
Yogurt
Shelf Life
8 Days
15 Days
Company Size
Large
Company
Small
Company
Company involvement
with favorite local events
Not involved
Helps make
them possible
Calories
160
110
Price for 8 oz. container
$0.59
$0.71
U
JJ
Please indicate your
choice (check only one)


181

APPENDIX G
SAMPLE CHOICE PROFILE USED IN EXPERIMENT 3C
Kobayama Auto
Insurance
Wilson Auto
Insurance
Customer service
Hours
8AM-6PM
Mon-Sat
24 Hours 7
days/week
Company Size
Large
Company
Small
Company
Company involvement
with favorite local events
Not involved
Helps make
them possible
Deductible
$1000
$500
Price for 6 months
coverage
$570.00
$615.00
U
u
Please indicate your
choice (check only one)


182

APPENDIX H
SAMPLE EXPOSURE STIMULI FOR EXPERIMENTS 1 A, IB, 1C,2A AND 2B
This appendix contains example of the posters used to expose subjects to the
sponsorship associations in the first five experiments. Two examples are given for
experiments la, lb and lc. One representing the condition in which manipulated
variables are all high (i.e. large event size, high fit, long duration of association and
high share of presence) and one condition in which all manipulated variables are low (i.e.
small event size, low fit, short duration of association and low share of presence). The
high fit conditions are given for 2a and 2b.
183

184
EXPERIMENT 1A ALL HIGH
The Euro-Capital
International Marathon
Brussels, Belgium
Sunday October 19th 1997 at 9:30 am
Proud to be a Sponsor for 5 Years

185
EXPERIMENT 1A ALL LOW
The Leuven
Student Fun-Run
Leuven, Belgium
Sunday October 19th 1997 at 9:30 am
Proud to be a New Sponsor

186
EXPERIMENT IB ALL HIGH
International Literature Festival
and Poetry Competition
Brussels, Belgium
October 24*-26T 1997
The European Union
Attended by:
Her Royal
Highness Queen
Fabiola
Speakers include:
Kurt Vonnegut
Hugo Claus
Poet Laureate of
Belgium
and
Seamus Heaney
Winner of the
Events Include:
Live Poetry
Readings
Performance Art
Happenings
and
Presentations of
European Union
Poetry
Scholarships
Sponsors:
Proud to be a sponsor
for 5 Years
Coolkat Books
Generate Bank

187
EXPERIMENT IB ALL LOW
Leuven City Library Literature
Festival and Poetry Competition
Leuven Belgium
October 241h-26,h 1997
Speakers
Include:
Wilfried
Vanhonacker
Head libraran
Luk Warlop
Poetry Club
President
and
Christoph Van
Den Bulte
Literature Professor.
Events Include:
Student Poetry
Readings
Amateur
Performance Art
and
Presentations of
the Leuven
Poetry Club
Award Ribbons
Sponsors:
Proud to be a New
Sponsor

188
EXPERIMENT 1C ALL HIGH
The Extram Euro-Capital
International Marathon
Brussels, Belgium
Sunday October 19th 1997 at 9:30 am
Proud to be a Sponsor for 5 Years

189
EXPERIMENT 1C ALL LOW
The Rombouts
Student Fun-Run
Leuven, Belgium
Sunday October 19th 1997 at 9:30 am
Proud to be a Sponsor for 5 Years

190
EXPERIMENT 2 A HIGH FIT
Win a Trip to
The 1998 US Open
August 24th-September 5th
In New York City
See Bottle for Details
Evian is proud to be an official sponsor of the 1998 US Open

191
EXPERIMENT 2B HIGH FIT
Win a Trip to
The 1998 Indy 500
May 23rd-25th 1998
At the Indianapolis Motor Speedway
From
See Store for Details
Texaco is proud to be an official sponsor of the 1998 Indy 500

APPENDIX I
ONE POSSIBLE FORM OF A SPONSORSHIP VALUATION FUNCTION FOR A
SPECIFIC BRAND
Value = MVSBI + MVE
Where:
Value=Change in brand equity from sponsorship
MVSBI= Marginal Value of Sponsorship Based Inferences
MVE= Marginal Value of Exposure Through Sponsorship
Assume that efficient markets lead to similar costs per 1000 exposures for various
sponsorships. Therefore, exposure per dollar becomes a constant, and deciding which
sponsorship opportunity is best for a given brand reduces to a comparison of the MVSBI
associated with each sponsorship opportunity.
4
MVSBI=Marginal value of sponsorship based inferences
R=Reach (number of people aware of the sponsorship association)
M=Margin (net present value of marginal revenue per customer)
%i=% inference (percentage of R who make inference x)
CF=Choice factor (change in choice probability given inference x)
Inferences:
x=l: size
x=2: legitimacy
x=3: facilitation
x=4: commercial motive
192

REFERENCES
Aaker, David A (1990) Brand extensions: the good, the bad and the ugly. Sloan
Management Review. Summer, 47-57
(1991) Managing brand equity: capitalizing on the value of a brand name. New
York: The Free Press
and Kevin Lane Keller (1993) Interpreting cross-cultural replications of brand
extension research. International Journal of Research in Marketing. 10, 55-59
and Kevin Lane Keller (1990) Consumer evaluations of brand extensions
Journal of Marketing. 54, 1, 27-41
and Donald E. Bruzzone (1985) Causes of irritation in advertising. Journal of
Marketing. 49, 2, 47-57
Abratt, Russel, Brian C. Clayton, and Leyton F. Pitt (1987) Corporate objectives in
sports sponsorship International Journal of Advertising. 6, 299-311
Alonzo, Vincent (1994) Misuse of koran k.o.'s world cup promo. Incentive. 168, 8, 8
Andrews, Jim (1996) Sponsorship overview: what you need to know. Paper presented
at the 1995 IEG Event Marketing Conference, Chicago Illinois.
Armstrong, Clive (1988) Sports sponsorship: a case-study approach to measuring its
effectiveness. European Research. May, p.97-103
Barsalou, Lawrence W. (1982) Context-independent and context-dependent information
in concepts. Memory & Cognition. 10, 1, 82-93
(1983) Ad hoc categories. Memory & Cognition. 11, 3, 211-227
(1985) Ideals, central tendency, and frequency of instantiation as determinants
of graded structure in categories. Journal of Experimental Psychology: Learning,
Memory. & Cognition. 11. 4. 629-654 =
193

194
Bawa, Kapil, Jane T. Landwehr, and Aradhna Krishna (1989) Consumer response to
retailers' marketing environments: an analysis of coffee purchase data. Journal of
Retailing. 65, 4, 471-495
Belk, Russell W., and Melanie Wallendorf (1990) The sacred meanings of money.
Journal of Economic Psychology. 11, 1, 35-67
, Melanie Wallendorf, and John F. Sherry Jr.(1989) The sacred and the profane
in consumer behavior: Theodicy on the Odyssey. Journal of Consumer Research.
16, 1, 1-38
Berger, Seymour (1965) Persons as mediators: a four-stage paradigm. Journal of
Personality & Social Psychology, 2, 3, 467-469
Bhattacharya, C B., Hayagreeva Rao, and Mary Ann Glynn (1995) Understanding the
bond of identification: an investigation of its correlates among art museum
members. Journal of Marketing, 59, 4, 46-57
Bimbaum, Michael H., Rebecca Wong, and Leighton K. Wong (1976) Combining
information from sources that vary in credibility. Memory & Cognition, 4, 3,
330-336
Blackston, Max. (1995) The qualitative dimension of brand equity. Journal of
Advertising Research. 35, 4, RC-2-7
Bloch, Peter H. (1986) The product enthusiast: implications for marketing strategy
Journal of Consumer Marketing. 3,3,51-62
Boulding, William, and Amna Kirmani (1993) A consumer-side experimental examination
of signaling theory : do consumers perceive warranties as signals of quality?
Journal of Consumer Research. 20, 1, 111-123
Brehm, Jack W., and David Lipsher (1959) Communicator-communicatee discrepancy
and perceived communicator trustworthiness. Journal of Personality. 27, 352-361
Brewer, Geoffrey (1993) Here's good news for advertisers. Sales & Marketing
Management. 145, 15, 20
Broniarczyk, Susan M., and Joseph W. Alba, (1994) The importance of the brand in
brand extension. Journal of Marketing Research. 31 (May), pg. 214-228
Bultez, Alain., Philippe Naert, Els Gijsbrechts and Piet Vanden Abeele (1989)
Asymmetric cannibalism in retail assortments. Journal of Retailing. 65, 2, 153-
192

195
Carpenter, Phil M. (1994) Some cobranding caveats to obey. Marketing News. (Nov
7), 28, 23, 4
Celsi, Richard L., Simeon Chow, Jerry C. Olson and Beth A. Walker (1992) The
construct validity of intrinsic sources of personal relevance: an intra-individual
source of felt involvement Journal of Business Research, 25, 2, 165-185.
Chaiken, Shelly (1979) Communicator physical attractiveness and persuasion Journal of
Personality and Social Psychology, 37, 8, 1387-1397
Chew, Fiona (1992) The advertising value of making possible a public television
program. Journal of Advertising Research. 36, 6, 47-52
Cohen, Jacob, (1977) Statistical power analysis for the behavioral sciences.
Orlando: Academic Press
Cohen, Joel B., and Kunal Basu (1987) Alternative models of categorization: toward a
contingent processing framework. Journal of Consumer Research. 13,4, 455-472
Crawford, C. Merle (1985) A new positioning typology. Journal of Product Innovation
Management. 2, 4, 243-253
Crimmins, James, and Martin Horn (1996) Sponsorship: from management ego trip to
marketing success. Journal of Advertising Research. 36, 4, 11-21
Crowley, Martin G. (1991) Prioritizing the sponsorship audience. European Journal of
Marketing. 25,11, 11-21
Dawar, Niraj, and Paul F. Anderson (1994) The effects of order and direction on
multiple brand extensions. Journal of Business research. 30, 119-129
DeCoursey, Jennifer (1995) Papal paraphernalia proffered. Advertising Age, 66, 41, 2
Deshpande. Rohit, Wayne D. Hoyer, and Naveen Donthu (1986) The intensity of ethnic
affiliation: a study of the sociology of hispanic consumption. Journal of Consumer
Research. 13, 2, 214-220
DiMingo, Edward P. (1988) The fine art of positioning. Journal of Business Strategy, 9,
2, 34-38
Ellen, Pam Scholder, Joshua Lyle Wiener, and Cathy Cobb-Walgren (1991) The role of
perceived consumer effectiveness in motivating environmentally conscious
behaviors. Journal of Public Policy & Marketing. 10,2,102-117

196
Elliot, Stuart (1995) Tanqueray pulls out the stops for bike event to benefit AIDS,
patients and Mr. Jenkins rides along New York Times, July 24.
Farquar, Peter H. (1989) Managing brand equity. Marketing Research. 1, 24-33
and Paul M. Herr (1992) The dual structure of brand association. In Building
Strong Brands. David A Aaker and Alexander L. Biel, eds. Hillsdale, NJ:
Lawrence Erlbaum Associates, 263-277
, Julia Y. Han, Paul M. Herr, and Yuji Ijiri (1992) Strategies for leveraging
master brands. Marketing Research. 4, 32-43
Frieden, JonB. (1984) Advertising spokesperson effects: an examination of endorser
type and gender on two audiences Journal of Advertising Research. 24, 5, 33-41
Friedman, Hershey H. and Linda Friedman (1979) Endorser effectiveness by product
type. Journal of Advertising Research, 19, 5, 63-71
Friestad, Marian, and Peter Wright (1995) Persuasion knowledge: lay people's and
researcher's beliefs about the psychology of advertising. Journal of Consumer
Research. 22, 1, 62-74
Gaedeke, Ralph M., and Dennis H. Tootelian (1988) Understanding how clients select
and evaluate law firms. Journal of Professional Services Marketing. 3, 4, 199-207
Gardner, Meryl Paula, and Phillip Joel Shuman (1987) Sponsorship: an important
component of the promotions mix. Journal of Advertising. 16,1,11-17
Gray, Edmund R. and Larry R. Smeltzer (1985) Corporate image-an integral part of
strategy. Sloan Management Review. Summer, 73-78
Hansen, Flemming, and Lene Scotwin (1995) An experimental inquiry into sponsorship:
what effects can be measured? Marketing and Research Today. 23, 5, 173-181
Hastie, Reid, Colin Schroeder and Renee Weber (1991) Creating complex social
conjunction categories from simple categories Bulletin of the Psvchonomic
Society. 28, 3, 242-247
Hastings, Gerald, B. (1984) Sponsorship works differently from advertising.
International journal of Advertising. 3, 171-176
Heider, F. (1946) Attitudes and cognitive organization. Journal of Psychology. 21, 107-
112

197
Hildreth, C. H., and Bernard C. Nalty (1969) 1001 questions about aviation history,
Dodd, Mead and Company, New York, 162-169
Hirschman, Elizabeth C. (1991a) Sacred, secular, and mediating consumption imagery in
television commercials. Journal of Advertising Research. 30, 6, 38-43
(1991b) Babies for sale: market ethics and the new reproductive technologies.
Journal of Consumer Affairs, 25, 2, 358-390
Hoek, Janet, Phililp Gendall, and Mark Stockdale (1993) Some effects of tobacco
sponsorship advertisements on young males. International Journal of Advertising.
12, 25-35
Horia, Joann, Nicholas Naccari and Elliot Fatoullah (1974) The effect of expertise and
physical attractiveness upon opinion agreement and liking Sociometrv. 37, 4,
601-606
Hovland, Carl I. and Wallace Mandell (1952) An experimental comparison of
conclusion-drawing by the communicator and by the audience. Journal of
Abnormal & Social Psychology, 47 581-588
Jacoby, Jacob, and David Mazursky, (1984) Linking brands and retailer images-do the
potential risks outweigh the rewards. Journal of Retailing 60, 2, 105-122
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
Johnson, Michael D. (1981) Context effects in product perception Advances in
Consumer Research. Vol. 8, Kent B. Monroe, ed. Chicago, IL: Association for
Consumer Research, 112-115
(1986) Consumer similarity judgments: a test of the contrast model Psychology
and Marketing. 3, 47-60
Kahle, Lynn R, and Pamela M. Homer (1985) Physical attractiveness of the celebrity
endorser: a social adaptation perspective Journal of Consumer Research. 11, 4,
954-961
Kamins, Michael A., and Kamal Gupta (1994) Congruence between spokesperson and
product type: a matchup hypothesis perspective. Psychology & Marketing, 11,6,
569-586

198
Kardes, Frank R. (1988) Spontaneous inference processes in advertising : the effects of
conclusion omission and involvement on persuasion. Journal of Consumer Research. 15,
2, 225-233
Keller, Kevin Lane, and David A. Aaker (1992), The effect of sequential introduction of
brand extensions. Journal of Marketing Research, 29, 1, 35-50.
Keller, Kevin Lane (1993) Conceptualizing, measuring, and managing customer-based
brand equity. Journal of Marketing. 57, 1, 1-22
Kirmani, Arnna (1990) The effect of perceived advertising costs on brand perceptions.
Journal of Consumer Research. 17, 2, 160-171
and Peter Wright (1989) Perceived advertising expense and expected product
quality. Journal of Consumer Research. 16, 3, 344-353
Laforet, Sylvie, and John Saunders (1994) Managing brand portfolios: how the leaders
do it. Journal of Advertising Research. 64, 5, 64-76
Landy, David, and Harold Sigall (1974) Beauty is talent: task evaluation as a function of
the performers physical attractiveness. Journal of Personality & Social
Psychology, 29, 3, 299-304
Leonard, Myron, and Austin Spencer (1991) The importance of image as a competitive
strategy: an exploratory study in commercial banks. International Journal of Bank
Marketing. 9, 4, 25-29
Lindberg, Charles A. (1953) The spirit of St. Louis. New York: Charles Schribners Sons
Linder, Darwyn E., and Stephen Worchel (1970) Opinion change as a result of efifortfully
drawing a counterattitudinal conclusion. Journal of Experimental Social
Psychology, 6, 4, 432-448
Loken, Barbara, and Deborah Roedder John (1993) Diluting brand beliefs: when do
extensions have a negative impact? Journal of Marketing. 57, 3, 71-84
Macchiette, Bart., and Abhijit Roy (1992) Affinity marketing: what is it and how does
it work? Journal of Services Marketing. 6, 3, 47-57
MacKenzie, Scott B. and Richard J. Lutz (1989) An empirical examination of the
structural antecedents of attitude toward the ad in an advertising pretesting
context. Journal of Marketing. 53, 2, 48-65

199
, Scott B., Richard J. Lutz, and George E. Belch (1986) The role of attitude
toward the ad as a mediator of advertising effectiveness : a test of competing
explanations Journal of Marketing Research. 23, 2, 130-143
Markman, Arthur B., and Dedre Gentner (1993) Structural alignment during similarity
comparisons Cognitive Psychology. 25, 4, 431-467
Marshall, D. W., and G. Cook (1992) The corporate (sports) sponsor. International
Journal of Advertising. 11, 307-324.
McCracken, Grant (1989) Who is the celebrity endorser? cultural foundations of the
endorsement process. Journal of Consumer Research, 16, 3, 310-321
McDonald, Colin, (1991) Sponsorship and the image of the corporate sponsor.
European Journal of Marketing. 25,11,31-38
Meenaghan, Tony (1991a) The role of sponsorship in the marketing communications
mix. International journal of Advertising. 10, 35-47
(1991b) Sponsorship-legitimizing the medium European Journal of Marketing.
25, 11, 5-10
(1994) Ambush marketing: immoral or imaginative practice? Journal of
Advertising Research. 34, 5, 77-88
Meerabeau, Elizabeth, Roy Gillett, Michael Kennedy, Johnson Adeoba, Michael Byass,
and Kingsley Tabi (1991) Sponsorship and the drinks industry in the 1990s.
European Journal of Marketing. 25,11, 39-56
Mervis, Car olyn B., and Eleanor Rosch (1982) Categorization of natural objects.
Annual Review of Psychology. 32, 89-115
Meyers-Levy, Joan, and Alice M. Tybout (1989) Schema congruity as a basis for product
evaluation. Journal of Consumer Research. 16, 1, 39-54
Misra, Shekhar, and Sharon E. Beatty (1990) Celebrity spokesperson and brand
congruence: an assessment of recall and affect. Journal of Business Research. 21,
2, 159-173
Murphy, Gregory L., and Douglas L. Medin (1985) The role of theories in conceptual
coherence. Psychological Review. 92, 3, 289-316
Nebenzahl, Israel D., and Eugene D. Jafife (1991) The effectiveness of sponsored events
in promoting a countrys image. International Journal of Advertising. 10, 223-237

200
Nicholls, J.A.F., Henry A. Laskey, and Sydney Roslow (1992) A comparison of
audiences at selected hallmark events in the united states. International Journal of
Advertising. 11, 215-225
OConnor, Robert (1996) Thumbs up for cobranding Credit Card Management. 9, 6,
150-154
Ortony, Andrew (1979) Beyond literal similarity. Psychological Review. 86, 3, 161-180
Otker, Ton (1988) Exploitation: the key to sponsorship success European Research.
May, p.77-86
Park, C. Whan, Bernard J. Jaworski, and Deborah J. Maclnnis (1986) Strategic brand
concept-image management. Journal Of Marketing. 50, 4, 135-146
, Sandra Milberg, and Robert Lawson (1991) Evaluation of brand extensions:
the role of product feature similarity and brand concept consistency Journal of
Consumer Research. 18, 2, 185-193
, Bernard J. Jaworski, and Deborah J. Macinnis (1986) Strategic brand concept-
image management. Journal of Marketing. 50, 4, 135-145
, Sung Youl Jun, and Allen D. Shocker (1996) Composite branding alliances: an
investigation of extension and feedback effects Journal of Marketing Research.
33, 4, 453-466
Parker, Ken (1991) Sponsorship: the research contribution. European Journal of
Marketing. 25, 11, 22-30
Petty, Richard E., John T. Cacioppo, and Rachel Goldman (1981) Personal involvement
as a determinant of argument-based persuasion. Journal of Personality & Social
Psychology, 41, 5, 847-855
Pham, Michel Tuan (1993) Cue representation and selection effects of arousal on
persuasion Journal of Consumer Research. 22, 4, 373-387
Pracejus, John W. (1995) Is more exposure always better? Effects of incidental exposure
to a brand name on subsequent processing of advertising Advances in Consumer
Research. Vol. 22, Frank Kardes and Mita Sujan, eds. Provo, UT: Association for
Consumer Research, 319-322
Proctor, James (1996) Why L.A. Gear is relying on sponsorship as the first step in its
repositioning. Paper presented at the 1995 IEG Event Marketing Conference,
Chicago Illinois.

201
Rajaretnam, J. (1994) The long term effects of sponsorship of corporate and product
image: findings of a unique experiment. Marketing and Research Today. 22, 1,
62-74
Rangaswamy, Arvind, Raymond R. Burke, and Terrence A. Oliva (1993) Brand equity
and the extendibility of brand names. International journal of Research in
Marketing. 10, 61-75
Rao, Akshay A., and Robert W. Ruekert (1994) Brand alliances as signals of product
quality. Sloan Management Review. Fall, 87-97
Reddy, Srinivas K., Susan L. Holak, and Subodh Bhat (1994) To extend or not to
extend: success determinants of line extensions Journal of Marketing Research.
31 (May), 243-262
Rosch, Eleanor (1976) Basic objects in natural categories. Cognitive Psychology, 8, 3,
382-439
Roslow, Sydney, J.A.F Nicholls, and Henry A. Laskey (1992) Hallmark events and
measurement of reach and audience characteristics. Journal of Advertising
Research. 32, 4, 53-59
Rosner, Hillary (1995) Dream team's sisters. Brandweek. 36, 42, 24-26
Ryssel, Christian, and Erich Stamminger (1988) Sponsoring world-class tennis players
European Research. May. 110-116
Sanbonmatsu, David M. and Frank R. Kardes (1988) The effects of physiological
arousal on information processing and persuasion Journal of Consumer Research.
15, 3, 379-385
Sandler, Dennis, and David Shani (1989) Olympic sponsorship vs. ambush marketing:
who gets the gold? Journal of Advertising Research. 29, 4, 9-14
Sawyer, Alan G. and Danial J. Howard (1991) Effects of omitting conclusions in
advertisments to involved and uninvolved audiances. Journal of Marketing
Research. 28, 4, 467-474
Scott, D. R., and Hazel T. Suchard (1992) Motivations for australian expenditure on
sponsorship-an analysis. International Journal of Advertising. 11, 325-332
Sedikides, Constantine, Nils Olsen, and Harry T. Reis (1993) Relationships as natural
categories. Journal of Personality & Social Psychology. 64, 1,71-82

202
Simon, Carol J. and Mary W. Sullivan (1993) The measurement and determinants of
brand equity: a financial approach. Marketing Science. 2, 1, 28-52
Sissors, Jack Z. and Jim Surmanek (1986) Advertising media planning Lincolnwood, IL:
NTC Business Books
Shavitt, Sharon (1990) The role of attitude objects in attitude functions. Journal of
Experimental Social Psychology, 26, 124-148
Sheinin, Daniel A, and Bemd H. Schmitt (1994) Extending brands with new product
concepts: the role of category attribute congruity, brand affect, and brand
breadth. Journal of Business Research. 31, 1-10
Smith, Daniel C., and C. Whan Park (1992) The effect of brand extensions on market
share and advertising efficiency Journal of Marketing Research. 29, 3, 296-313
Smith, Edward E. and Douglas L. Medin (1981) Categories and concepts. Cambridge,
MA: Harvard University Press.
Speck, Paul S., David W. Schumann, and Craig Thompson (1988) Celebrity
endorsements scripts, schema, and roles: theoretical framework and preliminary
tests Advances in Consumer Research. Vol. 15 Michael J. Houston, ed. Provo,
UT: Association for Consumer Research, 69-76
Stipp, Horst, and Nicholas P. Schiavone (1996) Modeling the impact of Olympic
sponsorship on corporate image. Journal of Advertising Research. 36,4, 22-28
Sunde, Lorraine, and Roderick J. Brodie Consumer evaluation of brand extensions:
further empirical results. International Journal of Research in Marketing. 10,
47-53
Thomas, Howard (1985) Sponsorship, an advertisers guide. International Journal of
Advertising. 4, 319-326
Thwaites, Des (1995) Professional football sponsorship-profit or profligate.
International Journal of Advertising. 14, 149-164
Tversky, Amos (1977) Features of similarity. Psychological Review. 84, 4, 327-352
and Itamar Gati (1982) Similarity, separability, and the triangle inequality.
Psychological Review. 89, 2, 123-154
Varadarajan, Rajan P., and Anil Menon (1988) Cause related marketing: a coalignment of
marketing strategy and corporate philanthropy. Journal of Marketing. 52, 3, 58-
74

203
Walker, Mary, Lynn Langmeyer, and Daniel Langmeyer (1992) Celebrity endorsers: do
you get what you pay for? Journal of Consumer Marketing. 9, 2, 69-76
Walshe, Peter and Peter Wilkinson (1994) Pompeii revisited: IBM digging for success.
Marketing and Research Today. 22, 1, 89-96
Weisz, Pam (1995) G-P's Angel Soft gets kid-conscious. Brandweek. 6, 39, 4
Westbrook, Robert A., and William C. Black (1985) A motivation-based shopper
typology. Journal of Retailing. 61, 1, 78-103
Williams, Michael (1996) Gen X files: how sony uses sponsorship to reach previously
unattainable goals. Paper presented at the 1995 IEG Event Marketing
Conference, Chicago Illinois.
Withcer, B., G. Craigen, D. Culligan, and A. Harvey (1991) The link between objectives
and functions in organization sponsorship International Journal of Advertising.
10, 1, 10-33
Wolton, Claude (1988) Arts sponsorship: harmony or discord? European Research.
May, 87-96
Worthington, Steve, and Suzanne Home (1992) Affinity credit cards: card issuer
strategies and affinity group aspirations. International Journal of Bank
Marketing. 10, 7, 3-10
Wright, Peter (1985) Schemer schema: consumers intuitive theories about marketers
influence tactics in Advances in Consumer Research. Vol. 13, ed. Richard J.
Lutz, ed. Provo, UT: Association for Consumer research, 1-3
Wright, Rodger (1988) Measuring awareness of British football sponsorship. European
Research. May, p.104-108
Zaichkowsky, Judith Lynne (1985) Measuring the involvement construct. Journal of
Consumer Research. 12, 3, 341-352.
Zajonc, Robert B. (1968) Attitudinal effects of mere exposure. Journal of Personality
and Social Psychology 9, 2, 1-27.
Zinkhan, George M., Jae W. Hong and Robert. Lawson (1990) Achievement and
affiliation motivation: changing patterns in social values as represented in american
advertising. Journal of Business Research. 20, 2, 135-143

BIOGRAPHICAL SKETCH
John W. Pracejus was bom in Bedford, Ohio, a suburb on the southeast of
Cleveland. He was raised with three younger sisters, Joan, Anne, and Sue by his father
Walter, a civil engineer, and his mother Rosemary, a primary school teacher. He spent
his first eight years of school at St. Marys of the Immaculate Conception where he
formed many friendships that last to this day.
He graduated from Bedford Senior High School where he was a member of the
swim team, ski club and National Honor Society. He earned two bachelors degrees from
Miami University, one in business and one in communications. He then worked selling
advertising space and video production services before deciding to return to college to
pursue a masters degree at the University of Illinois. It was there that he was exposed to
the stimulating environment of an active research faculty. To learn more about a career
in academic research and teaching, he asked the advice of Tom OGuinn, who
encouraged John to pursue a Ph.D. in marketing. After completing his masters degree at
Illinois, John joined the University of Florida as a doctoral student. He is currently an
Assistant Professor of Marketing at the University of Alberta in the city of Edmonton.
204

I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and
quality, as a dissertation for the degree of Doctor of Philosophy
Richard J. Lutz, Chairman
Professor of Marketing
I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and
quality, as a dissertation for the degree of Doctor of Philosophy
Barton A. Weitz
J. C. Penney Eminent Scholar
of Marketing
I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and
quality, as a dissertation for the degree of Doctor of Philosophy
Associate Professor of Marketing
jCCi.'i, .'Csl
Christopher^. Jarus^ewski
I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and
quality, as a dissertation for the degree of Doctor of Philosophy
Michael F. Weigold
Associate Professor of
Journalism and Communications

This dissertation was submitted to the Graduate Faculty of the Department of
Marketing in the College of Business Administration and to the Graduate School and
was accepted as partial fulfillment of the requirements for the degree of Doctor of
Philosophy.
August 1998
Dean, Graduate School



178
Quick Tennis Quiz
(If not sure, make your best guess)
1 When a player hits a ball into the net, or outside the court, or fails to hit the ball before
it bounces twice, his/her opponent .
2 Returning the ball over the net before it bounces is called a .
3 For what percentage of the game does the starting server serve %
4 To serve, must the server stand behind the base line or the service line (circle one)?
5 When serving from the right, must the server hit the ball into the left or the right
service court (circle one)?
6 A serve that is not good is called a ?
7 A serve that touches the top of the net and lands in the proper service court is called a
?
8 Is it OK for the receiver to return a serve before it bounces once? Yes No (circle one)
9 A score of zero or nothing is called .
10-A score of 30-30 is called Deuce. True False (circle one)
11 You must win at least games to win a set.
12 In championship matches of big tournaments, you must win sets to win the
match.


CHAPTER TWO
SPONSORSHIP WITHIN A BROADER CATEGORY OF COMMERCIAL
ASSOCIATIONS
As alluded to in the introduction, sponsorship is not unique as a commercial
endeavor which seeks to associate two things. The core idea behind brand extension, for
example, is that a new brand can sometimes benefit from an association with an existing
brand. Likewise for celebrity endorsers, an underlying assumption is that the brand can
sometimes benefit from being associated with a human being. Other areas of study in
marketing which involve commercial associations include co-branding (OConnor, 1996;
Carpenter, 1994), the use of one brand as an ingredient in another (Park, Jun and
Shocker, 1996), brand alliances (Rao and Ruekert, 1994), and retail brand assortment
(Laforet and Saunders, 1994; Bawa, Landwehr and Krishna, 1989; Bultez et al., 1989).
There is even anecdotal evidence of the importance of customers as commercial associates
in terms of providing prestige (Smith and Park, 1992).
The two areas which have received the most attention in terms of understanding
the impact of such associations on consumers are the brand extension and celebrity
endorsement literatures. While brand extensions and celebrity endorsers share with
sponsorship issues related to association, they also differ in some conceptually important
ways. The goal of this chapter is to fit sponsorship into a larger framework of commercial
associations. Due to the different assumptions about the underlying mechanisms of brand
12


91
TABLE 1
MEAN PERCEIVED PRESTIGE AND SIZE OF 111 EVENTS AND
ORGANIZATIONS
Eywit/OnuntaBon
fiat
Pristine
1
1
c
1
fia
Pristine
m2
Praatiaa
AMA Proffering
4.00
3.33
Florida State Fat
2.88
2.62
Pittsburgh Three Rives Regatta
289
4.X
AMC Cancar Research Center
5.36
6.00
Heed of tvs Chales Regsito
3.00
3.43
Poiarts Amphitheater
267
3.33
American Cancer Satisfy
6.11
6.76
Henry Ford Mutaun
3.44
4.10
Poxfcys Rating Te^n
200
1.X
American Diabetes Assoczcn
5.94
6.57
Houston Uvestodc Show
2.72
Z29
Power Team
267
2.67
American Hsert Ctitf
5.17
6.00
IMAX Corporation
4.50
3.95
Raserarch Hospital /ALSAC
4.S6
5.14
American Horae Shows Acsce.
2.44
3.62
IndyCar
5.00
4.14
Ringing Bra. Grajo
6.56
4.10
American LumbaJeck
1.86
229
inf^ Open LA (Tennis)
4.06
5.14
Rockdeilar Canter
5.61
6.14
American PooJpiaycra
222
2.81
Intemstocd Fcoilvd of lighto
4.22
4.19
Rocky Mountain Bk Fcund^cn
3.11
3.52
American Quarter Hcm>
2.36
3.43
Jeffrey Ballet
Z67
4.71
Roys! Academy of Arts
3.44
4.48
American Racing Series
3.61
3.33
Junior Adhlov'smcnt Inc.
3.78
3.81
San Diego Street Scene
2.X
2.76
Arthritis Foundation
5.36
5.52
Kennedy Center for Perf. Arts
4.56
5.48
San Diego Zoo
3.83
4.81
Baltimore Zoo
4.11
4.43
Kennedy Krieger institute
3.33
4.43
Sat Fmtdsco Jazz Festive
3.X
4.43
Big Apple Cira*. Ltd.
3.36
3.33
KO Drugs Boring
3.36
3.14
Science Museum of Minnesota
2.78
4.X
Brookfield ZooObcs^o
3.86
420
Lttdcnd Prowlers Hockey Teem
2.56
Z43
Shea's Performing Arts Caite-
3.17
4.33
Builders Squsr Alamo 0c .71
4.28
3.57
Lbarty Soane* Carter
2.86
3.X
Shed Aquarium
2.94
3.67
Calgary Stenpato
2.83
2.62
nctri Pssrk Zoo Chicago, IL
3.22
4.62
SL Marys Food Bank
4.28
4.71
Calgary Zoo
3.50
4.24
Martin of tornos
5.61
6.X
lEsiecfCckrado
239
2.81
Canadian Figure Skating Aaaoc.
322
4.62
Mary Bridge Children's Hospital
4.11
5.29
TadeGattoy
3.17
4.10
Canadian Olympic Aaaoddtan
4.33
5.67
Ete&te! Museum of Rna Arts
2.86
4.76
Taam Danmark DENMARK
Z44
3.81
Capital City RtariTaat
3.00
3.62
Multiple Sclerosis Society
4.78
5.X
The ChMerms Thtoter
3.44
4.X
Carnegie Hal
5.86
6.48
Museum of Science & Industy
4.X
4.X
Tho Nsiur Conafflvuncy
4.X
4.X
Chertctt Motor Speedway
4.44
3.00
Mystic Seaport Museum
3.11
3.71
Tcdttbe Tannis Ctasaic
3.72
5.X
Cherry Creak Arte Festival
2.94
3.62
Names
3.17
3.86
US. 10K Ctasaic
ZX
3.76
CMdrVs Home Sodoty
3.50
4.57
Nd Assoc, of tntsrcoQ. Athletics
4.78
4.81
U.S. Luge Association
Z94
3.33
Childrens Hospital o Alabama
4.11
4.71
Nal AsBsoc.Seccndsry Schools
322
3.20
U.S. Taekwondo Union
Z78
3.38
Children's Hospital of Pittsburgh
4.67
5.06
National Captioning Institute, Inc.
ZX
3.43
United States Sailing
3.17
4.10
Children's Miracle Network
6.17
6.00
NCAA
6.X
5.71
Untied States Termte Association
4.61
5.29
Children's Museum of todtonopoHs
3.67
4.14
National Geographic Society
5.44
5.38
Untied Way
6.39
5.X
Child's Play Touring Trieste
3.33
4.24
National Wiidife Federation
5.17
5.24
University Musical Society
3.61
4.52
Cirque duSoia
3.22
4.57
Nsw Jaraay Bdtoon Fes&v^
Z18
Z57
USA shooting
2.94
3.24
City of Hope
3.39
4.62
NMMA Bod Tour
Z22
3.X
USA Wresting
4.X
3.19
Caches v. Cancer
5.00
5.38
Nynet
4.X
4.14
Vancouver Art GaNery
2.78
429
Cranbroek Institu of ScteK
3.00
4.81
Optyland
2.67
3.29
Venetian Festival
289
4.X
Downtown District Houston. TX
3.50
3.43
QrotgeBowfl
6.22
5.X
Walt Dianeys World on ice
522
4.57
Bvis Preey^ Gracaiand
3.44
3.52
Qr^on hGJtespsi? Fodivd
267
3.71
Washington DC Paries
3.00
4.29
Epilepsy Fcundslon o America
4.72
5.81
Osfem America
239
3.14
Western Idaho Fair
272
Z24
Feed the Children
5.17
5.76
PedAc Nrtional Exhtotiicn
283
3.24
Wildlife Conservation Society
4.00
4.52
The idea that sponsoring an event can influence the perceived size of the sponsor
has been examined and has received some empirical support. McDonald (1991) did not
find evidence of such an effect. This may have been due to the fact that perceived size of
companies studied was already very high (e.g., Phillips and Cadbury). Evidence that
sponsorship can influence perceived size was found by Rajaretnam (1994) who reports
that a tire maker that spent almost its entire marketing budget on sponsorship
dramatically increased its perceived size.


145
experience (Hoch and Deighton, 1989). Experience ambiguity was defined as the degree
to which a consumer could know quality immediately after purchase.
Pretest
Forty-three products were pretested for experience ambiguity along a nine point
scale anchored by you can perfectly know quality immediately after trial and you can
never fully know quality. Thirty-three pretest subjects evaluated the 43 brands for
perceived experience ambiguity. These categories and the mean experience ambiguity
ratings are provided in Table 7. The high experience ambiguity product class chosen was
auto insurance (mean=4.0). The low experience ambiguity category chosen was yogurt
(mean=2.33). This difference is statistically significant (F=8.59, p<.01).
Because of the radically different attributes in the two product categories to be
examined, it is not recommended that these two categories be jointly modeled. Study 3a,
therefore, investigates the impact of perceived size and perceived event facilitation on auto
insurance choice. Study 3b investigates the impact of these variables on yogurt choice.
Discussion of these two studies is left until after the results section of study 3b.
Study 3a
Overview
Subjects made a series of eight choices between two brands of auto insurance (952
total choices). Each choice profile consisted of five brand attributes (including size and
event facilitation) for each of two brands. These attributes were manipulated orthogonally
through a fractional factorial design. Each subject completed half (8 cells)of this 16 cell


43
The ways in which implied endorsement can impact perceptions, attitudes, and
choice, have been well delineated in the source effects literature. Factors affecting the
ability of a source to persuade include trustworthiness (Brehm and Lipsher, 1959),
expertise (Petty, Cacioppo and Goldman, 1981), attractiveness (Landy and Sigall, 1974),
and credibility (Bimbaum, Wong and Wong, 1976). While the literature on source effects
is rich with predictive constructs, proposing the impact of these constructs in the context
of sponsorship first requires some empirical investigation. For example, it is not known
whether a credible or trustworthy event exists in the minds of consumers. If such things
do exist, it is not obvious what the determinants of these source characteristics would be.
Further elaboration on the role of source characteristics in sponsorship, therefore, requires
prior empirical investigation.
Reciprocity
Reciprocity is the mechanism which requires the highest level of cognitive
elaboration. The message of reciprocity is the sponsor supports events you care about,
so you should patronize the sponsor. The reciprocity mechanism is engaged when a
consumer makes a conscious decision to go out of her way to support the brands that
support the events which she cares about. While this may not be the most common
response to sponsorship, it is certainly a powerful one.
Surprisingly, there is evidence to support that fans of certain events demonstrate
significant reciprocity (or so they claim). Crimmins and Horn (1996) present what they
call gratitude whereby the fan of an event goes out of his way to buy the products of an
event sponsor. They cite a survey of NASCAR fans of whom 48% say they would


REFERENCES
Aaker, David A (1990) Brand extensions: the good, the bad and the ugly. Sloan
Management Review. Summer, 47-57
(1991) Managing brand equity: capitalizing on the value of a brand name. New
York: The Free Press
and Kevin Lane Keller (1993) Interpreting cross-cultural replications of brand
extension research. International Journal of Research in Marketing. 10, 55-59
and Kevin Lane Keller (1990) Consumer evaluations of brand extensions
Journal of Marketing. 54, 1, 27-41
and Donald E. Bruzzone (1985) Causes of irritation in advertising. Journal of
Marketing. 49, 2, 47-57
Abratt, Russel, Brian C. Clayton, and Leyton F. Pitt (1987) Corporate objectives in
sports sponsorship International Journal of Advertising. 6, 299-311
Alonzo, Vincent (1994) Misuse of koran k.o.'s world cup promo. Incentive. 168, 8, 8
Andrews, Jim (1996) Sponsorship overview: what you need to know. Paper presented
at the 1995 IEG Event Marketing Conference, Chicago Illinois.
Armstrong, Clive (1988) Sports sponsorship: a case-study approach to measuring its
effectiveness. European Research. May, p.97-103
Barsalou, Lawrence W. (1982) Context-independent and context-dependent information
in concepts. Memory & Cognition. 10, 1, 82-93
(1983) Ad hoc categories. Memory & Cognition. 11, 3, 211-227
(1985) Ideals, central tendency, and frequency of instantiation as determinants
of graded structure in categories. Journal of Experimental Psychology: Learning,
Memory. & Cognition. 11. 4. 629-654 =
193


75
result in poor evaluation and even outrage is equally attractive. When combined, the two
notions form a complex and largely unexplored area in marketing. This area is referred to
here as sacrilege.
The notion that consumption objects can take on dimensions of being sacred or
profane has received some attention in the consumer behavior literature. Belk,
Wallendorf and Sherry (1990) describe how the processes of the secularization of religion
and the sacralization of the secular are at work in contemporary society. They find that
consumers maintain the sacred in consumption through separation of sacred and profane,
and through the performance of rituals. Belk and Wallendorf (1990) explore the sacred
and profane dimensions of money itself. They find that the sacred and profane aspects of
money depend upon its sources and its uses.
Hirschman (1991a) explores the way advertisers utilize sacred and profane
consumpti on imagery to persuade. In an exploration of process of the secularization of the
sacred, Hirschman (1991b), describes how the commercial transactions involved with
surrogate motherhood and in-vitro fertilization can be made less profane. She finds that
this is achieved by stressing the sacred aspects of the acts of childbirth and family creation.
While the sacred and profane nature of consumption objects has received attention,
there has been no examination of these issues in the context of commercial associations.
Some might argue that words like sacred, profane and sacrilege are too strong to be
used in the context of commercial associations. There are several recent examples,
however, which, seem to meet the more traditional notion of sacrilege. DeCoursey (1995)
reports on the widespread feeling of over-commercialization associated with the Popes
1995 visit to the United States. Brewer (1993) describes attempts to associate the Pilot


160
sponsorship impacts behavior. This dissertation is the result of an attempt to address these
issues.
More specifically, the purpose of this dissertation has been to explore the variety
of ways in which sponsorship works to enhance the equity of the sponsoring brand. In
particular, it has investigated how engaging in sponsorship influences various inferences
about the brand, and how these inferences impact brand choice. It began by presenting a
review of extant sponsorship research, which was found to be largely descriptive and
speculative in nature. A summary of relevant research from related commercial
association domains, like celebrity endorsement and brand and line extension followed.
A conceptual framework for commercial associations in general was then
developed. The similarities in findings from celebrity endorsement, brand extension and
other commercial association literatures were presented. Likely implications for
sponsorship of the common findings of these literatures were developed into formal
propositions. The unique nature of sponsorship, and its likely impact on brand equity and
positioning, was then provided. The marketing variables which may be affected by
sponsorship were examined, and specific proposals about the impact of sponsorship on
brand equity and positioning were presented.
The specific mechanisms through which sponsorship works were then considered.
Exploratory research and managerial intuition as to how sponsorship works were
developed into seven testable mechanisms through which sponsorship may function in
various situations. In approximate order of necessary cognitive elaboration, these
mechanisms were posited to be simple awareness, affect transfer, image transfer,
affiliation, implied size, implied endorsement, and reciprocity. The psychological


54
F44: Given peoples low motivation to form associated brands and events
into categories, attribute fit is determined primarily by overlap in chronically
available attributes of the two associates.
Having explored several potential antecedents of attribute fit, possible properties
of attribute fit are now considered. As mentioned above, the attribute fit construct is quite
close in definition to Tverskys similarity. It therefore seems appropriate to consider
other properties of similarity which make sense in the context of attribute fit. One
seemingly important fact is that Tversky (1977) generally argues that similarity should not
be treated as symmetrical. He proposes and demonstrates, for example, that North Korea
is perceived to be more similar to China than China is to North Korea. This is an example
of a more general idea, that the variant is more similar to the prototype than the
prototype is to the variant, because the prototype is generally more salient than the
variant (Tversky, 1977, p.333). This has several implications for the perceived fit
between sponsor and event.
P45: When the brand is larger (more salient) than the event, fit is perceived
to be greater when the association is stated in the form of the event fits with
the brand than the brand fits with the event
P46: When the event is larger (more salient) than the brand, fit is perceived
to be greater when the association is stated in the form of the brand fits
with the event than the event fits with the brand.
It should be noted that similar results have been obtained in a marketing context.
Johnson (1981) finds Shasta was judged to be more similar to Coke than Coke was to
Shasta. Here again, asymmetries are demonstrated such that an object with many


114
subjects exposed to low fit associations(mean=.31) Some caution must be used, however,
before attempting to interpret this result (see study lb).
Discussion
Results of study la are quite encouraging. Support was found for HI (large event
leads to large brand inference), and H2 (long duration leads to large brand inferences).
The finding that aspects of a sponsorship association can significantly impact perceived
size of the brand is totally new in an experimental setting. If replicable, this finding should
be quite interesting for brand categories in which large size has a positive impact on choice
probability (see chapter 10).
H5 (large event leads to brand legitimacy(quality) inferences) was also supported.
Somewhat surprisingly, this effect was not moderated by brand/event fit. It made no
difference whether there was good or poor fit with the event. For both sports drink and
coffee, sponsoring a large running event led subjects to infer greater quality than
sponsoring a smaller one.
H10 (long duration leads to event facilitation inferences) received weak support.
It certainly makes sense that associating ones brand with an event for a long period of
time would lead people to infer greater facilitation on the part of the sponsor. Here again,
and counter to prediction, fit did not moderate the effect. For both sports drinks and
coffee, sponsoring a running event for a long time led to more facilitation inferences than
did being a new sponsor.
The open ended data yielded considerable evidence of spontaneous inference
making about a sponsoring brand. The number of each of the four specific inferences,


151
TABLE 9
IMPACT OF FIVE MANIPULATED VARIABLES ON BINARY LOGIT MODEL OF
YOGURT CHOICE
Parameter
Estimate
SE of
Parameter
Asymptotic
t-Stat
Pr(Z>|t|)
intercept
.10824559
. 72024387E-01
1.503
.1329
price
85104661E-01
.61957660E-02
-13.736
.0000
calories
-.21349918E-01
.20470950E-02
-10.429
. 0000
facilitation
.16591370
. 48397839E-01
3.428
.0006
company size
.31744675
.10012614
3.170
.0015
shelf life
.10856916
. 14422567E-01
7.528
.0000
As can be seen in the table, all five of the attributes had a significant impact on
choice. As before, the sign of the parameter estimate indicates direction of impact on
choice. Therefore, in this data set, as price and calories increase, choice probability
decreases. As event facilitation, company size, and shelf life increase, choice probability
increases. In the yogurt category, both event facilitation and company size had a
significant, positive impact on choice. Large companies and those that facilitated events
were preferred.
This pair of studies has shown that the impact of sponsorship-based inferences can
differ by brand category. Both studies manipulated company size and event facilitation.
Perceptions of both of these attributes have been shown in previous studies to be
influenced by sponsorship. In the auto insurance category (study la), company size had a
measurable impact on choice. Event facilitation had no measurable impact.
For the yogurt category, however, both event facilitation and company size had a
significant impact on choice. These findings lend support to the idea that specific
sponsorship-based inferences can have different effects on choice in different brand


CHAPTER ONE
EXTANT SPONSORSHIP RESEARCH: WHERE ARE WE NOW?
Since 1984, there have been at least 35 journal articles which deal, either directly
or indirectly, with sponsorship. Hundreds more can be found in the trade press. Despite
these facts, very little work beyond describing the occurrence frequency of sponsorship
activities has taken place. What has gone beyond description of frequency often involves
the reporting of levels of awareness, attitude or purchase intention following a
sponsorship. A few tangential topics have also received some attention. Finally, there
have been a good number of papers which put forth some ideas of how sponsorship
works. For the most part, though, these papers rely entirely upon anecdotal evidence for
support of their assertions. These assertions, which come largely from papers written by
practitioners and consultants, however, do provide a wealth of insight into potential
mechanisms through which sponsorship may work. The systematic development of such
mechanisms follows in a subsequent chapter.
One of the questions most frequently addressed in the sponsorship literature is
why do companies engage in sponsorship? Abratt, Clayton, and Pitt (1987) found that
among 60 sponsoring organizations in South Africa, 11% of respondents claimed that
their sponsorship was altruistic to a large extent, 53% said it was altruistic to some
4


198
Kardes, Frank R. (1988) Spontaneous inference processes in advertising : the effects of
conclusion omission and involvement on persuasion. Journal of Consumer Research. 15,
2, 225-233
Keller, Kevin Lane, and David A. Aaker (1992), The effect of sequential introduction of
brand extensions. Journal of Marketing Research, 29, 1, 35-50.
Keller, Kevin Lane (1993) Conceptualizing, measuring, and managing customer-based
brand equity. Journal of Marketing. 57, 1, 1-22
Kirmani, Arnna (1990) The effect of perceived advertising costs on brand perceptions.
Journal of Consumer Research. 17, 2, 160-171
and Peter Wright (1989) Perceived advertising expense and expected product
quality. Journal of Consumer Research. 16, 3, 344-353
Laforet, Sylvie, and John Saunders (1994) Managing brand portfolios: how the leaders
do it. Journal of Advertising Research. 64, 5, 64-76
Landy, David, and Harold Sigall (1974) Beauty is talent: task evaluation as a function of
the performers physical attractiveness. Journal of Personality & Social
Psychology, 29, 3, 299-304
Leonard, Myron, and Austin Spencer (1991) The importance of image as a competitive
strategy: an exploratory study in commercial banks. International Journal of Bank
Marketing. 9, 4, 25-29
Lindberg, Charles A. (1953) The spirit of St. Louis. New York: Charles Schribners Sons
Linder, Darwyn E., and Stephen Worchel (1970) Opinion change as a result of efifortfully
drawing a counterattitudinal conclusion. Journal of Experimental Social
Psychology, 6, 4, 432-448
Loken, Barbara, and Deborah Roedder John (1993) Diluting brand beliefs: when do
extensions have a negative impact? Journal of Marketing. 57, 3, 71-84
Macchiette, Bart., and Abhijit Roy (1992) Affinity marketing: what is it and how does
it work? Journal of Services Marketing. 6, 3, 47-57
MacKenzie, Scott B. and Richard J. Lutz (1989) An empirical examination of the
structural antecedents of attitude toward the ad in an advertising pretesting
context. Journal of Marketing. 53, 2, 48-65


147
TABLE 7
PERCIEVED EXPERIENCE AMBIGUITY OF SELECTED PRODUCT
CATEGORIES
Product Category mean stdev
razors 2.03 1.79
icecream 2.09 2.08
frozen dinner 2.18 2.19
beer 2.33 1.98
deodorant 2.33 1.90
yogurt 2.33 2.26
athletic shoes 2.52 2.08
encyclopedia 2.64 2.36
carwax 2.70 1.90
cellular phone 2.70 2.24
film 2.70 1.94
shampoo 2.79 1.96
airline 2.88 1.98
hand tools 2.91 2.08
paint 2.94 1.90
cielingfan 3.00 2.19
exercise equip. 3.03 2.32
personal watercraft 3.03 1.85
DVD 3.06 2.01
computer 3.30 2.31
charcoal 3.42 2.60
car 3.48 2.44
Internet service 3.48 2.02
long distance 3.52 1.99
music club 3.52 2.15
credit card 3.58 2.25
condom 3.61 2.41
cable 3.70 1.91
cell phone service 3.73 2.23
bank 3.82 2.04
battery 3.91 2.31
medical insurance 3.97 2.39
car insurance 4.00 2.36
motor oil 4.24 2.19
gasoline 4.30 2.19
investment company 4.30 1.83
dogfood 4.33 2.37
mutual find 4.36 2.29
renters insurance 4.45 2.24
car battery 4.52 2.40
catfood 4.55 2.20
sparkplugs 4.61 1.97
life insurance 5.06 2.38


11
anecdotal evidence drawn from colorful examples (Hastings, 1984; Meenaghan, 1991b;
Thomas, 1985; Crowley, 1991; Parker, 1991; McDonald, 1991; Otker, 1988; Ryssel and
Stamminger, 1988; Schoch, 1994; Walshe and Wilkinson, 1994). They have titles like
Professional Football SponsorshipProfit or Profligate (Thwaites 1995) and Arts
sponsorship: Harmony or discord? (Wolton 1988). While there is little formal conceptual
development, construct definition or theory, they do contain a good deal of intuition about
the processes involved. These papers are drawn upon heavily for examples of constructs
proposed in the following chapters


ACKNOWLEDGEMENTS
I would first like to thank my chair, Rich Lutz, for guidance, support, and for
allowing me to find my own way. I would also like to thank Chris Janiszewski for
helping me to think more conceptually about the topic, Bart Weitz for helping me to
focus on my strengths, and Mike Weigold for helpful comments and suggestions. Special
thanks go to Joffre Swait for providing me with a new perspective on this project and
research in general, and for tremendous help with development and analysis of the choice
studies.
Others to whom I am grateful for contributing to my intellectual development
include .Alan Sawyer, Joel Cohen, Joe Alba, John Lynch, and especially Tom OGuinn,
without whom I would never have started this process.
I would also like to acknowledge all of the doctoral students who have made my
life at Florida the wonderful experience it has been. Many thanks go to Luk Warlop,
Michel Pham, Susan Fournier, and Prasad Niak for providing maps which allowed me to
successfully follow the path, and to Norma Mednoza, Andre Menck, and Stijn Van
Osselaer, for being so supportive while walking that path with me. Thanks also go to
Corinne Faure, Francis Hollman Anne Stringfellow, Stacy Wood, Americus Reed, Lisa
Bolton, Kevin Bradford, Tom Meyvis and Velitchka Kaltcheva for ideas, talk and coffee.
Finally, I am thankful to my parents, Walter and Rosemary, without whom none
of this would have been possible.
iii


41
While both Kirmani (1990) and Kirmani and Wright (1989) report evidence of an
inverted U relationship between perceived advertising cost and perceived quality,
whereby beyond a certain point, quality perceptions actually go down. The data reported
in both cases could, however, simply reflect diminishing returns. Asymptotic functions
could have produced the significant quadratic terms reported. Also, no means or least
squared means were reported which might have shown the true shape of the
expenditure/quality relationship.
While it is not difficult to imagine situations under which perception of firm size
would have an impact on perceived quality, very little research has directly investigated
this effect. Within the context of the banking industry there is evidence that perceived size
of a bank can affect consumer beliefs about the likelihood of receiving a loan (Leonard and
Spencer, 1991). In the arena of legal services, however, there is evidence to suggest that
service providers believe perceived firm size to be more important than do clients
(Gaedeke and Tootelian, 1988).
Given the results of this research, it can be proposed that
P31: The sponsorship impact on perceived size has a positive impact on
perceived quality of the brand.
Implied Endorsement
The implied endorsement mechanism requires the most cognitive elaboration on
the part of the consumer, save reciprocity. The message of the implied endorsement
mechanism is this brand must be OK, or people wouldnt participate in an event named


62
that freely elicited attributes of the incongruent compound include many attributes which
are not elicited by either simple category. They suggest that evaluation of incongruent
combinations may be based upon a more complex judgment process than evaluation of
simple categories.
Put another way, their results seem to indicate that when incongruent things are
paired, evaluation of the pair is extreme, and attribute elicitation about the pair includes
attributes which are not elicited by either of the paired items alone. Assuming the pairing
of event with brand engages similar mechanisms,
P62: Lack of attribute fit between sponsor and event leads to more extreme
evaluation of the association along trait adjective scales.
P63: Lack of attribute fit between sponsor and event leads to the elicitation
of attributes which are not elicited for either the event or sponsor alone.
P64: Evaluation of high lack-of-attribute-fit sponsorships requires consumers
to engage in more complex judgment processes than do sponsorships which
have better attribute fit
Finally, it is important to qualify all of the assumptions of a linear relationship
between attribute fit and consequences in light of findings which suggest more complex
relationships. Meyers-Levy and Tybout (1989), for example, demonstrate that products
that are somewhat congruent with their associated category schemas will stimulate
processing that leads to a more favorable evaluation relative to products that are either
congruent or extremely incongruent. In addition to the preexisting notion that very
incongruent pairings are unnatural, weird, and/or poorly liked, this research suggests the
possibility that extremely similar pairings are seen as mundane or boring, and therefore,


3
various situations. These mechanisms are simple awareness, affect transfer, image
transfer, affiliation, implied size, implied endorsement, and reciprocity. Constructs which
are hypothesized to have an impact upon the functioning of these mechanisms are then
developed. These constructs include various types of fit between sponsor and event, as
well as sacrilege or consumer outrage at an association which is perceived to be
inappropriate. A theoretical discussion of how other communications can be deployed to
maximize the impact of the sponsorship association wraps up the background and theory
chapters.
Three empirical studies of how the event and associational factors (event size,
brand-event fit, duration of association and share of presence) of a sponsorship impact
consumer perceptions of unfamiliar brands are presented in chapter eight. Chapter nine
reports the results of two studies which investigate the role of audience characteristics
(involvement with and knowledge about the event) on consumer response to a pair of
familiar brands. In chapter ten, the impact of sponsorship-based inference on brand choice
is investigated for several categories of brands. Chapter 11 provides an overview of the
findings of chapters eight through ten and suggests some interesting areas for further study
in this emerging area.


154
TABLE 10
PREFERENCE FOR DOMESTIC VS. FOREIGN BRANDS
Cateaorv
mean*
stdev
credit card
1.94
1.30
car insururance
2.03
1.47
medical insurance
2.24
1.84
long distance
2.27
1.51
life insurance
2.30
1.49
bank
2.36
1.88
cable
2.48
1.92
renters insurance
2.48
1.62
cell phone service
2.61
1.69
airline
2.70
1.81
investment company
2.73
1.82
encyclopedia
2.79
1.90
mutual fund
2.79
1.82
frozen dinner
3.18
1.83
internet service
3.30
1.83
yogurt
3.30
1.79
music club
3.42
1.75
condom
3.48
1.91
deodorant
3.58
1.84
ice cream
3.70
2.02
athletic shoes
3.76
2.05
gassoline
3.82
1.96
car battery
4.06
1.90
battery
4.09
2.11
dogfood
4.18
1.53
catfood
4.24
1.70
film
4.27
2.11
shampoo
4.27
1.51
razors
4.36
1.54
tools
4.45
1.80
spark plugs
4.64
1.45
motor oil
4.67
0.99
carwax
4.76
1.32
charcoal
4.76
1.09
exercise equip.
4.85
1.72
ceiling fan
4.88
1.14
paint
4.91
1.13
computer
5.03
2.32
personal watercraft
5.27
1.75
cellular phone
5.67
2.15
beer
6.24
2.22
DVD
6.33
2.10
car
6.48
2.50
1= Always prefer American; 9=Always prefer foreign
As can be seen in the table, auto insurance is a category in which American brands
are almost always preferred. This means that, relative to a domestic company, a foreign


174
Please answer the following:
How many events do you think EVIAN (TEXACO) sponsors each year?
Very few events I 1 I | 1 1 1 [ Lots of events
Compare to other waters, how many events do you think EVIAN (TEXACO) sponsors
Far fewer than others | | 1 | | | 1 | Many more than others
Overall, how large do you perceive the US Open (Indy 500) to be
a very small event I I I I I 1 I I a very large event
Overall, how would you rate the fit between EVIAN (TEXACO) and the US Open
Very poor fit I I I I 1 1 I I very good fit


64
dominance). The authors also point out the implications of different levels and types of
dominance in the domain of brand extension. They point out that many researchers either
ignore directionality or look at the wrong associative direction. The authors also hint at
other brand aissociates, such as usage situations, benefits and attributes, but do not discuss
the implications of these associations in any detail.
They then discuss three methods for assessing category dominance. They first
describe naming methods, which involve a subject naming as many brands in a category as
possible in some short time, or naming some pre-specified number of brands. Brands
named, and especially named early-on are said to have high category dominance. Such
methods could be applied to sponsorship by asking subjects to list events sponsored by
X or brands sponsoring Y, using unaided recall as a surrogate for fit.
Farquhar and Herr (1992) also describe latency methods, which use reaction time
measures to determine category dominance. According to the authors, brands which are
quickly and correctly identified as members of the category have high category dominance.
Here a slight modification of the measurement method might be appropriate for
sponsorship. Cueing a consumer with an event, and measuring his/her reaction time and
accuracy to whether or not a brand is a sponsor would be a fairly good method of
measuring the spontaneous availability component of attribute fit. Reaction time
measures of this type have been demonstrated to discriminate between spontaneously
available properties and context dependent properties (Barsalou, 1982, experiment one).
Finally, the authors discuss facilitation methods, which involve providing
subjects with a category, then presenting them with a brand covered up with dots. The
authors claim that brands which are recognized earlier are category dominant. In the


137
between tennis and experiences in my life., tennis is of high personal relevance to me,
and tennis helps me express who I am
Subjects then listed the number of times they had played and watched tennis in the
last week, month, and six months. Finally, subjects took a 12 question quiz about the
rules of playing and scoring tennis. This represented the objective knowledge measure.
Results
Manipulation Check
On the same scale used in previous studies, the high fit brand (mean=5.65) was
rated as having better fit with the event than the low fit brand (mean=2.89) This
difference is statistically significant (F=49.69, p<0001)
Close Ended Data
The overall analysis strategy was to determine if fit impacted inference making,
and if so, if any of the audience factors moderated the relationship. Fit had no measurable
impact on inferences about the commercial motive of the sponsor (F=1.82, p>. 15). It did
however, have a significant, positive impact on event facilitation inferences (F=6.64,
p<02). The high fit brand (Evian) was perceived as doing more to help the event than
was the low fit brand (Texaco).
A series of ANCOVAs were then run to test which, if any, of the audience factors
might moderate this relationship. These models revealed no moderation by frequency of
watching, frequency of playing, ISPR, PH, or felt involvement. Significant moderation
was found, however, for both subjective (F=4.15, p<05), and objective (F=12.89,


29
but not limited to, greater attention paid to subsequent commercial communications (or
less counterarguing, see Pracejus, 1993), and greater chance of inclusion in the
consideration set. While there is little doubt that such things take place, to end here with
awareness ignores the rich possibilities involved in the association. In short, if simple
awareness were the only mechanism in operation, then it would, for example, make no
difference whether a brand name was seen on the side of a race car or on the side of a
bam.
As mentioned above, there is little doubt that awareness functions as an important
mechanism in sponsorship. Examples of simple awareness, however, where there is little
or no chance of cognitive elaboration are not easy to cite. One which comes close is the
practice of soft drink bottlers paying for convenience store signage, provided it
prominently displays the appropriate brand logo. This example meets the criterion for
simple awareness because of the low likelihood of elaboration by a store patron as to the
origins of the large Coke logo on the ABC Beverage sign. Even if there were some
attribution on the part of a customer, it is unlikely that he would have enough involvement
with the store to care that the owners received a new sign courtesy of the local Coke
bottler. The most likely impact (if any) is that the customer may have high top of mind
awareness of Coke as he faces his soft drink selection. There is little or no difference
between the logo appearing on the sign or on the side of a nearby building. The only thing
accomplished is simple awareness of the brand name.
Despite the relative simplicity of this mechanism, two basic propositions can be
offered. Given the reasonably well established mere exposure effect,


167
pointed out that the inferences recorded in the thought listing task were the result of a
single, 60 second exposure to a poster for an event. If compared to multiple pre-event
promotional exposures, and/or several hours of actual event viewing or attendance, the
percentage of people drawing inferences about a sponsoring brand in these experiments
could easily be underrepresented.
Determining the actual amount of inference making generated by sponsorship is
key to using the inference-based model to assess the actual value of a sponsorship for a
particular brand. Techniques similar to those employed in 3a,3b,and 3c should be able to
evaluate the change in choice probability once a certain inference is made. Being able to
calculate what percentage of consumers exposed to an association will make each
inference would allow precise calculations of the value of a sponsorship to be made. Some
combination of experimental evidence (to establish causality) and correlational field
research (to establish magnitude) must be brought together in order to make reasonable
estimates of the number of consumers who will draw certain inferences when exposed to
an association.
Once methods for making such estimations exist, calculating the value of a
sponsorship to a brand becomes a straightforward endeavor. The percentage of people
drawing an inference can be multiplied by an estimate of change in choice probability given
the inference. One possible form for such an evaluative function is provided in Appendix
I.


CHAPTER FIVE
DIMENSIONS OF FIT.
The better the fit of all elements, the better the chance that the sponsorship is
effective, because the relations between the sponsor and sponsored activity are
logical, natural, and easy to understand Otker (1988, p. 83)
The notion of fit is seemingly ubiquitous in the extant sponsorship
literature.(Roslow, and Laskey,1992; Thomas,1985; Crimmins and Hom,1996; Otker,
1988; Walshe and Wilkinson, 1994). This chapter attempts to define fit between an event
and a brand. Generic or respondent-reported perceived fit requires no development or
investigation of elaborate antecedents and component structures. It is simply a measure of
whatever a respondent thinks it is. Operationally, generic fit is measured by a subjects
responses to a question like how good is the fit between event X and brand Y. While
this measure may be a bit crude, the importance of quickly getting a handle on the fit
construct makes it useful.
A brief review of how the fit construct has been presented in sponsorship and other
commercial association domains is reviewed. Other, more technically defined types of fit,
such as attribute fit, audience fit and positioning fit are then discussed and defined.
The chapter concludes with a discussion of measurement issues.
Most sponsorship research refers to fit as the degree to which an association
makes sense. Walshe and Wilkinson (1994), for example, report that IBMs
45


146
design. While the fractional design employed does not allow testing of all higher order
interactions, it is extremely efficient. The full factorial design associated with this study
would have 65,536 cells!
Subjects
One hundred nineteen undergraduates in an introductory marketing class at a
major southeastern university took part. Each received course extra credit for
participation.
Independent Variables
Five independent attribute variables were manipulated orthogonally. They were
price (four levels), deductible (two levels), company size (two levels), event facilitation
(two levels), and hours of customer service (two levels). Each is described below (see
Appendix E).
Price level was determined by measuring perceived auto insurance cost. Pretesting
was used to find a range of perceived prices for six months auto insurance coverage.
Twenty-two subjects estimated cost for six months insurance coverage. The mean
estimate was $570.45, and the standard deviation was $89.52. The four manipulated price
levels were set at the mean minus one half standard deviation ($525), the mean ($570), the
mean plus half a standard deviation ($615) and the mean plus one standard deviation
($660). The amount of the deductible was manipulated at two levels. The low deductible
level was set at $500. The high deductible level was set at $1000.


TABLE OF CONTENTS
page
ACKNOWLEDGEMENTS iii
ABSTRACT vii
INTRODUCTION 1
CHAPTER ONE EXTANT SPONSORSHIP RESEARCH:
WHERE ARE WE NOW? 4
CHAPTER TWO SPONSORSHIP WITHIN A BROADER
CATEGORY OF COMMERCIAL ASSOCIATIONS 12
CHAPTER THREE THE POTENTIAL IMPACT OF
SPONSORSHIP ON BRAND EQUITY AND POSITIONING 21
CHAPTER FOUR SEVEN MECHANISMS THROUGH WHICH
SPONSORSHIP MAY FUNCTION 27
Simple Awareness 28
Affect Transfer 30
Image Transfer 32
Affiliation 36
Implied Size 38
Implied Endorsement 41
Reciprocity 43
Conclusion 44
CHAPTER FIVE DIMENSIONS OF FIT 45
Attribute Fit 49
Audience Fit 65
Positioning Fit 70
CHAPTER SIX SACRILEGE 74
CHAPTER SEVEN ASSOCIATION-ENHANCING
COMMUNICATIONS (AEC 80
iv


197
Hildreth, C. H., and Bernard C. Nalty (1969) 1001 questions about aviation history,
Dodd, Mead and Company, New York, 162-169
Hirschman, Elizabeth C. (1991a) Sacred, secular, and mediating consumption imagery in
television commercials. Journal of Advertising Research. 30, 6, 38-43
(1991b) Babies for sale: market ethics and the new reproductive technologies.
Journal of Consumer Affairs, 25, 2, 358-390
Hoek, Janet, Phililp Gendall, and Mark Stockdale (1993) Some effects of tobacco
sponsorship advertisements on young males. International Journal of Advertising.
12, 25-35
Horia, Joann, Nicholas Naccari and Elliot Fatoullah (1974) The effect of expertise and
physical attractiveness upon opinion agreement and liking Sociometrv. 37, 4,
601-606
Hovland, Carl I. and Wallace Mandell (1952) An experimental comparison of
conclusion-drawing by the communicator and by the audience. Journal of
Abnormal & Social Psychology, 47 581-588
Jacoby, Jacob, and David Mazursky, (1984) Linking brands and retailer images-do the
potential risks outweigh the rewards. Journal of Retailing 60, 2, 105-122
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
Johnson, Michael D. (1981) Context effects in product perception Advances in
Consumer Research. Vol. 8, Kent B. Monroe, ed. Chicago, IL: Association for
Consumer Research, 112-115
(1986) Consumer similarity judgments: a test of the contrast model Psychology
and Marketing. 3, 47-60
Kahle, Lynn R, and Pamela M. Homer (1985) Physical attractiveness of the celebrity
endorser: a social adaptation perspective Journal of Consumer Research. 11, 4,
954-961
Kamins, Michael A., and Kamal Gupta (1994) Congruence between spokesperson and
product type: a matchup hypothesis perspective. Psychology & Marketing, 11,6,
569-586


129
TABLE 5
SUMMARY OF SUPPORT FOR THE MAIN EFFECTS HYPOTHESES IN STUDIES
1A, IB, AND 1C
Study
la
Study
lb
Study
lc
open
ended
Any
Support
HI Large event leads to large brand inference
**
**
*
YES
H2 Long duration leads to large brand inference
**

**
YES
H3 Large share leads to large brand inference

NO
H5 Large event leads to legitimate brand
inference
**
***
YES
H6 Long duration leads to legitimate brand
inference
NO
H7 Large share leads to legitimate brand
inference

NO
H9 Small event leads to event facilitation
inference
NO
H10 Long duration leads to event facilitation
inference
*

YES
HI 1 Large share leads to event facilitation
inference

**
YES
*=p< 10 **=p< 05 ***=p<01 = Not Tested
As the table makes clear, the hypothesis receiving the most robust support is HI.
In all three studies, evidence was found that associating a brand with a larger event leads
people to believe that it is a larger brand. The other hypothesis to receive support in
multiple studies is H5. Associating a brand with a larger event appears also to improve
perceptions of quality. These results seem to indicate that event size does matter. Being
associated with large events can lead to at least two positive inferences about the brand.
Duration of association also was found to impact inference making. Using pooled,
open ended data, a main effect of duration on perceived brand size was found. This result
is especially interesting because it was obtained without cueing subjects about any brand
attribute. The inferences were spontaneously generated. The impact of duration on


28
Increase in required cognitive elaboration
( Simple J f*? ( JT ( affiliation( implied Y^A
\ awareness^ transfer transfer l size V endorsement J J
FIGURE 1
RELATIONSHIP BETWEEN THE SEVEN MECHANISMS
The only of these mechanisms to receive significant direct acknowledgment are
simple awareness, affect transfer and, to a lesser extent, image transfer. Evidence for the
existence of the affiliation, implied size, implied endorsement, and reciprocity mechanisms
is largely intuitive and/or anecdotal. The seven mechanisms are described below.
Propositions related to the main effects of these mechanisms are also provided. Potential
moderators of these mechanisms are provided in later chapters.
Simple Awareness
For who is going to buy a product or service from a company that one has never
heard of? Wright (1988, p. 104)
Awaireness, as a mechanism through which sponsorship is likely to function, is
widely accepted by practitioners and academics alike. A typical explanation of how this
mechanism might function is as follows. A consumer goes to an event or watches it on
TV and is exposed to the brand name. Since repeated exposure to an object has been
found to lead to positive affect toward the object (Zajonc 1969), the consumer feels better
about the brand. This good feeling then has all sorts of benefits to the brand, including,


26
are important to customers. By his definition, there is no ideal position for all brands.
Positioning must be based upon advantages of the brand. He contends that these
advantages should be communicated through a consistent, integrated communications
strategy involving brand name, packaging, and advertising. While he does not specify a
role for sponsorship, the idea that positioning should be used primarily for differentiation
from competition suggests that to the extent to which a sponsorship represents a point of
differentiation from competitors, it will be an effective positioning tactic.
This can be formally stated as:
P17: Sponsorship will have a greater impact on perceived brand positioning
when there is less competitor sponsorship of similar events.
This chapter has examined the marketing variables which may be affected by
sponsorship. Specific proposals about the impact of sponsorship on brand equity and
positioning were presented. The mechanisms through which sponsorship can effect these
variables, however, was not addressed. This issue is covered in the following chapter.


188
EXPERIMENT 1C ALL HIGH
The Extram Euro-Capital
International Marathon
Brussels, Belgium
Sunday October 19th 1997 at 9:30 am
Proud to be a Sponsor for 5 Years


20
the association can be perceived as either an anomaly or a separate category. With
regard to sponsorship, therefore, it can be proposed:
P13: Effects of an event/hrand association can impact the event as well as
the brand.
The goal of this chapter was to provide a conceptual framework for commercial
associations. The similarities in findings from celebrity endorsement, brand extension and
other commercial association literatures were presented. Likely implications of these
common findings for sponsorship were given as formal propositions. Overall, this chapter
has demonstrated the ways in which sponsorship is like other commercial associations.
The unique aspects of sponsorship and their likely impact on brand equity and positioning
are provided in the following chapter


CHAPTER THREE
THE POTENTIAL IMPACT OF SPONSORSHIP ON BRAND EQUITY AND
POSITIONING.
As stated in the introduction, with sponsorship, the association is the message.
There is no argument with which to measure agreement, no claims, attributes or benefits
on which to measure recall. It seems appropriate, therefore, to examine marketing
variables which may be affected by sponsorship and which do not rely entirely upon formal
argument. Two such variables are brand equity and brand positioning. The goal of this
chapter, therefore, is to investigate what sponsorship can do. It does not deal with how it
does it nor when it will work best. These issues are addressed in subsequent chapters.
While the complex associative mechanisms through which sponsorship may impact
brand equity have received little attention to date, the act of associating a brand with a
socially intrusive activity possessed of its own personality in the eyes of the receiving
audience (Meenhaghan, 1991b, p. 8) should certainly be able to impact the things one
associates with a brand. This is relevant because several authors have conceived of brand
equity as consisting largely of brand associations.
For instance, Keller (1993) describes what he calls customer-based brand equity.
He defines this as the differential effect of brand knowledge on consumer response to the
marketing of the brand (p. 1). The keys to this effect, according to Keller, are familiarity
with the brand and holding favorable, strong, and unique brand associations in memory
21


191
EXPERIMENT 2B HIGH FIT
Win a Trip to
The 1998 Indy 500
May 23rd-25th 1998
At the Indianapolis Motor Speedway
From
See Store for Details
Texaco is proud to be an official sponsor of the 1998 Indy 500


196
Elliot, Stuart (1995) Tanqueray pulls out the stops for bike event to benefit AIDS,
patients and Mr. Jenkins rides along New York Times, July 24.
Farquar, Peter H. (1989) Managing brand equity. Marketing Research. 1, 24-33
and Paul M. Herr (1992) The dual structure of brand association. In Building
Strong Brands. David A Aaker and Alexander L. Biel, eds. Hillsdale, NJ:
Lawrence Erlbaum Associates, 263-277
, Julia Y. Han, Paul M. Herr, and Yuji Ijiri (1992) Strategies for leveraging
master brands. Marketing Research. 4, 32-43
Frieden, JonB. (1984) Advertising spokesperson effects: an examination of endorser
type and gender on two audiences Journal of Advertising Research. 24, 5, 33-41
Friedman, Hershey H. and Linda Friedman (1979) Endorser effectiveness by product
type. Journal of Advertising Research, 19, 5, 63-71
Friestad, Marian, and Peter Wright (1995) Persuasion knowledge: lay people's and
researcher's beliefs about the psychology of advertising. Journal of Consumer
Research. 22, 1, 62-74
Gaedeke, Ralph M., and Dennis H. Tootelian (1988) Understanding how clients select
and evaluate law firms. Journal of Professional Services Marketing. 3, 4, 199-207
Gardner, Meryl Paula, and Phillip Joel Shuman (1987) Sponsorship: an important
component of the promotions mix. Journal of Advertising. 16,1,11-17
Gray, Edmund R. and Larry R. Smeltzer (1985) Corporate image-an integral part of
strategy. Sloan Management Review. Summer, 73-78
Hansen, Flemming, and Lene Scotwin (1995) An experimental inquiry into sponsorship:
what effects can be measured? Marketing and Research Today. 23, 5, 173-181
Hastie, Reid, Colin Schroeder and Renee Weber (1991) Creating complex social
conjunction categories from simple categories Bulletin of the Psvchonomic
Society. 28, 3, 242-247
Hastings, Gerald, B. (1984) Sponsorship works differently from advertising.
International journal of Advertising. 3, 171-176
Heider, F. (1946) Attitudes and cognitive organization. Journal of Psychology. 21, 107-
112


96
previously mentioned, there is some evidence from both the sponsorship literature
(Rajaretnam 1994), and other literatures (Kirmani 1990), that perceived expenditure can
influence perceived firm size. Within the sponsorship context, perceived brand/firm size
inferences are hypothesized to be driven by perceived financial ability to sponsor. The
idea is that the association cues some response that the brand must have substantial
resources to engage in such an association. It can therefore be predicted:
HI: Sponsoring larger events leads to more inferences on the part of consumers
that the brand is large.
H2: Longer duration of association between brand and event leads to more
in ferences on the part of consumers that the brand is large.
H3: Engaging in a sponsorship with a large, as opposed to small, share of
presence leads to more inferences on the part of consumers that the brand is
large.
Given that inferences about brand/company size are expected to be driven strictly by
perceived ability to engage in sponsorship, the congruence between brand and event
should not impact perceived size. Therefore:
H4: Brand/event fit does not moderate the relationships in H1-H3
Because of the prediction of no effect inH4, a liberal significance value is
necessary to test it credibly (i.e. p>. 15).


51
F36: When attributes which overlap between event and brand are context
dependent, exposing consumers to the relevant context increases attribute fit.
Sponsorship associations are likely to be a context-dependent property initially,
but may, over time become context-independent. Initially, for example, Miller beer may
be a context-dependent property of NASCAR. At this initial stage, one would expect
Miller to cue NASCAR only when the sponsor is also cued, perhaps while one is watching
a race. As the association becomes stronger (as might be the case for a serious fan who is
spontaneously aware that Miller sponsors his favorite driver), Miller might become a
context independent property of NASCAR. In this latter case, one of the spontaneous
associations that comes to mind when this person free-associates about Miller beer would
be NASCAR.
It therefore seems possible that
P37: High involvement with an event leads to greater spontaneous awareness
for the brand-event association.
P38: Frequent attendance at an event leads to greater spontaneous awareness
for the brand-event association.
P39: Recent attendance at an event leads to greater spontaneous awareness
for the brand-event association.
In another study, Barsalou (1985) found that central tendency did not predict
graded structure in goal-derived categories. Goal-derived categories are categories which
are formed for a specific purpose, like assembling things to sell at a garage sale. From this
example, it is clear that goal derived categories are often categories which are formed in


152
categories. Therefore, managers need to determine the attribute perceptions which impact
choice in the category before setting sponsorship strategy.
It is also possible that brand specific factors might affect the impact sponsorship-
based inferences have on choice. That is, within a category, certain inferences might
impact some brands more than others. To investigate this possibility, a third study was
run, where the two brands in the same category differed along a relevant dimension.
Study 3c
Overview
Subjects made a series of eight choices between two brands of auto insurance (984
total choices). Each choice profile consisted of the same five brand attributes at the same
levels as in 3a. Once again these choices were made between two brands. The two brands
used in this study differed, however, in that one was a domestic auto insurance company,
and the other was a Japanese auto insurance company. The five attributes were
manipulated orthogonally, using the same fractional design as in 3 a. As before, each
subject completed half (8 cells) of the 16 cell design.
Subjects
One hundred twenty-three undergraduates in an introductory marketing class at a
major southeastern university took part. Each received course extra credit for
participation.