|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
INFLUENCE OF PRODUCT-ENDORSER MATCH-UP ON
CONSUMER' S PURCHASE INTENTIONS OF
(NON-SPORT) ENDORSED PRODUCTS
JESSICA ROBIN BRAUNSTEIN
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
Jessica Robin Braunstein
This document is dedicated to my family.
I want to thank my committee, the sport management faculty, my advisor/chairs,
my classmates/colleagues, and especially my friends. Most importantly, I wish to
acknowledge my family. I want to thank those by my side today and those that are no
longer with us. They have all taught me more than they will ever realize. Their
unwavering support has allowed me to take chances and let my heart to guide me along
this crazy journey. Without their love, support, and encouragement this would not have
been possible or worth it! They are my rocks, my best friends, and my biggest
cheerleaders. They have been my inspiration and my tour guides. Without them I would
not be where I am or, more importantly, who I am today. I could not have accomplished
this without any of the people in my life, and for that I am forever grateful.
TABLE OF CONTENTS
ACKNOWLEDGMENT S .............. .................... iv
LI ST OF T ABLE S ............. ..... ._ .............. vii...
LIST OF FIGURES ............. ......___ ..............viii...
AB STRAC T ................ .............. ix
1 INTRODUCTION ................. ...............1.......... ......
Athlete Endorsers............... .. ...............
Influence on Purchase Intentions ................. ...............3................
Theory Development .............. ...............5.....
Purpose .................. .......... .... ...............11.......
Significance of the Study ................. ...............13.......... .....
Delimitations ................. ...............15...............
Limitations ................. ...............15.................
Definition of Terms .............. .....................16
Overview ................. ...............17.................
2 REVIEW OF LITERATURE ................. ......... ...............18. ....
Introducti on ................. ...............18.................
Attitude Theory............... ...............24.
Perceived Value ................. ...............27.................
End orser Effectiveness .............. ...............29....
Match-Up Hypothesis............... ...............3
Match-Up: Characteristics............... ............3
Product characteristics............... ............3
Endorser characteristics................... ..........3
Endorser-Product Congruency/Match-Up ................. .............................40
Identity Theory ................ .........................4
Identity Theory and Sport Consumption ................ ......___ ........._ ....46
Identity Theory and Points of Attachment ............. ......__. ............. ...47
Hero W orship .............. .............................4
Endorser-Consumer Congruency/Match-Up ................ .......... ................49
Sum m ary ................. ...............52.......... ......
3 M ETHOD .............. ...............54....
Participants .............. .. ...............54.
Questionnaire Development .............. ...............55....
Points of Attachment Index ............... .. .. ......... .. ...............56....
Scale of Athletic Star Power Consumer' s Perspective ................. ................. 57
Perceived Value Scale ................ ...............59........... ....
Validity Screening ................. ...............60.................
Athlete and Product Selection .............. ...............61....
Proc edure s ................ ...............62........... ....
Data Analyses ................. ...............63.......... .....
4 RE SULT S .............. ...............65....
Measurement Model Analysis .............. ...............65....
Structural Model Analysis .............. ...............72....
5 DI SCUS SSION ........._.__....... .__. ...............75...
A ITEM CORRELATIONS FOR THE MEASUREMENT MODEL.............._._. ........89
B SPORT MARKETING SURVEY ................. ...............94......__._....
LIST OF REFERENCES ................. ...............97..............
BIOGRAPHICAL SKETCH ................. ...............107....... ......
LIST OF TABLES
1. Definition of Terms. ........... ........... ...............16....
2. Descriptive Statistics for the Demographic Variables ........._._. .........._._............55
3. Descriptive Statistics for the Measurement Model ...._._._._ ... .....____ ........._.....66
4. Factor Loadings for the Measurement Model............... ...............68.
5. Factor Correlations for the Measurement Model ....._.._................. ............_... .71
6. Relationships in the Structural Model............... ...............73.
LIST OF FIGURES
1. Influence of Product-Endorser Match-Up on Consumer' s Purchase Intentions of
(Non-Sport) Endorsed Products .............. ...............12....
2. Product-Endorser Advertisement ................. ...............62........... ...
3. Structural M odel .............. ...............74....
4. Recommended Model for Future Research .............. ...............88....
Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
INFLUENCE OF PRODUCT-ENDORSER MATCH-UP ON
CONSUMER' S PURCHASE INTENTIONS OF
(NON-SPORT) ENDORSED PRODUCTS
Jessica Robin Braunstein
Chair: James J. Zhang
Cochair: Galen T. Trail
Major Department: Tourism, Recreation and Sport Management
The practice of using athletes to endorse sport and non-sport products has
increased drastically since Miller Lite' s highly successful "Tastes Great, Less Filling"
campaign in the 1970's. As a result of media coverage and increased social visibility, star
athletes have embraced their celebrity status and benefited financially from endorsing
products. Numerous studies have indicated that a star athlete's association with a brand
may help to define and enhance the brand's image; however, negative characteristics of
an endorser could also have a deleterious effect (Horrow, 2002; Pitts & Stotlar, 2002). To
a great extent, the success of an endorsement depends on product-endorser congruency.
The purpose of this study was to develop a model to examine the relationships among
identification with an athlete and his/her sport (Robinson & Trail, 2005), product-
endorser congruency (Match-Up = Expertise, Trustworthiness, and Image; Aaker, 1997;
Braunstein & Zhang, 2005a, 2005b; Tenser, 2004), perceived value of the product
(Emotion, Quality, Price, and Social; Sweeney & Soutar, 2001), and consumer purchase
intentions. Participants (N= 400) were college students, who responded to an online
questionnaire that measured their perception of product-endorser congruency,
identification with the athlete and/or her sport, perceived value of the product, and
purchase intentions after viewing an advertisement with Maria Sharapova endorsing a
Canon Power Shot digital camera. A confirmatory factor analysis revealed a lack of
discriminant validity among the first-order latent variables for Match-Up; thus, an
adjustment was made to allow all items to load directly on the general Match-Up factor.
One Perceived Value subscale (Emotion) was eliminated from further analyses because
of a lack of discriminant validity with the Quality subscale. The structural model showed
adequate fit of the model, with the largest amount of variance being explained by the
relationships of Match-Up to Perceived Value (3 8%) and of Perceived Value to Purchase
Intention (52%). Identification (both Athlete and Sport) was found to have a small
influence on Match-Up, with only 7% of the variance explained. The Einal model
provides preliminary information on socio-psychological factors that influence the
purchase intentions of endorsed products, and can be used as a reference by corporations
when choosing athlete endorsers.
As the sport market has grown, so have the various mediums that make it possible
to be a sport spectator. Through various magazines, newspapers, video games, television
shows, and the Internet, it is easy for individuals in different regions of the country, or the
world, to become familiar with a variety of different sports and, in turn, a wide array of
athletes. As a result of the high skill-level of professional athletes and the media attention
that sport receives, consumers are aware of many athletes' abilities and achievements.
However, when distinguishing one athlete from another, it is the athlete's individual
characteristics that set him or her apart from the others (Brooks & Harris, 1998; Chalip,
1996; Howard, 1979; Stevens, Lathrop, & Bradish, 2003). Some athletes create a name
for themselves based on athletic talent alone, while others make an impact on the
industry, both positive (Cornwell, Roy, & Steinard, 2001; Javalgi, Traylor, Gross, &
Lampman, 1994; McDonald, 1991) and negative (Horrow, 2002), with their attitudes and
actions on and off the playing field. Many athletes receive public attention, but some
bridge the gap between sport star and general celebrity. A celebrity, or an individual
whose name has the ability to attract attention and interest, has the potential to influence
consumers as a result of an individual's response to that celebrity (Rein, Kotler, &
Stoller, 1997). Athletes that have an elevated level of celebrity may also have star power,
made up of a star athlete's overall characteristics, both ability and personality (Braunstein
& Zhang, 2005a).
As a result of the metamorphosis of athletes into celebrities, the use of an
identifiable and reliable athlete or coach to endorse a product may encourage a bond to
form between the product and consumers. Advertisers take advantage of the loyalty that
forms as a result of an individual's allegiance to the product (Wansink & Ray, 2000) and
the athlete (Braunstein & Zhang, 2005a). James (2002) discussed the importance of
matching a product and an endorser. Typically, the effectiveness of an endorser increases
by matching the qualities of the athlete to those of the product (Brooks & Harris, 1998;
Chalip, 1996; Howard, 1979). These characteristics would then be reinforced as traits of
both the endorser and the product in the mind of the consumer. Choosing the right athlete
for the product enforces this congruence. Well-known pairs include the groupings of
Tiger Woods (golf) and Nike, Mia Hamm (soccer) and Gatorade, and Shaun White
(skateboarding and snowboarding) with Mountain Dew. The traits of these athletes are
widely viewed as similar with the traits of the products that they promote. This fit, or
match-up, is the perceived congruency between the characteristics of the endorser and the
product. Corporations are looking for a pair that fits in the eyes of the consumer. The
consumers that organizations generally target with these relationships are typically the
youngest generation with purchasing power (Holton, 2000). Presently, Generation Y, or
individuals approximately 13-27 years-old, is the market segment that is influenced the
most by the use of athletic stars in promotional campaigns (Rein et al., 1997; Stevens et
al., 2003; Stotlar, Veltri, & Viswanthan, 1998). Since this is the age range within which
people typically form lasting product preferences, corporations are trying to create
endorser-product matches that solidify their products) as the product of choices) for the
consumer in the future.
Marketers create endorser-product matches with the intention that consumers will
understand this relationship when recalling their productss. However, there are
exceptions to this rule. There are instances in which an athlete acts in a manner that is not
congruent with the product. This is an inherent risk when a corporation becomes involved
with any celebrity as an endorser. Although the product may not be involved, if an athlete
makes controversial decisions in his or her personal life, the bad publicity that the athlete
receives may transfer to negative exposure for the product that he or she endorses (Pitts
& Stotlar, 2002; Stevens et al., 2003). Such instances may affect the individual as well as
any products) that he/she is associated with, becoming a liability to the products) (Till
& Shimp, 1998). A controversial figure may create an initial impact on consumer
purchase intentions; yet, some researchers believe that it is in the best interest of the
marketer to align a product with an athlete that maintains the values of the company and,
in turn, the product itself (Kahle & Homer, 1985; Kamins, 1989, 1990). It is important for
marketers to be aware of the factors that influence a consumer' s purchase intentions,
including the potential influence of athlete endorsers.
Influence on Purchase Intentions
Marketers often use celebrity endorsers in advertising campaigns as a tool to
influence consumer's purchase intentions (e.g., Kamins, 1990; Ohanian, 1991). Attitude,
or an individual's feelings for or against the attributes of an obj ect (Fishbein & Ajzten;
1975), is developed through a process comprised of three phases. The phases (cognition,
affect, and conation; e.g., Lavidge & Steiner, 1961; Palda, 1966; Ray, 1973) follow an
individual's unique process as he or she thinks about the product (cognition), develops
feelings for or against the product (affect), and the behavior intentions that follow
(conation). Studies have shown that attitudes may be altered through the techniques used
to market a product. Specifically, advertisements have been found to mediate a
consumer' s perception of a brand (e.g., Lutz, 1975; MacKenzie & Lutz, 1989;
MacKenzie, Lutz, & Belch, 1986). In order to understand the influence of athlete
endorsers, the consumer' s level of identification toward both the athlete and the sport in
which the athlete participates must be observed (Fink, Trail, & Anderson, 2002; Kwon,
Trail, & Anderson, 2005; Robinson & Trail, 2005; Robinson, Trail, & Kwon, 2004; Trail,
Robinson, Dick, & Gillentine, 2003; Wann, 2002). Identification, stemming from identity
theory, examines the "parts of a self composed of the meanings that persons attach to the
multiple roles they typically play in highly differentiated contemporary societies"
(Stryker & Burke, 2000, p. 284). Therefore, identity theory examines the individual's
perception of his or her role in different social contexts, not the general social
categorization that society places on specific roles. While this has not been studied on an
individual level, or in the context of examining the influence of an individual athlete on a
consumer' s purchase intentions outside the realm of sport, it is an integral part of the
relationship that may help to explain a consumer' s purchasing rationale. Once the level of
influence that these factors exert has been examined, how the individual will perceive the
fit of the pair based on the characteristics of both can be observed (e.g., Aaker, 1997;
Austin et al., 2003; Azoulay & Kapferer, 2003; Erdogan, Baker, & Tagg, 2001; Fink,
Cunningham, & Kensicki, 2004; Goldsmith, Lafferty, & Newell, 2000; Hovland, Janis, &
Kelley, 1953; Ohanian, 1991). This fit, or congruency between the characteristics of the
endorser and the product, has often been examined to evaluate the effectiveness of
endorsers in advertising. Numerous researchers have found that this fit, when appropriate,
does influence a consumer's purchase intentions (e.g., Ohanian; Tripp, Jensen, &
Carlson, 1994; Wansink & Ray; 2000).
In addition to understanding the influence of an individual's level of identification
with the endorser/sport and perception of the endorser-product fit, the consumer' s
perception of the product must be examined. Ultimately, an individual must believe that
there is some level of value in owning the product in order to purchase it. Perceived value
has been defined as "consumers overall assessment of the utility of a product (or service)
based on perceptions on what is received and what is given" (Zeithaml, 1988, p. 14).
Examined in terms an individual's emotional attachment, its price, its quality, and/or a
feeling of social acceptance received from obtaining the item (Lee, Trail, Kwon, &
Anderson, 2005; Sweeney & Soutar, 2001), perceived value involves a consumer's
overall perception of the product. If a consumer does not believe that the product holds
any value, the individual will not purchase the product. Therefore, the perceived value of
the product is an integral aspect of the relationship.
Through the course of studying the phenomenon of using endorsers in advertising,
four main ideas have been developed. They include the Source Credibility Model
(Hovland et al., 1953; Hovland & Weiss, 1951; McGuire, 1968), the Source
Attractiveness Model (McGuire, 1985), the Meaning Transfer Model (Langmeyer &
Walker, 1991a; 1991b; McCracken, 1989), and the Product Match-Up Hypothesis (Kahle
& Homer, 1985; Kamins, 1989, 1990). When synthesized, these concepts define a
specific set of factors that are deemed appropriate for use in the selection of the most
effective endorser of a product. The selection criterion lies with the thought process
stemming from the first model, the Source Credibility Model (Hovland et al.; Hovland &
Weiss; McGuire, 1968), which suggests that an individual's perception of an endorser' s
Expertise and Trustworthiness leads to an influential campaign. The Source
Attractiveness Model (McGuire, 1985) expanded upon the Source Credibility Model
(Hovland et al.; Hovland & Weiss; McGuire, 1968), claiming that an individual's
acceptance of a message stems from his/her similarity to, familiarity with, and liking of
the endorser that is relaying the message. In an effort to consolidate the findings of these
studies, Ohanian (1990) determined that perceived Expertise, Trustworthiness, and
Attractiveness were the most appropriate factors to use in examining the effectiveness of
an endorser. These factors extended to the Match-Up Hypothesis (Kahle & Homer;
Kamins, 1989, 1990), with researchers stating that the most effective campaign results
from an appropriate fit between the endorser and the product. As an extension of the
Match-Up Hypothesis (Kahle & Homer; Kamins, 1989, 1990), the Meaning Transfer
Model (Langmeyer & Walker, 1991a; 1991b; McCracken) was developed in order to
explain the process that is conducted when an endorser is used to promote a product. The
model, focusing on Perceived Image, illustrates that the image of the endorser is first
solidified in the mind of the consumer, the appropriate endorser is then chosen to pass the
image on from the endorser to the product, and, finally, the image is transferred from the
product to the consumer. These concepts have been synthesized into four main factors
referent to the selection of an appropriate endorser: perception of the endorser' s
Attractiveness (e.g., McGuire, 1968; Ohanian), Expertise (e.g., McGuire, 1968;
Ohanian), Trustworthiness (e.g., McGuire, 1985; Ohanian), and Likeability (e.g.,
Braunstein and Zhang (2005a) conducted a study to examine the aforementioned
characteristics, specifying the use of athletes as endorsers, which resulted in the
formation of the Scale of Athletic Star Power (SASP). This scale included five factors,
characterizing athlete endorsers based on consumer perceptions: Professional
Trustworthiness (e.g., genuineness and integrity), Likeable Personality (e.g., behavior),
Athletic Expertise (e.g., knowledge, experience, and ability), Social Attractiveness (e.g.,
image and attractiveness), and Characteristic Style (e.g., controversial and flamboyant).
Although Characteristic Style had not been distinguished as a unique factor in previous
celebrity endorser studies, the attributes themselves had been identified throughout
previous studies (e.g., Erdogan, 1999; Erdogan et al., 2001; Stevens et al., 2003).
Development of the scale was based on previous endorser effectiveness research (e.g.,
Baker & Churchill, 1977; Boyd & Shank, 2004; Brooks & Harris, 1998; Erdogan;
Hovland & Weiss, 1951; McGuire, 1968; McGuire, 1985; Miciak & Shanklin, 1994;
Ohanian, 1991), and it expanded the literature to include characteristics that may be
unique to athlete endorsers.
As a follow-up to their initial study, Braunstein and Zhang (2005b) reexamined the
5-factor SASP to look at the characteristics that consumers deemed important in athlete
endorsers. The reevaluation concluded in the creation of the Scale of Athletic Star Power
- Consumer Perspective (SASP CP). The SASP CP included many of the original
factors, including the collapse of the Social Attractiveness and Characteristic Style into
one factor: Public Image. An individual's overall image is often overlooked in the sport
literature as a result of the focus on attractiveness and expertise (Brooks & Harris, 1998;
Chalip, 1996; Fink et al., 2004). This may be an important factor when considering its
practical implications. For example, the influence of public image can be observed by
looking at the endorsement contracts for athletes on the Professional Golf Association
(PGA) Tour. Athletes on the PGA Tour, such as Tiger Woods, can have extensive
endorsement deals based on the athletic star power factors. However, if overall image of
the athlete endorser is not taken into account, a number of important factors could be
deleted from the equation. An athlete, such as Vijay Singh, who is a highly-ranked golfer
and performs on the same level as Woods, does not have the overall image necessary to
create an athlete endorser who will influence a consumer' s intention to purchase a
product. Therefore, although Singh is an expert in his sport, corporations do not believe
he will have a maj or impact in driving sales, leaving him behind in the race for
endorsement dollars (Seligman, 2005).
As previously mentioned, studies on athlete endorsers tend to focus on the
attractiveness and expertise of endorsers (Brooks & Harris, 1998; Chalip, 1996; Fink et
al, 2004). These studies lack the integration of the other factors (e.g., Trustworthiness and
Likeability) that previous researchers believed to be significant (Hovland et al., 1953;
Hovland & Weiss, 1951; Kahle & Homer, 1985; Kamins, 1989, 1990; Langmeyer &
Walker, 1991a; 1991b; McCracken, 1989; McGuire; 1968; 1985). Additionally, the
characteristics deemed important to the fit have not often included the "personality traits"
of the product itself (Aaker, 1997; Austin, Siguaw, & Mattila, 2003; Azoulay & Kapferer,
2003; Govers & Schoormans, 2005). These factors, including Trustworthiness, Expertise,
and Image, of both the endorser and the product may provide a more complete overview
of the influence of the fit between an athlete and a product on a consumer' s purchase
Several researchers have indicated the importance of understanding the needs and
wants of consumers in the selection of an athlete endorser (Greenstein & Marcum, 1981;
Hansen & Gauthier, 1989; Schofield, 1983; Zhang, Lam, & Connaughton, 2003; Zhang,
Pease, Hui, & Michaud, 1995). By choosing a star athlete or coach who appeals to a
specific market segment, the effectiveness of the sport endorsement is enhanced (Brooks
& Harris, 1998; Chalip, 1996; Howard, 1979). The consumer is a key factor to consider
when distinguishing among ideas regarding marketing a new, modified, or previously
established product. A company must understand the consumer group that will most
likely purchase its product. Although this information is often considered, many key
characteristics that pertain to athletes, the product, and the consumers have often been
overlooked in studies involving athlete endorsers.
While previous studies have shed considerable light on the general perception of
athlete endorsers and the reasons for their use, a number of limitations have been
identified that are associated with the applicability of the proposed theories to the
selection process of athlete endorsers. First, an integrated approach to studying the
phenomenon of athlete endorsers has not been devised. Based on professional intuitions,
many concepts (e.g., attractiveness, credibility, and meaning transfer) have been
proposed and superficially investigated (Hovland et al., 1953; Hovland & Weiss, 1951;
McCracken, 1989; McGuire, 1968, 1985). While Erdogan et al. (2001) stressed the
importance of measuring the characteristics of endorsers, products, and consumers in
order to reach a strategy, a theoretical framework that takes into consideration various
proposed theoretical frameworks and systematically measures the influence of athlete
endorsers has not been identified. It has often been stated that match-up is an integral part
of the process in determining the appropriate endorser of a product (e.g., Kahle & Homer,
1985; Kamins, 1989, 1990); however, while the concept itself is often assumed to be
integral in influencing a consumer' s purchase intentions, a framework has never been
developed that fully examines the hypothesis. Second, most studies have been conducted
in various business settings rather than in a sport context. According to Baumgartner,
Jackson, Mahar, and Rowe (2003), Nunnally and Bernstein (1994), and Thomas and
Nelson (2001), research Eindings are population and context specific. Therefore,
generalizations may be problematic and it is uncertain if research Eindings derived from
other business segments are applicable to sport endorsement. Third, the entire market
environment, not merely the celebrity and the endorsement, has been inadequately taken
into consideration. The impact of media value and the characteristics of the product and
the consumers are integral to the relationship formed in the process of endorsement. The
potential of endorsement is maximized through the understanding of specific elements
that facilitate consumption. Therefore, it is essential to enhance the match between the
product and the endorser in order to maximize the potential of an endorsement. Expertise
is generally deemed an important factor in the fit between endorsers and products (Boyd
& Shank, 2004; Brooks & Harris, 1998; Chalip, 1996; Erdogan, 1999; Fink et al., 2004;
Goldsmith et al., 2000; Hovland et al., 1953; Hovland & Weiss, 1951; Langmeyer &
Walker, 1991a; 1991b; McCracken, 1989; McGuire, 1968; Miciak & Shanklin, 1994;
Stevens et al., 2003). While these Eindings remain fairly consistent and logical for the
promotion of sport products, the practice of using athletes to endorse non-sport products
must come into question. While an athlete might fit into the appropriate endorser
category for a sneaker or a sports drink, the level of acceptance of an athlete endorsing a
car or a camera has not been extensively examined. Finally, those who have previously
studied the role of celebrity endorsers in sport have focused their attention mainly on the
opinions of advertising executives (e.g., Erdogan et al.; Miciak & Shanklin, 1994). These
studies opted to focus on the opinion of the producers, and often overlooked the
importance of the perspectives of the consumers (Kahle & Homer; Kamins, 1989, 1990;
McCracken; McGuire, 1968, 1985).
A celebrity of sport or non-sport is used in one out of four print or broadcast
commercials (Howard, 1979; Shimp, 1979). As a result, it is important to have a
framework that explains the factors that may lead to the success of a product' s campaign.
This study was designed to develop a model that included the factors that have been
derived from previous studies on product-endorser match-up (Expertise, Trustworthiness,
and Attractiveness/Image; e.g., Braunstein & Zhang, 2005a, 2005b; Kahle & Homer,
1985; Kamins, 1989, 1990; McGuire, 1968, 1985; Ohanian, 1990), examining the
influence of product-endorser match-up on a consumer' s purchase intentions as a result
of an individual's identification with a sport or athlete (Fink et al., 2002; Kwon et al.,
2005; Robinson & Trail, 2005; Robinson et al., 2004; Trail et al., 2003; Wann, 2002) and
perceived value of the endorsed product (Lee et al., 2005; Sweeney & Soutar, 2001).
To simultaneously test a variety of relationships among variables, the first step was
to test a model (Figure 1) that examined the relationships of these factors together in the
context of an athlete endorser promoting a non-sport product.
Figure 1. Influence of Product-Endorser Match-Up on Consumer's Purchase Intentions of (Non-Sport) Endorsed Products
This study tested the following research questions:
1. Did the model fit the data?
2. Which relationship explained the largest amount of variance in the model?
a. Direct Effects
i. Identification with Athlete/Sport Purchase Intention
b. Indirect Effects
i. Identification with Athlete -Congruency: Athlete/Product -
Attitude Toward Product Purchase Intention
ii. Identification with Sport -, Congruency: Athlete/Product -
Attitude Toward Product Purchase Intention
Significance of the Study
Previous studies focusing on the characteristics of endorsers have helped to
determine the main qualities that marketers and consumers deem important for
individuals that endorse products (e.g., Braunstein & Zhang, 2005a, 2005b; Erdogan,
1999; Erdogan et al., 2001; Hovland et al., 1953; Hovland & Weiss, 1951; Langmeyer &
Walker, 1991a; 1991b; McCracken, 1989; McGuire, 1985; Shimp, 1979; Smith, 1973;
Stevens et al., 2003). However, these studies have not fully analyzed the unique
characteristics of athlete endorsers. The current study was intended to fill the void by
reevaluating the factors deemed relevant in the match between an endorser and a product
to determine if this match does influence a consumer' s purchase intentions.
In this study, the inclusion of an individual's level of identification with the
sport/athlete and perceived value of the product, allowed the model to examine match-up
in a manner that resembled the reality of the marketing environment. In order to enhance
the generalizability of the research findings, the model examined each of these
relationships simultaneously in an effort to simulate and test the reality of the exposure.
With the addition of these factors, the model provided a better understanding of the
influence of athlete endorsers as promotional tools. The model may provide a helpful tool
to use when developing interventions, promotional strategies, and procedures.
Additionally, the use of structural equation modeling allowed for the control of
measurement and inferential errors, taking into account inaccuracy and compounding
factors when testing inter-factor relationships.
As previously mentioned, match-up (e.g., Kahle & Homer, 1985; Kamins, 1989,
1990), or endorser-product fit, is often noted and deemed important in the selection and
retention of endorsers in general, but has not been appropriately measured. Instead of
testing congruency, researchers typically measure a consumer' s level of identification
with an athlete (e.g., Boyd & Shank, 2004; Fink et al., 2004). While organizations are
investing a considerable amount of money on athletes to endorse their products (e.g.,
Tiger Woods' endorsement contracts are worth $80 million per year and Michael
Jordan's are worth $33 million per year; Weil, 2005), a model has not been developed to
examine this influence, or lack of influence, on an individual's purchase intentions. Both
practitioners (e.g., Tenser, 2004) and academicians (e.g., Fink et al., 2004), who have
examined the characteristics that they believe to be important in selecting an athlete
endorser, often include expertise in the mix. If expertise is a selection criterion, how does
this translate to the fit between a tennis player and a digital camera? Therefore, while this
practice is still rampant, it would be prudent to examine this model for future use with
both non-sport and sport products. The current research provides a model that may
answer a number of other research questions in the future. Does star power itself make
the difference and to what extent (e.g., high-profile athlete, average athlete, and non-
athlete)? Do different types of advertisements (e.g., humor and informational) make a
difference on the level of influence? Is there a difference in the effectiveness of the
medium used (e.g., television, print, and radio)? Is match-up more effective when used
for a sport product than a non-sport product? These are questions that should be answered
to better understand the influence of athlete endorsers. The model in the current study
provides a foundation with which these questions can be explored in the future.
This study was confounded by a number of factors. Because this investigation
focused on the initial testing of a model, the findings are limited to the present sample as
well as those individuals in the sample that are currently aware of (a) the particular
endorser, (b) the sport with which the endorser wais associated with, and (c) the product
included in the study. Therefore, the findings of this single study will not provide
generalizable results that pertain to all situations involved in determining the
effectiveness of a product-endorser match-up.
There were a number of limitations associated with this study. This study was not
exempt from limitations often found as a result of the use of a convenience sample. The
sample consisted of only students from one region of the country. There were also a
number of limitations stemming from the purpose of the study. Because the purpose was
to test the model, only one athlete, one sport, and one (non-sport) product (therefore, one
overall match) were included in the study. Therefore, level of star power associated with
the individual athlete was not examined, nor was the impact of different athletes from
different sports or different products from different product categories. As a result, path
coefficients may be different in other studies for other pairings, yielding different
findings regarding the most influential relationship in the model. However, the findings
may provide a model that can be used or adapted for a number of combinations in the
Definition of Terms
While key terms are defined in the context of the model in Chapter Two, a number
of basic terms that used throughout the text must first be defined (Table 1; Merriam-
Table 1. Definition of Terms
A person who is trained or skilled in exercises, sports, or games
requiring physical strength, agility, or stamina.
The state of being personally attached, or to bind by personal ties
(as of affection or sympathy).
A feeling or emotion toward a fact or state.
A state or habit of mind in which trust or confidence is placed in
some person or thing.
Marked or enhanced by harmonious agreement among constituent
Offering reasonable grounds for being believed.
To approve openly or to express support or approval of publicly
One with the special skill or knowledge representing mastery of a
particular subj ect.
A largely unconscious process whereby an individual models
thoughts, feelings, and actions after those attributed to an obj ect
that has been incorporated as a mental image.
A pair suitably associated.
To regard as being such.
Table 1 Continued
The act or process or an instance of persuading (to move by
argument, entreaty, or expostulation to a belief, position, or course
An outstandingly talented performer or a person who is preeminent
in a particular field.
To convey from one person, place, or situation to another.
Something (as a principle or quality) intrinsically valuable or
In Chapter One, the purpose of the study and the model that was tested has been
explained. The remaining chapters focus on the theoretical framework of the model, how
the study was conducted, and the findings of the research project. Therefore, Chapter
Two will provide an overview of the pertinent research regarding the theoretical
framework that supports the model. Chapter Three will explain the research methods
involved in the creation of the questionnaire and the steps that were taken to complete the
study. Chapter Four will then provide the results of the study and, in conclusion, Chapter
Five will expand upon the results provided in Chapter Four, presenting the discussion and
conclusion of the study.
REVIEW OF LITERATURE
According to the American Marketing Association, consumer behavior is "the
dynamic interaction of affect and cognition, behavior, and the environment by which
human beings conduct the exchange aspects of their lives" (Marketingpower. com, 2006).
Therefore, in order to understand what influences a consumer' s behavioral intentions and
ultimate behavior, the decision-making process concerning the purchase must be
examined based on the specific decision that the individual is making.
In order for a consumer to ultimately make a purchase, the individual must first be
aware of the product. After an individual is aware of the product, a marketer must then
take the steps necessary to lead the consumer from awareness, to interest, then desire, and
finally, action. In order to examine this phenomenon, Lavidge and Steiner (1961)
developed the Hierarchy of Effects model to measure advertising effectiveness. This
included six "steps" and three general processes, to be performed in order: (a) cognition
(i.e., awareness-knowledge), (b) affect (i.e., liking-preference), and (c) conation (i.e.,
conviction-purchase). While the model was initially designed as a step-by-step process,
the time that each individual spends at each phase may depend on the specific product.
Additionally, the level of investment required to consume the product may alter the
process. If less of an investment is required, steps may be skipped in the decision-making
process. Palda (1966) questioned Lavidge and Steiner' s model, claiming that there was
no empirical evidence to support its claims. Palda did not believe that the cognition-
affect-conation hierarchy was the only way to define decision-making. Ray (1973)
supported Palda's assertion, suggesting that there was not one process, but three: the
traditi onal hi erarchy (cogniti on-affect-conati on), the di ssonance-attributi on hi erarchy
(conati on-affect-cogniti on), and the low involvement hi erarchy (conati on-cogniti on-
affect). As a result of each individual's unique internal motives, or motivators, it is
important for marketers to understand what drives an individual's attitude and, in turn,
attitude change (Eagly & Chaiken, 1993).
The "awareness" that Lavidge and Steiner (1961) discussed can be accomplished
through various promotional tactics (e.g., print, radio, television, Internet). Reis and Trout
(1981) believed that these methods are used to position a product or entity in the minds of
both current and potential consumers. Positioning, a tool used to communicate with
consumers in a crowded marketplace, has often been used to set an organization apart
from its competitors. One way of positioning a product is through an organization's
promotional strategies (e.g., advertising campaigns). An organization can choose to use
the type of advertisement that they believe will reach their target market. In order to set
oneself apart, an organization may use a number of different positioning tactics, including
the use of a celebrity endorser in their advertising (e.g., Javalgi, Traylor, Gross, &
Lampman, 1994; McCracken, 1989).
In North America, the public maintains a steady interest in those deemed to have
celebrity status. Therefore, the individuals that make up this population of celebrities
remain a staple in the marketing strategies of executives in a wide array of industries,
ranging from automobile manufacturers to the makers of sports equipment. Due to the
extensive use of endorsers (including celebrities) in advertising and promotions,
numerous researchers have attempted to understand the impact that these tactics have on
their markets (e.g., Friedman & Friedman, 1979; Kahle & Homer, 1985; Kamins, 1990;
Ohanian, 1990). The use of celebrity endorsers is believed to (a) generate a greater
likelihood of a customer choosing the endorsed brand or product (e.g., Heath, McCarthy,
& Mothersbaugh, 1994; Kamins; Ohanian, 1991), (b) enhance message recall (e.g.,
Friedman & Friedman), (c) aid in the recognition of brand names (e.g., Petty, Caciopo, &
Schumann, 1983), and (d) create a distinct personality for the endorsed brand or product
(e.g., Javalgi et al., 1994; McCracken, 1989).
Individuals in sport (e.g., athletes and coaches) are not exempt from the American
public' s obsession with celebrity. As a result of professional sport' s evolution from
newspaper to radio to television and the Internet, it has long held a place in the public
spotlight (Schaaf, 2004). Athletes and coaches, particularly those with unique
personalities, levels of talent, and/or longevity in the industry, have become celebrities in
their own right. This level of celebrity has provided an opportunity for both athlete and
manufacturer to work with one another to promote a variety of products. As the number
of high-profile athletes has grown, so has the number of multi-million dollar endorsement
contracts (King, 2005). Stemming from Miller Lite' s "Lite All-Stars" and "Tastes Great,
Less Filling" campaigns in the 1970's, both athletes and coaches have evolved into
mainstream celebrities, endorsing everything from beer and sports equipment to beauty
products. The use of athletes and coaches to endorse products gained credibility after the
success of Miller Lite' s initial campaign using "masculine" male-athletes to promote
lower calorie beer (Miller, 2002). Because celebrity endorsements have the ability to help
build a brand's image (Cornwell, Roy, & Steinard, 2001; Javalgi et al., 1994; McDonald,
1991), athletes such as John Madden and Bubba Smith were used in Miller' s campaigns
to change the image of a non-masculine product (Miller).
As this practice perpetuates, and athletes and coaches continue to sign large
endorsement contracts, understanding the effectiveness of these tactics becomes
extremely important. With contracts for current athletes (e.g., Tiger Woods $80 million
per year; Weil, 2005) and retired athletes (e.g., Michael Jordan $33 million per year;
Weil) on the rise, the organizations that invest large amounts of money want to leave as
little to chance as possible in terms of return on investment (ROI). While recent studies
have been conducted regarding the use of athlete endorsers in terms of expertise and
attractiveness (e.g., Boyd & Shank, 2004; Fink, Cunningham, & Kensicki, 2004), the
maj ority of the previous research regarding endorsers does not focus on sport and athletes
as endorsers. Therefore, it is necessary to develop a model that includes the unique
attributes of sport. A model that examines these relationships may help marketers to
make a more educated decision regarding the selection of an endorser.
The relationship between an endorser and a product has been termed congruency,
fit, and match-up. No matter the terminology, a match between the characteristics of the
two entities is often deemed to be highly relevant when attempting to emit an effectively
endorsed message (e.g., Kahle & Homer, 1985; Kamins, 1989, 1990; Kanungo & Pang,
1973; Lynch & Schuler, 1994; Peterson & Kerin, 1977). Typically, fit has been examined
in terms of the endorser' s Attractiveness (e.g., physical appearance; McGuire, 1968;
Ohanian, 1990), Expertise (e.g., specialized knowledge; McGuire, 1968; Ohanian),
Trustworthiness (e.g., believability; McGuire, 1985; Ohnaian) and Liking (e.g., personal
feelings about that individual; McGuire, 1985).
Although the congruency between endorsers and consumers has often been defined
as an integral part of the match-up process, it has been measured in a variety of ways.
Employing endorser-consumer match-up to define congruency, researchers have often
used an individual's liking of, or identification with, a particular endorser to measure this
construct (e.g., Aronson & Worchel, 1966; Berscheid, 1966; Kelman, 1961). While these
findings do not provide a measurement model to fit the current study, they validate the
characteristics used in the endorser-product studies (e.g., Kahle & Homer, 1985; Kamins,
1989, 1990; McGuire, 1968; McGuire, 1985; Ohanian, 1990).
An individual's identification with an endorser and/or an endorser' s sport can be
traced back to identity theory (e.g., James, 1890; Mead, 1934; Stryker & Burke, 2000).
Based on the belief that each individual has multiple role identities, seven points of
attachment have been proposed as motives for spectator attendance at sporting events
(Kwon, Trail, & Anderson 2005; Robinson & Trail, 2005; Robinson, Trail, & Kwon,
2004; Trail, Robinson, Dick, & Gillentine, 2003). While these factors have yet to be
examined in terms of the effectiveness of athlete endorsers, two of the seven points of
attachment, or role identities (attachment to the sport and attachment to the
player/athlete) may help to define the effectiveness of the match and, ultimately, the
consumer' s intent to purchase the endorsed product.
Both scholars and practitioners claim that celebrities (including athletes and
coaches) help in the branding of a product (Brooks & Harris, 1998; Chalip, 1996;
Howard, 1979). A number of concepts and theories have been developed to examine the
use of endorsers in advertising, and the characteristics that work to create the most
effective message (e.g., Kahle & Homer, 1985; Kamins, 1989, 1990; McGuire, 1968,
1985; Ohanian, 1990). Once these characteristics have been established, the next step is
to use a method that will synthesize the important characteristics from each element of
the consumption process (i.e., endorser, product, and consumer) and determine how they
work together to increase an individual's intent to purchase endorsed products. As
marketing executives rely on endorsers as promotional tools, it is important that they
fully understand and appropriately use the information that may lead to a consumer' s
potential attitude change (Eagly & Chaiken, 1993). In order to have a greater
understanding of the actual influence of athlete endorsers on purchase intentions, an
individual's level of identification with the athlete and the sport (e.g., Kwon et al., 2005;
Robinson & Trail, 2002, 2005; Robinson et al., 2004; Trail et al., 2003) as well as their
perceived value of the product (e.g., Lee, Trail, Kwon, & Anderson, 2005; Sweeney &
Stoutar, 2001) must be examined. Once these factors have been discussed, the models
and hypotheses that support the use of endorsers as persuasive tools in advertising must
be examined. In order to determine which relationships have the greatest influence on a
consumer' s ultimate purchase decision, both direct and indirect effects of these factors
must be considered. Refer to Figure 1.
There are a number of limitations with the existing literature that have led to this
proposed model. No model has been proposed to examine the precedents and antecedents
to congruency. According to the Eindings of this literature review, congruency theories do
not typically measure fit. Instead they tend to examine the level of identification that an
individual has with another individual or product. Thus, a model that examines
congruency, based on attitude and identification theories, was proposed and tested. This
was accomplished by examining the influence of an individual's level of identification
with the athlete and the sport (e.g., Kwon et al., 2005; Robinson & Trail, 2002, 2005;
Robinson et al., 2004; Trail et al., 2003) on an individual's perception of endorser-
product congruency and purchase intentions. When using endorsers, this relationship may
include the influence of the congruency between the endorser and product (e.g., Kahle &
Homer, 1985; Kamins, 1989, 1990; Kanungo & Pang, 1973; Lynch & Schuler, 1994;
Peterson & Kerin, 1977), leading to an individual's final perceived value of the product
(e.g., Lee et al., 2005; Sweeney & Stoutar, 2001). Refer to Figure 1.
As previously mentioned, an individual's attitude toward an object is developed
through a process composed of three phases (cognition, affect, and conation; e.g.,
Lavidge & Steiner, 1961; Palda, 1966; Ray, 1973). According to Fishbein (1963), attitude
itself is an individual's affect for or against an obj ect. An individual's reaction, positive
or negative, is the result of that individual's belief about the obj ect and the evaluative
aspects of the belief. Fishbein and Ajzen (1975) developed the Expectancy-Value Model
of attitudes based on a cognitive approach to attitude formation. The model stems from
the belief that an individual takes an information-processing approach to developing an
attitude. The "thinking" is based on that individual's belief about the attributes of the
obj ect. The individual, already having positive or negative beliefs about different
attributes, then creates an attitude toward the obj ect (positive or negative). Therefore, it is
not merely the obj ect or the information that is provided at the decision-making stage that
influences the overall attitude or attitude change leading to a purchase decision. An
individual's belief about the attributes of the obj ect predisposes all responses (cognitive,
affective, and conative) toward the object (Krebs & Schmidt, 1993).
Zimbardo and Ebbesen (1970) expanded upon Fishbein' s (1963) theory, stating that
an individual's perception of an attribute is learned, not innate, and therefore capable of
change. Although attitudes are not easily altered, practices that are known to influence
beliefs about attributes may be applicable in eliciting an altered response. Researchers
have suggested that advertising is a tool that can be used to implement attitude change
(e.g., Lutz, 1975; MacKenzie & Lutz, 1989; MacKenzie, Lutz, & Belch, 1986). Lutz
(1975) hypothesized that (a) altering an individual's belief about a brand attribute leads to
a change in that individual's attitude towards using the brand in question and (b) altering
an individual's evaluation of a brand attribute leads to a change in that individual's
attitude toward using a the brand in question. Examining these hypotheses with a
previously nonexistent brand, the experimental manipulations worked towards attitude
formation as well as attitude change. Although the experiment did not provide an
indication of the response to a preexisting brand, the study's findings supported the
hypotheses (based on multiple attribute theories) that a cognitive stimulant can work to
alter an individual's initial attitude with a new brand. Lutz (1985) provided a model to
observe the phenomenon, examining four sets of conditions: (a) pure affect transfer, (b)
message-based persuasion, (c) contextual evaluation transfer, and (d) dual mode
persuasion. His findings suggested that an individual's attitude toward the advertisement
in general was an important predictor of the individual's attitude toward the brand. Lutz
also concluded that long-term effectiveness was enhanced when pertinent information
about the brand was presented in an entertaining or distinctive fashion (dual mode
persuasion). MacKenzie et al. expanded upon Lutz's (1985) framework, using a low-
importance product (toothpaste) in a single-ad pretest exposure setting to observe the
influence of an advertisement as a causal antecedent of brand attitude. The Eindings
supported Lutz's (1975; 1985) earlier work, with attitude toward the advertisement
strongly influencing attitude toward the brand and moderately influencing brand
perceptions. MacKenzie and Lutz reevaluated Lutz's (1985) framework in an empirical
examination of the role that advertisements play in attitude change. The study strongly
supported Mackenzie et al.'s findings regarding the influence of the advertisement on the
individual's attitude toward the brand. However, that study did not align with the
previous findings regarding the advertisement' s influence on an individual's perception
of the brand.
For marketers, it is important to understand a consumer' s attitude regarding the
product that is being promoted. A consumer' s perceived value of the product is extremely
relevant in the formulation of a product' s marketing mix. An individual's attitude toward
an obj ect is based on the individual's beliefs about the attributes of the obj ect (Fishbein &
Ajzen, 1975). Therefore, in order to alter a consumer' s affect regarding a product, an
organization should understand dimensions that lead to an individual's initial perception
of a product. Sheth, Newman, and Gross's (1991) examination of consumption values led
to a set of Hyve constructs that influence consumer choice. The factors included:
Functional Value (i.e., ability for functional, utilitarian, or physical performance), Social
Value (i.e., association with one more specific social groups), Emotional Value (i.e.,
ability to arouse feelings or affective states), Epistemic Value (i.e., ability to arouse
curiosity, provide novelty, and/or satisfy a desire for knowledge), and Conditional Value
(i.e., result of a specific situation or set of circumstances facing the decision-maker).
When analyzing an individual's motivation to purchase a product, the value that the
individual perceives to be attached to the product itself must be examined. Therefore, an
integral step in the process is operationalizing perceived value as it relates to the potential
purchase intentions of a consumer. According to Lee et al. (2005), the critical distinction
between personal values (Richins, 1994) and perceived value (Sweeney & Soutar, 2001)
is often not distinguished appropriately. Accordingly, personal values are the general
beliefs that guide an individual's behavior (Richins), and perceived value is the worth of
an object to an individual in a specific situation (i.e., domain specific; Sweeney &
Soutar). The importance in the distinction between the two concepts is that an
individual's personal values are often a precursor to the perceived value of the obj ect
(Lee et al.). This distinction is necessary, and was not made by Sweeney and Soutar in
the development of the Perceived Value (PERVAL) scale. The scale, stemming from
Sheth et al.'s (1991) values study, was created to measure the perceived worth of
consumer products to the purchaser. Sweeney and Soutar argued that previous research
failed to capture all of the elements that a consumer values, focusing solely on the trade-
off between Quality and Price, or what is given and what is received (Cravens, Holland,
Lamb & Moncrieff, 1988; Monroe, 1990; Zeithaml, 1988). While the majority of
perceived value research has focused on Quality and Price, researchers believe that
focusing on these two factors does not provide a complete picture of the decision-making
process (e.g., Porter, 1990). Sweeney and Soutar agreed with these sentiments, believing
that individuals consumed products based on the way the product makes them feel, and
not simply the cost and quality of the product itself. As a result, they added Emotional
and Social constructs to the previously studied dimensions of Quality and Price. Lee et al.
argued that, while Price and Quality did fit Sweeney and Soutar' s definition of perceived
values, the factors of Emotional and Social should have been categorized as personal
In a reassessment of Sweeney and Soutar' s (2001) PERVAL scale, Lee et al.
(2005) conducted a study to examine consumers' decision-making processes regarding
the purchasing of licensed sport merchandise. They found that the model did not fit the
data well. The Price and Quality subscales did not show good construct reliability or
internal consistency. However, the Emotion and Social subscales did. While the model
did not fit the data well, there were a number of items within each factor that had high
factor loadings. The model in Lee et al.'s study did not support the original constructs. As
a result, the authors proposed revised dimensions, suggesting that the revised factors
should be reevaluated in terms of face and content validity before being used in future
Advertisers have also realized that using endorsers who have similar characteristics
as the promoted product (endorser-product congruency) might increase the likelihood of
the purchase (e.g., Ohanian, 1991; Tripp, Jensen, & Carlson, 1994; Wansink & Ray,
2000). Therefore, the theories and models used to examine endorser effectiveness provide
a framework to determine which characteristics that are deemed most relevant in the
relationship (e.g., Kahle & Homer, 1985; Langmeyer & Walker, 1991a; 1991b;
McCracken, 1989; McGuire, 1968; 1985).
McGuire's Source Credibility Model (1968) was the first approach used in the
attempt to understand the characteristics that create an effective advertising/marketing
campaign. Source credibility is a term commonly used to imply a communicator' s
positive characteristics that affect the receiver' s acceptance of a message. The Model
depicted the influence of perceived level of Expertise and perceived level of
Trustworthiness on the effectiveness of the message. McGuire operationalized the first
factor, Expertise, as the perceived level of knowledge, experience, or skills of an endorser
and the second factor, Trustworthiness, as the intended audience's belief in the honesty,
integrity, and believability of the endorser. He suggested that these factors be used in a
Hyve-step process to change the attitude of a consumer: attention, comprehension,
yielding, retention, and action.
Building on the concepts from the Source Credibility Model (McGuire, 1968),
McGuire (1985) developed the Source Attractiveness Model. The model depicted that an
individual's acceptance of a message relies on the Similarity (resemblance between
source and receiver), Familiarity (knowledge of the source through previous exposure),
and Liking (affection of the source as a result of the physical appearance or behavior) of
an endorser. McGuire (1985) theorized that the attractiveness of the source determined
the overall acceptance of the message being conveyed.
Ohanian (1990) combined the Source Credibility Model (McGuire, 1968) and the
Source Attractiveness Model (McGuire, 1985) to create a single scale to measure
celebrity endorsers' perceived Expertise, Trustworthiness and Attractiveness. Ohanian
defined Expertise as the characteristics that define an individual who is trained, informed,
educated, an authority, and competent. She suggested that Trustworthiness was the
potential consumer' s degree of confidence in, and level of acceptance of, the speaker and
the message (i.e., a consumer' s trustt in a speaker). Ohanian added the Attractiveness
dimension based on research that suggested that physical attractiveness is an important
cue in an individual's initial judgment of another person. The findings supported the
creation of a 15-item differential scale to measure the above traits (attractiveness, AVE =
.63 .65; trustworthiness, AVE = .63 .63; expertise, AVE = .61-.62). With the
development of a three-dimensional scale and further validation of the factors and items,
researchers now had a more valid and reliable approach to use in assessing each
component of celebrity endorsers' effectiveness and persuasiveness.
Kahle and Homer (1985) altered the original concepts to examine the use of
endorsers to develop the Product Match-Up Hypothesis. The hypothesis highlighted the
importance of the match between celebrity endorsers and products. Kahle and Homer
examined the influence of Physical Attractiveness on the attitudes and purchase
intentions of consumers by manipulating three independent variables: celebrity-source
physical attractiveness, celebrity-source likeability and participant-product attractiveness.
Results showed that physical attractiveness was more influential than likeability on a
consumer' s attitude toward the brand and purchase intentions.
McCracken (1989) suggested that the effectiveness of a celebrity endorser was
derived from the cultural meanings associated with the individual rather than simply the
endorser' s physical attractiveness and credibility. Therefore, he designed the Meaning
Transfer Model to further explain the endorsement process. He suggested that celebrity
endorsement involves a general process of meaning transfer through three stages in which
a celebrity passes on a message regarding a product. In the first stage, the consumer
forms an image of the celebrity. Second, to transfer the meaning or image of the celebrity
to the product, the organization selects a celebrity that represents the intended image of
the product. Finally, the meaning is then transferred from the product to the consumer
(Langmeyer & Walker, 1991a; 1991b).
Although the Source Credibility Model (McGuire, 1968), Source Attractiveness
Model (McGuire, 1968), Product Match-Up Hypothesis (Kahle & Homer, 1985), and
Meaning Transfer Model (Langmeyer & Walker, 1991a; 1991b; McCracken, 1989) are
all extremely relevant, the models are typically used to examine the perceptions of the
producers rather than the consumers. Therefore, executives need additional knowledge of
the factors that influence consumers in order to create a more successful marketing
As the above models and hypotheses state, the selection of a celebrity endorser is
often based on the profile of the endorser. The advertising and marketing literature often
define the selection of endorsers based on the consumer' s perception of an endorser' s
Attractiveness (e.g., physical appearance; McGuire, 1968; Ohanian, 1990), Expertise
(e.g., specialized knowledge; McGuire, 1968; Ohanian), Trustworthiness (e.g.,
believability; McGuire, 1985; Ohnaian) and Liking (e.g., personal feelings about that
individual; McGuire, 1985). As the public perceives athletes as celebrities, these
characteristics translate to athlete endorsers as well. The need for an examination of these
characteristics in the context of sport is necessary because athletes and coaches are
currently endorsing both sport and non-sport products. Therefore, it is necessary, as
numerous researchers have stated, to examine the importance of match-up between
product and endorser (e.g., Baker & Churchill, 1977; Kahle & Homer, 1985; Kamins,
1990; Lynch & Schuler, 1994; Till, 2000; Till & Shimp, 1998) and consumer and
endorser (e.g., Aronson & Worchel, 1966; Berscheid, 1966; Boyd & Shank, 2004;
Heider, 1958; Kelman, 1961; Mowen, Brown, & Schulman, 1979). Based on the above
information, the Match-Up Hypothesis was chosen because it represents the most
comprehensive approach to examining the effectiveness of the endorsement.
Numerous studies have examined the influence of an apparent match-up between
endorser and product to determine if there is an influence on the consumer' s purchase
intentions (e.g., Ohanian, 1991; Tripp et al., 1994; Wansink & Ray, 2000). Kanungo and
Pang (1973) used traditionally gender-specific advertising (men with stereos and women
with sofas) to propose that fittingnesss," or perceived congruence between
characteristics, existed between the endorser in an advertisement and the type of product
being advertised. Peterson and Kerin (1977) also suggested the need for product/endorser
congruency within an advertisement, if the purpose of the advertisement was to enhance
Many researchers have determined that the greater the congruence between the
image of the endorser and the image of the product being promoted, the more effective
the message (e.g., Kahle & Homer, 1985; Kamins, 1989, 1990; Kanungo & Pang, 1973;
Lynch & Schuler, 1994; Peterson & Kerin, 1977). However, one area that has been
disputed is how to measure effectiveness. Researchers have not agreed on a single
approach, and a number of methods have been used, including purchase intentions (e.g.,
Ohanian, 1991; Tripp et al., 1994; Wansink & Ray, 2000), attitude toward the
advertisement or brand (e.g., Goldsmith, Lafferty, & Newell, 2000; Till & Busler, 2000;
Tripp et al.), and increase in stock margins as the result of a contract announcement
(Agrawal & Kamakura, 1995). While match-up has been emphasized, the approach has
not been extensively examined in the realm of sport or with athlete endorsers. Therefore,
there is a need to examine (a) the characteristics of those involved in the match and (b)
the match itself, including the unique attributes of sport.
As the match-up hypothesis states, it is the fit between the characteristics of the
product and the endorser that is most relevant in creating an effective campaign.
Although numerous characteristics have been defined for celebrity endorsers in general,
for athlete endorsers, researchers tend to focus on Attractiveness and Expertise (e.g.,
Boyd & Shank, 2004; Fink et al., 2004). If athletes continue to be used as endorsers, sport
studies need to expand beyond the realm of Attractiveness and Expertise in order to
determine if other characteristics are relevant.
As a result of advertising campaigns, brands are often portrayed as having human
characteristics. Although this concept had been described, it had not been systematically
studied until Aaker' s (1997) creation of the Brand Personality Scale (BPS). After
examining 309 candidate traits based on previous literature (psychology, marketing, and
original qualitative research), the author ultimately looked at 114 traits over 37 brands.
The final BPS included 42 items/traits under 5 factors: Sincerity (e.g., honest, genuine),
Excitement (e.g., daring, spirited), Competence (e.g., reliable, responsible),
Sophistication (e.g., glamorous, charming), and Ruggedness (e.g., tough, strong). Aaker
believed, as a result of the rigor of the study, that the scale and framework were
generalizable across product categories. Govers and Schoormans (2005) confirmed
Aaker' s findings by conducting a longitudinal study examining the influence of product
personality on a consumer' s preference over time, finding that product-personality
congruency (P = .48) had the greatest impact on consumer preference.
Although Aaker (1997) claimed that the findings were generalizable as a result of
the breadth of the study, Austin, Siguaw, and Mattila (2003) argued that the BPS was not
generalizable to individual brands, as a result of the method of the study. Agreeing that a
measurement study was necessary to fully understand the yet unmeasured concept of
brand personality, Austin et al. (2003) reexamined Aaker' s BPS to determine the validity
of those findings. The authors found that, while the constructs were internally reliable
(Cronbach's alpha), they did not have construct validity (Average Variance Extracted >
.50 or discriminant validity). Thus, the findings did not support Aaker' s factor constructs.
Azoulay and Kapferer (2003) also agreed that the purpose behind Aaker' s study was
relevant; however, it was conducted in an inappropriate manner. The authors believe that
the study did not properly operationalize its concepts, measuring a number of different
brand identity constructs and claiming that they are all brand personality measures. While
the study was a review of current practices, and not an empirical investigation, Azoulay
and Kapferer agreed with Austin et al. that the concept should be studied, but perhaps
that the constructs should undergo a reexamination of both content and construct validity.
While the characteristics of the product, human or not, are integral to the marketing
mix, the endorser' s traits are equally as relevant when deciding to use a celebrity
endorser as a promotional technique. Kamins (1990) conducted a study to further
examine the Attractiveness aspect of the importance of congruency in an advertising
campaign. Subjects were shown advertisements containing physically attractive (Tom
Selleck) and physically unattractive (Telly Savalas) endorsers. Results indicated that
when the physically attractive celebrity was shown with an attractiveness-related product,
it enhanced measures of spokesperson credibility and attitude toward a brand, relative to
the use of the physically unattractive celebrity. However, the use of the physically
attractive celebrity had no effect on attractiveness-unrelated products. According to the
Kamins, the advertisement was deemed a success if there was an appropriate match-up.
Miciak and Shanklin (1994) explored the ways in which veteran celebrity
evaluators select effective endorsers based on certain criteria. Gathered from the 43
advertising agencies and companies that were surveyed, the most important criteria were
that the endorsers were trustworthy, readily recognizable by the target audience,
affordable, at little risk for negative publicity, and appropriately matched with the
intended audience. Five universal categories were also evaluated and ranked by
significance. These categories included Celebrity Credibility, Celebrity-Audience Match,
Celebrity-Product Match, Celebrity Attractiveness and Other Considerations.
Expanding upon the endorser characteristic research, Erdogan, Baker, and Tagg
(2001) investigated the importance of certain celebrity characteristics based on the
product type, from the point of view of the practitioner who is selecting the celebrity.
Erdogan et al.'s study examined British advertising agencies by first using exploratory
interviews and then progressing to mail surveys. The findings confirmed the hypothesis
that the importance of the measure used to choose an endorser for a product depended on
the type of product. Practitioners considered a set of criteria when choosing a celebrity
and that the importance of said criteria depended on the type of product being endorsed.
The researchers found that the most important characteristics included Trustworthiness,
Expertise, Familiarity, Likeability, and Physical Attractiveness, aligning with Miciak and
Shanklin's findings as well as those of other previous researchers (e.g., Kahle & Homer,
1985; Kamins, 1989, 1990; McGuire, 1968, 1985; Ohanian, 1990).
Javalgi et al. (1994) also examined the relationship between sponsorship and
corporate image. Their findings showed that sponsorship can improve the image of a
corporation, but it is not an automatic response. It can have either a positive or negative
impact, depending on the company's image prior to the sponsorship. If an organization
began with a positive image, the positive aura can increase. However, if a company's
image was negative prior to a sponsorship, its image may not benefit from the affiliation.
Therefore, as previously mentioned, the image of the product as well as the endorser
plays an integral role in determining the effectiveness of the match and, in turn, the
Effective endorser characteristics have been confirmed from a practical perspective
as well. Advertising Age (Tenser, 2004) examined the importance of specific endorser
characteristics from both the consumer and practitioner's perspective. They found a
number of characteristics that consumers deemed important in an endorser. Participants
did not believe that an endorser had to mirror the age, gender, or background
characteristics of a consumer. However, the constructs that were characterized as
pertinent in the selection of athlete endorsers relied heavily on the sport played and the
character of the individual. Some of the characteristics revealed as extremely important
included: "is same gender as me," "is good looking/stylish," "is hottest new star in his/her
sport," "has been playing or played the sport for a long time," "plays/played for one of
my favorite teams," "plays/played sport I follow," and "is someone I would like to be
like." While a number of the characteristics do overlap from those selected by consumers,
there are a few that are unique. These traits include: the right sport (sport your target
market likes), a winner (success on the field and post-sports career), clean living (not too
clean if product needs an "edge"), articulateness (endorser can speak in sentences and is
coachable), likes your brand (must be believable), personality (winning attitude,
outgoing, humility likeable), good looks (attractive), marketing savvy (understands
marketing goals), right portfolio (does not sign with competing brands), and right price
(based on budget). While practitioners may not have placed these characteristics in
categories, or factors, there is a distinct overlap between these Eindings and scholarly
research in the same domain.
Tripp et al. (1994) also conducted a study examining the effect that the number of
products endorsed by a celebrity had on consumer's attitudes and purchase intentions.
Through two experiments it was determined that as the number of products a celebrity
endorses increases, celebrity likeability and advertisement likeability both decrease.
Tripp et al.'s findings were in accordance with Hovland, Janis, and Kelley's (1953) study
regarding source credibility. The authors found both Expertise (the extent to which a
communicator is perceived to be a source of valid assertions) and Trustworthiness (the
degree of confidence in the communicator' s intent to communicate the assertions he or
she considers most valid) to be highly beneficial when choosing an endorser to influence
a consumer's attitude change.
Although previous studies had been conducted regarding athlete endorsers, they
typically focused on the attractiveness and expertise of the endorser (e.g., Boyd & Shank,
2004; Fink et al., 2004) or the heroism aspect of an athlete (e.g., Chalip, 1996; Stevens,
Lathrop, & Bradish, 2003), limiting the potential breadth of the Eindings. In order to adapt
the factors used in previous studies for use in sport studies, Braunstein and Zhang
(2005a) created the Scale of Athletic Star Power (SASP). Braunstein and Zhang believed
that it was crucial to understand the influence of additional characteristics that have been
used in the evaluation of endorsers in other disciplines. The purpose of the study was to
determine the characteristics that athlete endorsers encompass. The development of the
SASP was guided by celebrity endorser research Eindings that focused on
Trustworthiness, Likeability, Expertise, and Attractiveness. The first factor, Professional
Trustworthiness was defined as the genuineness and integrity of the endorser (Erdogan,
1999; Erdogan et al., 2001; Langmeyer & Walker, 1991a; 1991b; McCracken, 1989;
McGuire, 1985; Shimp, 1979; Smith, 1973; Stevens et al., 2003). It has been noted that if
an endorser is deemed untrustworthy, regardless of his or her characteristics, consumers
view the endorser's message as questionable (Smith, 1973). Second, Likeable
Personality, or the fondness for an endorser as a result of his or her behavior, pertains to a
consumer' s perception of the endorser as a genuine individual (Erdogan; Erdogan et al.;
McGuire, 1985; Stevens et al.). The third factor, Athletic Expertise, the perception that
the endorser is a credible source, refers to the perception that the individual chosen to
endorse a product has a base of knowledge, experience, and ability in his or her chosen
Hield (Boyd & Shank; Brooks & Harris, 1998; Chalip; Erdogan; Fink et al.; Goldsmith et
al., 2000; Hovland et al., 1953; Hovland & Weiss, 1951; Langmeyer & Walker, 1991a;
1991b; McCracken; McGuire, 1968; Miciak & Shanklin, 1994; Stevens et al.). The next
factor, Social Attractiveness, or the perceived image of an individual, does not
necessarily pertain to the physical attractiveness of an endorser. Instead, it is comprised
of personality, exposure, and an individual's public image (Baker & Churchill, 1977;
Boyd & Shank; Brooks & Harris; Chalip; Erdogan et al.; Fink et al.; McCracken;
McGuire, 1968, 1985; Ohanian, 1991). An additional factor, characteristic style, is the
distinctive characteristics of an individual. Although Social Attractiveness had not been
introduced in previous celebrity endorser research, it was included to further define the
unique characteristics of athlete endorsers (Erdogan; Erdogan et al.; Stevens et al.). The
SASP expanded the literature to include characteristics that may be unique to athlete
endorsers. The findings of the exploratory study resulted in a scale that included 5 factors
(Professional Trustworthiness, Likeable Personality, Athletic Expertise, Social
Attractiveness, and Characteristic Style) with a total of 40 items. The study showed that
all five factors influenced direct consumption of endorsed products, and when gender
differences were taken into consideration, the influence of athlete endorsers was greater
with female participants in terms of attractiveness and expertise.
In order to further validate the SASP, Braunstein and Zhang (2005b) reexamined
the scale in order to examine the characteristics of athlete endorsers that consumers
deemed important rather than the characteristics that consumers believe that athlete
endorsers possessed. The factor structure differed from the original SASP, combining
Social Attractiveness and Characteristic Style, renamed as Perceived Image, to create the
Scale of Athletic Star Power Consumer Perspective (SASP CP). When reassessing the
initial factors, and the content of the final factors, it was apparent that Perceived Image
examined the characteristics that pertained to both the athlete endorser' s attractive and
unique characteristics. Although an endorser's perceived image had been examined in
other areas of academic literature, the sport literature primarily focused on expertise and
attractiveness of athlete endorsers. When this was studied, attractiveness was typically
characterized as physical attributes (Boyd & Shank, 2004; Fink et al., 2004), not
necessarily public image (Baker & Churchill, 1977; Brooks & Harris, 1998; Chalip,
1996; Erdogan et al., 2001; McCracken, 1989; McGuire, 1968, 1985; Ohanian, 1991).
The addition of a factor that examined the overall image of an athlete endorser could be
beneficial to the sport literature in terms of its potential practical implications.
Consequently, Braunstein and Zhang (2005b) retained a total of 34 items representing
four factors (Professional Trustworthiness, Likeable Personality, Athletic Expertise, and
Perceived Image). While previous studies examined the fit of endorser characteristics
based on the practitioner' s perspective (e.g., Erdogan et al.), Braunstein and Zhang' s
(2005b) study confirmed that consumers also believe that these traits are relevant for
The congruency between the product and the endorser is an integral aspect of the
match-up relationship. Baker and Churchill (1977) were at the forefront of the match-up
phenomenon. They stated that the interaction between product and spokesperson
characteristics was most important in creating the most influential relationship. Friedman
and Friedman (1979) investigated the effects of matching endorser type and product type
by examining the characteristics of endorsers of products based on the level of risk
associated with the product being endorsed. Three types of endorsers (celebrity, expert,
and consumer) as well as three products defined by risk (low Einancial, performance, and
physical risk; high in Einancial, performance, and physical risk / low in psychological and
social risk; no risk) were included in the study. While examining trustworthiness,
awareness, attractiveness, and likeability, the researchers found that an effective match
was not limited to the physical characteristics of a spokesperson.
Till and Busler (1990) conducted two studies to examine the importance of two
match-up factors: Attractiveness and Expertise. The first study examined the influence of
attractiveness on brand attitude, purchase intention, and key brand beliefs. In the second
study, the authors examined the influence of expertise on the same dimensions by
manipulating the product and the type of endorser. The study found that expertise was
more important than attractiveness, as perceived expertise was found to have an
interaction with the endorser (e.g., athletes and energy bars versus actors and energy bars
or candy bars). Overall, results showed that expertise was a more important factor than
physical attractiveness when it came to brand attitude, purchase intent, and endorser-fit
with the product.
Lynch and Schuler (1994) used the schema theory to interpret the results of
previous inquiries into the match-up hypothesis and designed two experiments to provide
additional insight into how schema might be changed by a spokesperson/product match.
The first experiment addressed the effect of match-up between a spokesperson
characteristic and a product attribute on spokesperson credibility. The second considered
the effect of a match-up between a spokesperson characteristic and a product attribute on
the schema of the product. Both studies supported the match-up hypothesis between a
spokesperson' s characteristics and a product's attributes in that they could both initiate a
change in the spokesperson and product schemas.
It is often assumed that an endorser' s image often influences the image of the
product he or she is helping to promote. Till and Shimp (1998) used an associative
network model of memory to examine the influence of negative information about a
celebrity on the product that they endorsed. Three studies were conducted to examine the
relationship. In the first two studies, fictitious, but realistic, celebrity endorsers were
used. In the third study, an actual celebrity was used. Findings showed that negative
information about a celebrity resulted in a decline in attitude towards the endorsed brand
only for the fictitious celebrity. Overall, Till and Shimp's results differed from Kamins'
(1990) study in that there was no match-up interaction between endorser attractiveness
and products that enhance attractiveness. As a follow-up to Till and Shimp's study, Till
(2000) examined the effect of the product that an individual is endorsing on the overall
image of the endorser him or herself. Till found that both athlete and non-athlete
endorsers' images were negatively affected when they represented unsuitable products.
Furthermore, when an athlete endorsed a product that was viewed as unhealthy, the result
was an even stronger negative image.
In sport, the event is often the product that the individual is endorsing. Fink et al.
(2004) used two fictitious athletes to examine the use of endorsers in promoting
attendance at a sporting event on a university campus. The model used in the study
included the constructs of Attractiveness and Expertise, converging into a second-order
latent variable of overall fit. The relationship of fit with attitudes and intentions to
purchase a ticket to the event was then explored. The researchers found that expertise was
more important for athlete endorsers in predicting an individual's intention to attend a
sporting event. Their findings supported the concept that fit is key, stating that
"attractiveness has little to do with the event itself...level of expertise has a significant
impact on an event" (p. 363). Jones and Schumann (2000) used these theories to examine
trends, congruence, and type of promotional strategies used in advertising campaigns
involving athlete endorsers. The authors used a sport-oriented medium over the course of
its entire publication (1955-1998). The overall findings showed that although the
popularity of athlete endorsers was continually growing (more in sports-related medium),
congruency was only found in approximately half of the advertisements, and most
advertisements were "thinking" type ads.
In sum, attitude theory and its extension to the congruency hypothesis has provided
the framework for the proposed model. According to Lutz (1985), attitude toward the
brand is mediated by attitude toward the advertisement, or in the present case, the
perceived congruency between the endorser and the product. Additional support for the
influence of endorser/product congruency on attitude toward the product comes from
Baker and Churchill (1977), Caballero, Lumpkin, and Madden (1989), and Caballero and
Solomon (1984). Furthermore, research has also shown that aspects that make up
endorser/product congruency consistently have been linked to positive attitude changes as
well as increased purchase intentions (e.g., Braunstein & Zhang, 2005a; Kahle & Homer,
1985; Ohanian, 1991; Simon, Berkowitz, & Moyer, 1970).
As noted above, an individual typically develops an attitude toward the product.
This may translate into an individual's initial intention to purchase or not purchase the
product. However, advertisers have realized that an advertising campaign may modify an
individual's initial attitude toward the product, thus potentially increasing the likelihood
of purchase (e.g., Heath, McCarthy, & Mothersbaugh, 1994; Kamins, 1990; Ohanian,
1991). Moreover, when including an athlete endorser in an advertising campaign, a
consumer' s identification with the sport and the athlete become an additional
consideration (Kwon et al., 2005; Robinson & Trail, 2005; Robinson et al., 2004). In
order to further understand the influence that these factors have on an individual's
decision-making process, identity theory must be examined (e.g., James, 1890; Mead,
1934; Stryker & Burke, 2000).
According to James (1890), individuals have different identities for each of their
networks of relationships. In each relationship, an individual plays a different role based
on the mutually reinforcing and conflicting aspects of oneself (Mead, 1934). Both
identity theory and social identity theory have been used to examine this phenomenon.
Identity theory, as a model, is made up of "four central components: (a) the identity
standard, or set of (culturally prescribed) meanings held by the individual which defined
his or her role identity in a situation, (b) the person' s perceptions of meanings within the
situation match to the dimensions of meaning in the identity standard, (c) the comparator
or the mechanism that compares the perceived situational meanings with those held in the
identity standard, and (d) the individual's behavior or identity which is a function of the
difference between perceptions and standards" (Stryker & Burke, 2000, p. 287). In
contrast, social identity theory states that the social categories to which individuals are
members of are all part of a structured society, existing only in relation to other
contrasting categories (e.g., student versus faculty, Black/African-American versus
Caucasian, male versus female). An individual is born into a society that has already been
structured. Therefore, social categories have been determined before the individual
identifies this role as part of his or her life. Each individual, belonging to a unique set of
social categories, has his or her own set of social identities that work to create an
individual's unique self concept (Stets & Burke, 2000). Both traditions (identity theory
and social identity theory) recognize that individuals view themselves in terms of
meanings imparted by a structured society. However, while both theories involve the
roles that an individual plays in society, identity theory is an individual's own perception
of the role that he or she plays in a specific situation while social identity theory is based
on an individual's role as determined by others.
Stryker (1980) built on the original concept of identity theory, using it to explain
the influence that social structures have on an individual and how, in turn, social
structures influence that individual's social behavior. While Stryker examined the
external motivators of identity theory, Burke (1991) explored the internal motivators of
identity theory. These two strands of identity theory research were brought together in
order to create a greater understanding of how both internal and external motivators
influence an individual's belief of the roles that self and social structure play in creating
an individual's identity (Stryker & Burke, 2000). In an examination of the past, present,
and future of identity theory, Stryker and Burke conceptualized identity theory to
encompass each of these critical components, claiming that an individual's cognitive
comparisons, which ultimately lead to the individual's final behavior, are influenced by
that person's identity standard and perceived situational meanings.
Identity Theory and Sport Consumption
Trail, Anderson, and Fink (2005) used identity theory to explain that an
individual's identification with a sports team may lead to a change in behavior. They
used the Team Identifieation Index (TII; Fink, Trail, & Anderson, 2002; Trail & James,
2001) to measure an individual's identification with the team. Trail, Anderson and Fink
(2000) defined identification as "an orientation of the self in regard to other obj ects
including a person or group that results in feelings or sentiments of close attachment" (p.
165-166). This concept, and specifically this index, has been used extensively. Several
studies (e.g. Fink et al., 2002; Robinson et al., 2004; Trail, Fink & Anderson, 2003; Trail
et al., 2003) found that team identification was significantly related to motives for
following teams (e.g., Drama, Aesthetics, and Social Interaction) and Trail et al. (2003)
showed that team identification was related to expectancies about game outcome. Trail et
al. (2005) determined that it was related to basking in reflected glory (BIRGing) and
cutting off reflected failure (CORFing) behavior and also conative loyalty. However, the
above authors have not been alone in showing that team identification has an influence on
cognitive, affective, conative, and behavioral responses.
While others who have examined this phenomenon (e.g., Mahony, Madrigal, &
Howard, 2000; Milne, & McDonald, 1999; Wann, 2002) have not specifically used
identity theory, they have used measures that were representative of identity theory. For
example, Wann, in a preliminary attempt to validate a measure for assessing an
individual's identification as a sport fan, examined a fan's identification with his or her
role as a sport fan. This involved an individual's awareness and perception of themselves
as a sport fan. The creation of that instrument provided sport marketers with the ability to
assess the level of their fan's identification and, in turn, become more aware of their
consumer' s wants and needs.
Funk, Mahony, and Ridinger (2002) found that factors that pertain specifically to
the sport were the most influential regarding spectator support level. In a separate
application of the Sport Interest Inventory (SII), Funk, Ridinger, and Moorman (2003)
examined 18 factors regarding the aspects that lead to consumer support of the team. In
their study, 10 of the 18 factors explained 48% of the variance in consumer support for
the team, with Interest in the Team (team identification, P = .31) as the most influential
factor. In some of the earlier work that included variables that were similar to team
identification, findings showed that highly identified individuals differed from lowly
identified individuals on several behavior and affective factors (e.g., Branscombe &
Wann, 1991; Lever, 1983; Madrigal, 1995; Wann & Branscombe, 1993).
Identity Theory and Points of Attachment
Although most of the previous research has focused on an individual's
identification with the team (Fink et al., 2002; Funk et al., 2002; Funk et al., 2003;
Robinson & Trail, 2005; Trail et al., 2005), a number of other studies have looked at
other points of attachment, such as identification with the athlete/player or identification
with the sport. Trail et al. (2003) suggested that an individual's level of identification in
sport may not lie solely with a team, and that there are seven potential points of
attachment (Player, Team, Coach, University, Community, Sport, and Level), creating
the Points of Attachment Index (PAI) from the original research with the TII (Fink et al.,
2002). This builds on James (1890) and Mead's (1934) assumptions that individuals have
multiple role identities. The PAI has been retested numerous times (Kwon et al., 2005;
Robinson & Trail; Robinson et al., 2004) with the model showing at least adequate fit
each time. While the PAI has typically been tested on sport consumption, two of the
subscales may prove beneficial in the assessment of the effectiveness of athlete
endorsers: identification with the athlete/player and identification with the sport that the
While the current study analyzed the relationship between the endorser and the
consumer in terms of identification and identity theory (e.g., James, 1890; Mead, 1934;
Stryker & Burke, 2000), others have examined this relationship in terms of congruency,
match-up, or hero-worship. As athlete endorsers have become commonplace in
advertising, understanding the athlete's characteristics that consumers deem important
becomes extremely relevant. In order to understand the characteristics of athlete
endorsers that consumers identify with, the unique attributes of sport and athletes must be
taken into consideration.
There has always been the phenomenon of the sport hero. As sport moves into a
new era, individual athletes are garnering the attention that teams once boasted. Wann,
Melnick, Russell, and Pease (2001) discussed the pastime of placing athletes on pedestals
in the public spotlight. Through the creation of entities like Halls of Fame, where hero
worship reins supreme, sport heroes can live on through the years. Generations can pass
down the knowledge of their favorite athletes, linking them with the next, and keeping
the hero worship mentality alive. As identification with these heroes allows the athlete or
coach to have an elevated level of power, it is in the best interest of the public to
understand the means by which they attain their status.
It may be more difficult to worship an athlete today than in the past. As demand
for athlete salaries increase to outlandish levels, labor disputes are constantly pending,
franchises threaten to relocate, and many individual athletes extracurricular activities do
not fit those that put them in contention for hero worship (Stevens et al., 2003). If an
organization is attempting to utilize an athlete to endorse its new product, the executives
should realize that fan identification still runs strong (Milne & McDonald, 1999).
Through all of the ups-and-downs in sport, fans still identify with athletes and teams.
Therefore, as endorsement deals grow and brands vie for the affection of elite athletes, it
may be assumed that fans are identifying more with their heroes.
While, according to the Match-up Hypothesis (Kahle & Homer, 1985), it is
extremely important to recognize the congruency between the product and the endorser, it
is also integral to the match to understand the match-up between the consumer and the
endorser. If a particular endorser does not influence an organization's target market, the
money spent on the endorsement may not be worth the investment. Therefore, it is crucial
to the match to be aware of the characteristics that a specific market deems important in
an endorser in order to maximize the match' s effect. While this area of match-up has not
been as widely studied as the product-endorser match-up, a number of studies have been
conducted to examine this relationship. For example, Mowen et al. (1979) used Heider' s
(1958) balance theory to describe the relationships that exist among target audience and
endorser, product and endorser, and consumer and product. The authors found that when
a strong affective relationship exists between each of the pairs, an endorser will be
maximally effective. Mowen et al. concluded that the match itself is what led to a
maximum effect. Boyd and Shank (2004) took this concept one step further by exploring
gender differences. The authors examined gender and expertise match-up as the most
effective use of athlete endorsers. When observing influence on male and female
participants, the results showed that women were more sensitive to the match between
endorser and product. In general, the Eindings state that a greater level of trust exists
between men and male endorsers and women and female endorsers. Additionally, the
authors found that the most effective campaign targeted male viewers, included a male
athlete, and involved a sport-related product. A study conducted by Peetz, Parks, and
Spencer (2004) also examined the role of gender in the meaning-transfer process. The
Endings of the study coincided with Boyd and Shank' s findings, stating that male
endorsers led to a greater increase of purchase intentions. More specifically, a greater
increase of purchase intentions was seen for male participants in response to male
endorsers. These findings support a number of similar studies that found that both
similarity and likeability have been used as determinants of identification and
interpersonal attraction between the source and the message recipient (Aronson &
Worchel, 1966; Berscheid, 1966; Kelman, 1961). As match-up between the consumer
and endorser can be roughly defined through identification and attraction (Kahle &
Homer; Kamins, 1989, 1990; McGuire, 1968; McGuire, 1985; Ohanian, 1990), these
Endings are extremely relevant in the match-up process.
In order for level of celebrity and fan recognition to be effective in marketing, an
individual's perception of an athlete must be elevated above their view of others. This
means that a consumer must believe that an individual has a particular talent or trait that
makes them unique (Braunstein & Zhang, 2005a). They must then wish to identify with
that athlete. As a result of identification, a marketer believes that if the "identified"
athlete endorses a certain product, the consumer will react to this association, and
purchase that product over one that is not endorsed by an individual he or she feels
People can typically relate well to others that they believe to have the same values
and interests as themselves (i.e., other individuals in their social structure). For example,
as Generation Y (Gen Y; age 10-27) begins to form their views on the world, they will
identify with those that they relate to, which will be an important concept for marketers
(Rein, Kotler, & Stoller, 1997; Stevens et al., 2003; Stotlar, Veltri, & Viswanthan, 1998).
This is one of the reasons that action sports have become mainstream for Gen-Y
consumers. This generation is looking for extreme ways to live, in part because
individuality is one of their defining qualities (Bennett, Henson, & Zhang, 2003). For
these reasons, marketers are feeding on the individual athlete that will have the ability to
relate to Generation Y (Marlatt, 1999). This generation is not looking to football or
baseball the way that previous generations have. As the sport marketer begins to
understand this, star power moves from the team as one entity to the individual athlete,
whether extreme or mainstream.
Application of the TII established that an individual's identification with the team
might influence consumption behavior (Trail et al., 2005). Trail et al. (2003) also found
that an individual's level of identification may expand beyond identification with just the
team, including the player/athlete and sport as well. This level of identification may
influence the consumer' s belief regarding his or her role identity regarding both the
athlete and the sport. Therefore, an individual's role in a social structure (e.g., Generation
Y) may influence an individual's belief regarding the congruency between an endorser
and a product (e.g., Shaun White and Mountain Dew). If congruency is found, this may
influence the individual's consumption behavior regarding the endorsed product (e.g.,
Heider, 1958; Mowen et al., 1979; Peetz et al., 2004).
No research has examined the relationship from identification with a specific
athlete or sport to perceived congruency between an endorser (the athlete) and the
product. Therefore, when examining the affect of endorsers on a consumer' s purchase
intentions, it is important to also examine that individual's level of identification with
certain factors, as they may influence behavior (e.g., Burke, 1991; Stryker, 1980; Stryker
& Burke, 2000). In the case of an athlete endorsing a product, it is vital that the level of
identification with both the sport and the endorser are taken into account in the overall
Previous researchers have examined the importance of fit when including a
celebrity in an advertising campaign (e.g., Kahle & Homer, 1985; Kamins, 1989, 1990;
Kanungo & Pang, 1973; Lynch & Schuler, 1994; Peterson & Kerin, 1977). According to
the Product Match-Up Hypothesis (Kahle & Homer), it is the fit between the
characteristics of the endorser and the product that is the most relevant in creating an
effective advertising campaign. Therefore, it is important to understand the characteristics
of the endorser and product that must first be measured to examine the fit in order to
create the most effective campaign. Although numerous characteristics have been defined
for celebrity endorsers in general, researchers tend to focus on the attractiveness and
expertise of athlete endorsers (e.g., Boyd & Shank, 2004; Fink et al., 2004). Sport studies
need to expand beyond these two factors in order to determine if other characteristics are
relevant for athlete endorsers. In order to achieve a greater understanding of the actual
influence of the match, other factors that influence the relationship must be taken into
consideration. While match-up studies have primarily focused on the influence of the
match itself, they often lack the influence of other factors. Both identification (sport and
athlete; e.g., Kwon et al., 2005; Robinson & Trail, 2005; Robinson et al., 2004; Trail et
al., 2003) and perceived value (e.g., Cravens et al., 1988; Monroe, 1990; Porter, 1990;
Sweeney & Soutar, 2001; Zeithaml, 1988) influence an individual's purchase intentions.
Additionally, the overall influence of these factors may affect a consumer' s perception of
the endorser-product match-up (e.g., Baker & Churchill, 1977; Friedman & Friedman,
1979, Till & Busler, 1990; Till & Shimp, 1998). Therefore, in order to fully understand
the influence of an athlete endorser, each of these factors (i.e., identification with the
endorser or the sport, perception of endorser-product congruency, and perceived value of
the product) must be examined to understand the role that they play in a consumer' s
decision-making process. See Figure 1.
Based on the number of effective parameters in the model (N = 145), a total of 850
participants were recruited for this study (minimum N = 132 to achieve power of .80 with
df = 1 130; MacCallum, Browne, & Sugawara, 1996). The participants (N = 400) were a
convenience sample of undergraduate and graduate students at a large southeastern
university. The majority of the participants were male (61%), 20-23 years old (75%),
White/Non-Hispanic (64%), and upper-level university students (junior/senior 70%).
Marketers use athlete endorsers in campaigns in order to target those that are most
highly influenced by this practice (the youngest generation with purchasing power, or,
presently, Generation Y; Holton, 2000). As the older segment of Generation Y, university
students are an appropriate population for this study. At this time in their lives, these
individuals may not have a solidified brand preference, but they may have a loyalty to a
particular athlete that can be transferred to the brand that the athlete is endorsing
(Holton). This, coupled with the ability to purchase middle to high-end products, is what
marketing and advertising executives hope to capitalize on with athlete endorsements.
The sample fit the demographic characteristics of the Generation Y population (Table 2).
Table 2. Descriptive Statistics for the Demographic Variables
Variable Category N %
Gender Male 245 61.1
Female 156 38.9
Age 18-19 72 18.0
20-21 212 52.9
22-23 92 22.9
24 or Older 25 6.2
Ethnicity Afri can-Ameri can/Bl1ack 42 10.5
Asian 23 5.7
Hispanic/Non-White 25 6.2
White/Hi sp ani c 46 11.5
White/Non-Hi spanic 258 64.3
Other 7 1.7
Current Student Status Freshman 25 6.2
Sophomore 78 19.5
Junior 163 40.6
Senior 122 30.4
Graduate Student 13 3.2
For this study, the questionnaire was split into two segments (Observation #1 and
Observation #2). On both segments of the questionnaire, all items were measured using a
7-point Likert-type scale (1 = strongly disagree; 2 = disagree; 3 = slightly disagree; 4 =
neutral; 5 = slightly agree; 6 = agree; 7 = strongly agree). The first segment used two
dimensions of the Points of Attachment Index (PAI; Trail, Robinson, Dick, & Gillentine,
2003) to examine the participant' s level of identification with both the athlete endorser
and the athlete' s sport. The second segment of the questionnaire included two additional
scales. First, the participants were asked questions to examine the congruency between
the athlete and the endorser, including elements of the Scale of Athletic Star Power -
Consumer Perspective (SASP CP; Braunstein & Zhang, 2005b). The second part of the
second segment included the Perceived Value (PERVAL) scale (Sweeney & Stoutar,
2001) to examine the subj ect' s perceived value of the product once he/she has seen the
advertisement. The survey included questions regarding the participant' s purchase
intentions of the product that was used in the study (I believe that this product...is one
that I definitely will purchase, is one that I would consider buying, is one that there is a
high probability that I would purchase, and is one that I intend to purchase). Additionally,
for the purpose of sample description, questions on five sociodemographic variables
(gender, age, ethnicity, current student status, and country of citizenship) were included
in a multiple-choice format. Prior to the administration of the questionnaire, the
University's Institutional Review Board approved the research methods.
Points of Attachment Index
The PAI has been used in numerous studies to examine the influence a consumer's
level of identification on his or her consumption intentions (Kwon, Trail, & Anderson,
2005; Robinson & Trail, 2005; Robinson, Trail, & Kwon, 2004; Trail, Robinson, Dick, &
Gillentine, 2003). Originally developed to determine the factors that influence consumers
to attend sporting events, the PAI has tested various dimensions (Player, Team, Coach,
University, Community, Sport, and Level) in a variety of settings. While the PAI factors
have remained consistent in previous studies, the scale has not been examined in terms of
its influence on an individual's consumption intentions regarding endorsed products or
the belief of the congruency between an athlete endorser and a product. Although
attachment to an athlete has often been referred to as an influential factor in the selection
of an endorser and ultimately the purchase intentions of a consumer (e.g., Tenser, 2004),
these factors have not often been controlled for in athlete endorser studies. Therefore, it
would have been careless to further explore this relationship without examining these
factors to better understand their role in the attitude change process (Eagly & Chaiken,
1993). Consequently, it was necessary to include them in the first stage of the model.
As a result of the content of the current study, it was deemed that two dimensions
(Sport and Player/Athlete) of the original seven (Player, Team, Coach, University,
Community, Sport, and Level) PAI factors (Trail et al., 2003) were to be included in the
model. After evaluating the items in terms of the context of the study, eight items were
retained. The Sport sub-scale included four items: "first and foremost, I consider myself a
(sport) fan," "being a fan of (sport) is very important to me," "(sport) is my favorite
sport," and "I am a (sport) fan at all levels and tours." The Player/Athlete sub-scale also
included four items: "I identify with (athlete)," "when (athlete) loses a match, it feels like
a personal failure," "I am a big fan of (athlete)," and "I follow (sport) because I am a big
fan of (athlete)."
Scale of Athletic Star Power Consumer's Perspective
In Braunstein and Zhang's (2005b) study to examine the Scale of Athletic Star
Power (SASP) with a unique focus, the SASP CP focused on consumer' s beliefs
regarding the characteristics that they believed athlete endorsers should possess, instead
of the characteristics that they did possess. The original scale was constructed to examine
characteristics exhibited by athlete endorsers that potentially influence consumer
purchases of endorsed products and included 40 items under five factors (Professional
Trustworthiness, Likeable Personality, Athletic Expertise, Social Attractiveness, and
Characteristic Style). As a reevaluation of the original scale, the SASP CP was
decreased to 34 items under four factors (Professional Trustworthiness, Likeable
Personality, Athletic Expertise, and Perceived Image).
It was important that the factors, both the SASP and SASP CP (Braunstein &
Zhang, 2005a, 2005b), were reevaluated to examine their relevance in terms of athlete
endorsers and the use of these individuals in promoting non-sport products. For use in
this study, items were evaluated and included or removed based on three criteria. Items
were removed if they had low factor loadings in both of the previous studies (SASP and
SASP CP; Braunstein & Zhang, 2005a, 2005b), did not load on the same factors for the
two analyses, or if they did not have face validity with the newly established 4-factor
model. Once these factors had been established, items were modified based on newly
acquired research regarding athlete endorsers (e.g., Tenser, 2004) and brand personalities
(e.g., Aaker, 1997). The scale then had 33 items under 4 factors.
In order to further reduce the number of items, a test of construct validity was
conducted with 40 university students. The students were provided with two lists: one list
of the factors and one list of the items. They were then instructed to match the items with
the factor that they believed worked best to define the overriding construct. As a result of
these findings and the addition of the congruency between athlete and endorser to the
model, likeability was eliminated as a factor.
After the initial construct validity check, the scale was decreased to three factors
with 18 items. These factors included Trustworthiness (sincerity, hard-working, quality,
consistency, dependability, and reliability), Expertise (competence, performance over
time, accomplishments, ability to perform, superiority, and success), and Image
(appearance, style, attractiveness, trendy, recognizability, and distinctiveness). As
previous research has consistently linked these three factors (Expertise, Trustworthiness,
and Image/Attractiveness) to both positive attitude changes and increased purchase
intentions (e.g., Kahle & Homer, 1985; Simon, Berkowitz, & Moyer, 1970), they fit the
criteria used to test the model. The scale was preceded by the phrase "I believe that both
the endorser and the product match up well in terms of..." in order to examine the
congruency of the attributes between the athlete endorser and the product.
Perceived Value Scale
The PERVAL was originally developed to expand the previous literature that
examined consumer' s perceptions of the value of products (Sweeney & Soutar, 2001).
Originally, perceived value was operationalized as the trade-off between Quality and
Price (Cravens, Holland, Lamb, & Moncrieff, 1988; Monroe, 1990; Zeithaml, 1988). The
introduction of the PERVAL scale was intended to add the dimensions of Emotion and
Social to the two original factors (Sweeney & Stoutar). Lee, Trail, Kwon, and Anderson
(2005) reexamined the PERVAL scale in terms of the convergent and divergent
characteristics of the factors. While the results did not reveal high factor loadings for
many of the items or high average variance extracted (AVE) values for many of the
factors, there were a number of items that did load well and were reevaluated for use in
this study. Therefore, the PERVAL factors that were used in the study included Emotion
(is one that I would enj oy, is one that I would want to use, is exciting, is one that I would
feel comfortable using, is one that would make me happy, would make me feel good, and
would be fun), Quality (would perform consistently well, is high-quality, is well-made,
has an acceptable standard of quality, and would last a long time), Price (is reasonably
priced, is a good product for the price, is fairly priced, and is affordable), and Social
(would give its owner social approval, would make a good impression on other people,
would help me to feel accepted by others, and would improve the way I am perceived by
others). All factors retained from Lee et al.'s study had factor loadings of .63 or higher, of
which only one item was below .70 (Lee et al.). The items were preceded by the
following statement: "Regarding (product) I believe that this product is..."
Following its preliminary formulation, the questionnaire was submitted to a panel
of experts (n = 16) for a test of content validity. The panel included eight university
professors in sport management, tourism, and/or marketing and eight graduate students
maj oring in sport management. Based on 80% agreement among the panel members, the
items were evaluated on whether the format and content were: (a) appropriate, (b)
adequate/representative, and (c) accurate/clear (Lam, Zhang, & Jensen, 2005; Zhang,
Pease, Hui, & Michaud, 1995).
After the initial items were determined, a test of internal consistency was conducted
on the SASP CP, PERVAL, and purchase intention subscales (N = 49). The test was
not conducted on the PAI, as the subscales (identification with sport and identification
with player/athlete) have proven to have high alpha values in previous studies (Kwon et
al., 2005; Robinson & Trail, 2005; Robinson et al., 2004; Trail et al., 2003). The items
were retained based on Nunally and Bernstein' s (1994) criterion that alpha reliability
coefficient' s should be equal to or greater than .70. The alpha coefficients for the SASP
- CP were .76 (Expertise), .83 (Trustworthiness), and .44 (Image). If the item
"controversy" was eliminated from the Image dimension, the alpha coefficient increased
to .69. Therefore, the item was dropped and "appearance" was added to the Image factor.
The final scale included three factors with six items representing each factor.
For the PERVAL scale, the alpha coefficients were .56 (Emotion), .83 (Quality),
.82 (Price), and .83 (Social). As a result of the low alpha coefficient for Emotion, the
items were reexamined. In order to improve Emotion, "would make me want to use it,"
the most problematic item, was dropped (after, 0= .61) and a number of additional items
were added ("is exciting," "is one I would want to use," "is one that would make me
happy," and "would be fun"). The Einal scale included 4 factors with a total of 20 items.
The purchase intention subscale had a sufficient alpha coefficient of .87, retaining all four
items. The final questionnaire included four sub scales (PAI, SASP CP, PERVAL, and
purchase intentions) with 50 total items.
Athlete and Product Selection
In order to identify the athletes for the study, the same students used to test the
content validity of the scale were asked to list all of the athletes that they could think of in
three minutes. The celebrities were later ranked based on the number of times their names
were mentioned. Two of the athletes that were mentioned most often were Dale
Earnhardt Junior and Maria Sharapova. As these two athletes currently promote middle-
to high-end products, they were chosen as potential endorsers to use to test the model.
The Einal pairings of endorsers and products include Dale Earnhardt Junior with Wrangler
jeans and Maria Sharapova with Canon cameras. Once the final pairings were determined
and advertisements were selected, the same set of students was asked if they would
purchase either of these products. After an overwhelming response that they would
purchase a Canon camera and they would not purchase Wrangler j eans, the Maria
Sharapova/Canon advertisement was chosen for the study.
Figure 2. Product-Endorser Advertisement
Subj ects were asked to participate in the study, designed to examine the affect of
athlete endorsers on consumer' s purchase intentions for non-sport products. Online
testing procedures were uniform for all research participants. Data collection procedures
included: (a) an e-mail to all potential participants, including an introduction of the
researcher, the purpose of the study, and a link to the questionnaire (serving as the
consent form); (b) an explanation that participation is voluntary, anonymous, and there
would be no penalty for not participating or not completing the questionnaire; (c) a
welcome statement; (d) two screening questions in order to be included as a research
participant in this study ("I know who Maria Sharapova is by reading her name" and "I
know what the Canon Power Shot digital camera is by reading the name of the product");
(e) the first part of the questionnaire; (f) the advertisement with the athlete and the
product (Figure 2); and (g) the second part of the questionnaire. Once the questionnaire
was finished, the data was submitted via the survey service and the final screen thanked
~rcSi;n .....c.l...--1. I .... ~. .-..
'-'' L''' ~~~-"*' *~~-''~''LLlrl- C~IL
~ Gi~ ~i L~B~r(
culplraa .I-~pU ~Cbl~'9*
the individual for his/her time and support for the study. The questionnaire was
administered in an online format and took approximately 10 minutes to complete.
Of the 850 e-mails distributed, 400 completed (555 total) submissions were made,
constituting an overall response rate of 47%. Descriptive statistics for the
sociodemographic variables were calculated using SPSS 13.0 (Norusis, 2004). The
RAMONA (SYSTAT, 1997) program was used to test the overall model fit of the
measurement model (a confirmatory factor analysis CFA). In the CFA, the latent
variables were allowed to correlate, while the error terms were not. Model fit was
determined by examining the root-mean-square-error of approximation (RMSEA;
Steiger, 1989), a confidence interval (CI) for the RMSEA, the ratio of the chi-square to
its degrees of freedom, and the percent of residuals greater than .10 (Browne & Cudeck,
1992). Researchers often indicate the importance of additional goodness-of-fit indices
(e.g., Root Mean Square Residual and the Non-Normed Fit Index; Hair, Anderson,
Tatham, & Black, 2002); however, RAMONA does not provide additional indices.
Browne and Cudeck believed that the RMSEA is the least influenced by the sample size
and number of parameters included in the model. Therefore, additional indices were not
deemed necessary and, as a result, not included in the RAMONA program. The RMSEA,
bound by zero on the lower end and only equal to zero when there is perfect fit, was the
first measure of model fit. According to Hu and Bentler (1999), the following criteria
regarding RMSEA values should be used to determine the fit of the model: less than .060
indicates a close fitting model, .061 .080 indicates reasonable fit, .081 .100 indicates
mediocre fit, and > .100 is unacceptable. While it has often been suggested that chi-
square per degree of freedom values should range from 2.0 to 3.0, there is no consensus
on this figure, with Bollen (1989) also stating that "acceptable values can reach as high as
5.0" (p. 278). According to Bagozzi and Yi (1988), when less than 10% of the residuals
are greater than .10, this is another indicator of adequate model fit.
Factor loadings equal to or greater than .707 show that there is more common
variance than unique variance (Hair et al., 2002), thus indicating that the item represents
the construct well. In addition, the factors were tested for internal consistency against the
standard of I > .70 (Nunally & Bernstein, 1994) and tested for construct reliability
against the standard of AVE values > .50 (Hair et al.). Discriminant validity was assessed
to determine if the factors were distinct from one another. This was accomplished by
squaring the correlations of the two referent factors. According to Fornell and Larcker
(1981), if the findings are greater than the AVE score of either factor, then the factors are
After the fit of the measurement model had been examined, RAMONA was then
used to examine the fit of the structural model. The same criteria were used to examine
the structural model as were used with the measurement model. Path coefficients were
used to determine the most influential direct or indirect relationship on an individual's
intent to purchase the product.
Measurement Model Analysis
After reviewing the skewness and kurtosis coefficients, one item ("When Maria
Sharapova loses a match, it feels like a personal failure") did not meet the criteria of
approximately + 2.00 (Stevens, 1996; Tabachnick & Fidell, 1996). As a result, the item
was eliminated from the remainder of the analysis. The balance of the items did meet the
criterion, suggesting that the data did not deviate from normality for any given variable
(Table 3). Therefore, it was appropriate to test the measurement model. The analysis
indicated that the data fit the model adequately. The RMSEA was .078 (CI = .075 to
.082) and the chi-square per degree of freedom value (2750/798) was 3.45. While the chi-
square per degree of freedom value may be slightly higher than the typically accepted
value, as previously stated, Bollen (1989) suggested that a value as high as 5.0 may be
deemed acceptable. The number of residuals greater than .10 in the current analysis, 150
(12.76%), was an indicator that the model could be improved. The maj ority of the AVE
coefficients were above the acceptance criterion (.53 for Athlete, .69 for Sport, .45 for
Expertise, .49 for Trustworthiness, .50 for Image, .54 for Emotion, .65 for Quality, .70 for
Price, .57 for Social, and .68 for Purchase Intention). Alpha coefficients of internal
consistency for the factors were all above the acceptance criterion of .70 (Nunnally &
Bernstein, 1994). AVE values, alpha coefficients, and factor loadings can be found in
Table 3. Descriptive Statistics for the Measurement Model
1 I follow tennis because I am a big fan of
2 I identify with Maria Sharapova.
3 When Maria Sharapova loses, it feels like
a personal failure.
4 I am a big fan of Maria Sharapova
1 First and foremost, I consider myself a
2 Being a fan of tennis is very important
3 Tennis is my favorite sport.
4 I am a tennis fan at all levels and tours.
Attitude Toward Endorser-Product Congruency
Expertise (6 items)
2 Performance over time.
4 Ability to perform.
Trustworthiness (6 items)
Table 3. Continued
Imane (6 items)
M~ SD Skewness Kurtosis
Attitude Toward the Product
Emotion (7 items)
1- Is one that I would enj oy.
2 Is on that I would want to use.
3 Is exciting.
4 Is one that I would feel comfortable using.
5 Is one that would make me happy.
6 Would make me feel good.
7 Would be fun.
Quality (5 items)
1 Would perform consistently well.
2 Is high-quality.
3 Is well-made.
4 Has an acceptable standard of quality.
5 Would last a long time.
Price (4 items)
1 Is reasonably priced.
2 Is a good product for the price.
3 Is fairly priced.
4 Is affordable.
Social (4 items)
1 Would give its owner social approval.
2 Would make a good impression on other
3 Would help me to feel accepted by others.
4 Would improve the way I am perceived by
Table 3. Continued
M~ SD Skewness Kurtosis
1 Is one that I definitely will purchase.
2 Is one that I would consider buying.
3 Is one that there is a high probability that
I would purchase.
4 Is one that I intend to purchase.
Table 4. Factor Loadings for the Measurement Model
Variable Pt. Est. CI (L) CI (H) SE t aA VE
Athlete (3 items)
1 I follow tennis because I am a big fan
Of Maria Sharapova.
2 I identify with Maria Sharapova.
3 I am a big fan of Maria Sharapova.
Sport (4 items)
1 First and foremost, I consider myself
a tennis fan.
2 Being a fan of tennis is very important
3 Tennis is my favorite sport.
4 I am a tennis fan at all levels and tours.
0.773 0.033 22.13
0.683 0.037 16.59
0.872 0.028 29.00
0.803 0.769 0.837 0.021 38.68
0.928 0.014 66.05
0.787 0.025 30.05
0.888 0.017 51.94
Attitude Toward Endorser-Product Congruency
Expertise (6 items)
2 Performance over time.
Trustworthiness (6 items)
Table 4. Continued
Image (6 items)
t aA VE
Pt. Est. CI (L) CI (H) SE
Attitude Toward the Product
Emotion (7 items)
1 Is one that I would enj oy.
2 Is one that I would want to use.
3 Is exciting.
4 Is one that I would feel comfortable
5 Is one that would make me happy.
6 Is one that would make me feel good.
7 Would be fun.
Quality (5 items)
1 Would perform consistently.
2 Is high-quality.
3 Is well-made.
4 Has an acceptable standard of quality.
5 Would last a long time.
Price (4 items)
1 Is reasonably priced.
2 Is a good product for the price.
3 Is fairly priced.
4 Is affordable.
Social (4 items)
1 Would give its owner social approval.
2 Would make a good impression on
3 Would help me to feel accepted by
4 Would improve the way I am
perceived by others.
0.577 0.518 0.637 0.036 15.95
0.596 0.539 0.654 0.035 16.99
0.900 0.872 0.927 0.017 53.86
0.883 0.855 0.912 0.017 50.59
Table 4. Continued
Variable Pt. Est. CI (L) CI (H) SE t aA VE
Purchase Intentions (4 items) .89 .68
1 Is one that I definitely will purchase. 0.849 0.821 0.877 0.017 49.97
2 Is one that I would consider buying. 0.681 0.633 0.729 0.029 23.32
3 Is one that there is a high probability
that I would purchase. 0.866 0.841 0.892 0.016 55.15
4 Is one that I intent to purchase. 0.892 0.870 0.915 0.014 64.23
The three Match-Up subscales (Expertise, Trustworthiness, and Image) did not
evidence discriminant validity as per Fornell and Larcker (1981). Thus, the first-order
latent variables were eliminated and all items were forced to load directly on the latent
variable "Match-Up" for the remainder of the analysis. After collapsing the factors into
one (Match-Up), the new AVE value was .49 and the new alpha coefficient was .93.
While the balance of the factors showed adequate discriminant validity, Quality was not
distinct from Emotion. Based on theory, the factors were supposed to be unique
constructs in the examination of Perceived Value (i.e., attitude toward the product; e.g.,
Lee, Trail, Kwon, & Anderson, 2005; Sweeney & Soutar, 2001). However, in this data
set, Quality and Emotion were not distinct. It is suspected that a "Quality" product may
result in positive emotions. Therefore, as it was not theoretically feasible to combine the
factors, Emotion was dropped for the remainder of the analysis. In addition, Emotion was
a more recent addition to the PERVAL scale (Sweeney & Soutar) and had been
problematic in previous studies (Lee et al., 2005). The altered model contained eight
factors: seven first-order latent variables (Athlete, Sport, Match-Up, Quality, Price,
Social, and Purchase Intention) and one second-order latent variable (Perceived Value).
Item correlations are found in Appendix A and factor correlations are found in Table 5.
Table 5. Factor Correlations for the Measurement Model
Variable Pt. Est. CI (L) CI (H) SE t
Social 0.599 0.534 0.657 0.038 15.96
Price 0.644 0.584 0.697 0.034 18.75
Quality 0.488 0.413 0.556 0.043 11.22
Emotion 0.622 0.560 0.678 0.036 17.28
Image 0.163 0.071 0.253 0.055 2.95
Trustworthiness 0.367 0.283 0.445 0.049 7.45
Expertise 0.355 0.270 0.435 0.050 7.05
Athlete 0.232 0.138 0.322 0.056 4.13
Sport 0.197 0.108 0.282 0.053 3.72
Price 0.434 0.355 0.506 0.046 9.44
Quality 0.359 0.275 0.437 0.049 7.28
Emotion 0.481 0.405 0.551 0.044 10.89
Image 0.182 0.090 0.271 0.055 3.28
Trustworthiness 0.343 0.258 0.423 0.050 6.80
Expertise 0.348 0.262 0.429 0.051 6.82
Athlete 0.185 0.089 0.278 0.058 3.22
Sport 0.099 0.008 0.188 0.055 1.81
Quality 0.594 0.528 0.652 0.038 15.73
Emotion 0.602 0.537 0.659 0.037 16.14
Image 0.257 0.168 0.343 0.053 4.82
Trustworthiness 0.449 0.371 0.521 0.046 9.79
Expertise 0.458 0.379 0.530 0.046 9.91
Athlete 0.184 0.089 0.276 0.057 3.22
Sport 0.178 0.089 0.264 0.053 3.34
Emotion 0.930 0.904 0.949 0.014 68.59
Image 0.580 0.510 0.643 0.040 14.36
Trustworthiness 0.659 0.598 0.712 0.035 18.95
Expertise 0.678 0.618 0.731 0.034 19.77
Athlete 0.227 0.132 0.318 0.057 3.99
Sport 0.165 0.075 0.252 0.054 3.05
Table 5 Continued
Pt. Est. CI (L) CI (H)
0.639 0.570 0.699
Structural Model Analysis
Due to the elimination of the Emotion subscale and the highly kurtotic Athlete
Identification item, the structural model consisted of 42 items. Although the original
relationships among Athlete Identification, Sport Identification, Match-Up, Perceived
Value, and Purchase Intention were not altered, the structural model did vary from the
measurement model in terms of the number of items analyzed and number of first-order
latent variables. The structural model revealed adequate fit as well, with an RMSEA of
.080 (CI = .077 to .083) and chi-square per degree of freedom value of 3.55 (2873/810).
However, not all criterion were met, with the number of residuals greater than 0. 10
(234/861 = 27. 18%) above the criteria of approximately 10%.
In addition to model fit, specific variance was examined to determine the amount of
variance explained in the model. In order to determine the amount of variance that the
model explained for a specific factor, specific variance (i.e., error) was subtracted from
1.0 (i.e., 1 Z). The variance explained for each factor was: 7% (Match-Up), 39%
(Perceived Value), and 54% (Purchase Intention). For path coefficients, see Table 6 and
Table 6. Relationships in the Structural Model
Variable Point Est. CI (Low) CI (High) SE t
Sport 0.642 0.572 0.702 0.039 16.34
Match-Up 0.268 0.136 0.399 0.080 3.36
Purchase Intention 0.064 -0.012 0.141 0.047 1.39
Match-Up -0.013 -0.139 0.113 0.077 0.17
Perceived Value 0.622 0.558 0.687 0.039 15.87
Quality 0.769 0.716 0.822 0.032 23.82
Price 0.787 0.735 0.838 0.031 0.29
Social 0.587 0.517 0.657 0.042 13.84
Purchase Intention 0.723 0.665 0.782 0.036 20.36
Figure 3. Structural Model
Previous studies focusing on the use of endorsers in advertising have helped to
determine the general factors (i.e., Attractiveness, Expertise, and Trustworthiness) that
academicians and practitioners believe play a key role in understanding the effectiveness
of endorsers (e.g., Erdogan, 1999; Erdogan, Baker, & Tagg, 2001; Hovland, Janis, &
Kelley, 1953; Hovland & Weiss, 1951; Langmeyer & Walker, 1991a; 1991b;
McCracken, 1989; McGuire, 1985; Shimp, 1979; Smith, 1973; Stevens, Lathrop, &
Bradish, 2003). However, there are other factors that may influence consumer behavior in
terms of the perceived congruency between an endorser and a product (i.e., effectiveness
of endorsers). With the continued practice of using athlete endorsers to promote non-
sport products, it was necessary to examine this phenomenon to determine its relevance
and impact on consumer's purchase intentions. Previously, athlete endorser effectiveness
was only examined in terms of the attractiveness and expertise of the endorser (e.g., Boyd
& Shank, 2004; Fink et al., 2004). This study was conducted to examine factors in
addition to those included in traditional studies that may influence a consumer' s purchase
Model development, guided by Attitude Theory (e.g., Fishbein, 1963; Fishbein &
Ajzen, 1975; Zimbardo & Ebbesen, 1970) and Identity Theory (e.g., James, 1890; Mead,
1934; Stryker & Burke, 2000), was intended to include factors beyond those typically
addressed in the endorser literature. First, attitude theory was used to examine a
consumer' s perception of the match between an endorser and a product (Expertise,
Trustworthiness, and Attractiveness/Image; e.g., Braunstein & Zhang, 2005a, 2005b;
Kahle & Homer, 1985; Kamins, 1989, 1990; McGuire, 1968, 1985; Ohanian, 1990) and
the individual's perception of the value of the product in general (Emotion, Quality,
Price, and Social; Lee et al., 2005; Sweeney & Soutar, 2001). As a result of the inclusion
of athlete endorsers in the study, it was determined that an individual's level of
identification with both the athlete and the athlete's sport should be included in the model
as well. Identity theory was used as the foundation for this phenomenon (Athlete
Identification and Sport Identification; e.g., Fink, Trail, & Anderson, 2002; Kwon, Trail,
& Anderson, 2005; Robinson & Trail, 2005; Robinson, Trail, & Kwon, 2004; Trail,
Robinson, Dick, & Gillentine, 2003; Wann, 2002). Ultimately, a 5-factor (50 item) model
was developed to examine the relationships between Athlete/Sport Identification and
Match-Up (Expertise, Trustworthiness, Image), Athlete Identification and Purchase
Intention, Match-Up and Perceived Value (Emotion, Quality, Price, Social), and
Perceived Value and Purchase Intention.
The proposed measurement model was tested and found to fit the data adequately.
However, there was evidence of psychometric problems with some of the subscales. One
item used to measure athlete identification ("When Maria Sharapova loses a match, it
feels like a personal failure.") was eliminated as a result of its high skewness and kurtosis
coefficients. This may have been a result of the syntax used to construct the item, as the
remaining athlete identification items referred to being a fan, and this may have tried to
measure something unique from the other statements that were much more general. After
examining skewness and kurtosis coefficients, factor loadings were reviewed. In general,
the factor loadings were favorable. Discriminant validity between the dimensions was
examined and it was determined that there was no discriminant validity among the
Match-Up subscales (Expertise, Trustworthiness, and Image). In addition, there was a
lack of discriminant validity between two of the four Perceived Value subscales (Emotion
The lack of discriminant validity among the Match-Up subscales contradicted
previous findings (e.g. Baker & Churchill, 1977; Friedman & Friedman, 1979; Kamins,
1990; Lynch & Schuler, 1994; Till & Busler, 1990), which indicated that these were
distinct concepts and that each played a significant role in the effectiveness of endorsers.
Expertise and Trustworthiness may not have been distinct because participants may have
believed that endorsers that have a certain level of expertise in an area are therefore
trustworthhy, or vice versa. This may also stem from a number of factors, including (a) the
fact that the advertisement included an athlete with a non-sport product or (b) a halo
effect. Halo effect refers to a cognitive bias, using an individual's judgment of one
quality to influence the assessment of other qualities (Asch, 1946). In other words, the
participants may not have examined the advertisement closely enough to decide whether
or not they believed there to be a congruency between the endorser and the product.
However, as a result of the screening questions, it was clear that they did have previous
knowledge of the athlete endorser as well as the product. Therefore, participants may
have had preconceived notions regarding both elements of Match-Up, that is, the
participants may have assumed that the advertisement was included in the study because
there was a high level of congruency. This may have resulted in the slightly negatively
skewed responses. This speculation cannot be validated without further analyses
involving different types of endorser-product match-ups. Because of the lack of
discriminant validity among the dimensions in Match-Up and the fact that the purpose of
the study was to test the overall relationships in the model (i.e., the relationships with the
second-order latent variable Match-Up), the items from the three first-order latent
variables were allowed to load directly on Match-Up as a first-order latent variable for
the remainder of the analysis. Although the factors were not distinct in the initial analysis,
when the factor loadings on Match-Up as a first-order latent variable were examined
there was some indication that there may be more than one factor, as previous researchers
have stated (e.g. Baker & Churchill; Friedman & Friedman; Kamins, 1990; Lynch &
Schuler; Till & Busler). Therefore, the lack of discriminant validity between the original
factors may stem from a number of possibilities, such as preconceived notions of the
athlete and/or product, the wording of the items, the sample, the advertisement, and the
endorser. Therefore, this issue needs further examination.
Based on theory (Sweeney & Soutar, 2001), the dimensions of Quality and
Emotion are not similar. However, the measurement model showed a lack of discriminant
validity between these two factors. After further examination, it was determined that
there may be a distinct relationship between the factors that was not previously detected.
The factors may be related in that one is cognitive (Quality) and one is affective
(Emotion). If, as previous researchers have stated (e.g., Lavidge & Steiner, 1961),
cognition precedes affect, this may account for the high correlation between the two
factors. Therefore, instead of being a part of the dimensions that represent Perceived
Value, the Emotion construct may be the affective component and the consequence of, or
response to, the cognitive perception of the value of a product. Also, problems may have
arisen from the way that the items were stated. This may account for the problems that
the PERVAL scale has had in the past (e.g., Lee et al., 2005). As previously mentioned,
the two factors should be measuring different constructs, and thus should not be
correlated due to the content of the items. Therefore, if there is a hierarchical relationship,
with Price/Quality/Social being cognitive dimensions and Emotion being an affective
dimension, that might explain the high correlation between the two factors and alleviate
the problems associated with that scale. However, for the purposes of testing the
proposed structural model, the Emotion subscale was eliminated from the analysis.
Social, the other factor along with Emotion, that Sweeney and Soutar (2001) added
to represent Perceived Value, had some problems as well. Although the construct had an
adequate AVE value and was distinct from the other factors, only two of the items loaded
at the set criteria of .707 or higher (Hair, Anderson, Tatham, & Black, 2002). Upon
review, this may be due to inconsistent wording of the items. Some of the items asked the
participant to take ownership (e.g., "would improve the way I am perceived by others")
while others did not (e.g., "would give its owner social approval"). Therefore, the
wording of the questions should probably be modified for use in future studies.
Another aspect of the instrument that may benefit from a reevaluation is Purchase
Intention. Upon further examination of the factor, the mean score of the four items was
3.91 on a 7-point Likert-type scale. Therefore, the average response was just below
neutral (towards slightly disagree). This may be due to the fact that the statements were
worded very strongly. The item with the highest mean referred to a "consideration" of a
purchase, whereas the rest of the items were stronger statements (i.e., "will purchase").
As a result, this may not be an adequate measure for middle to high-end products.
Marketers may intend to create an interest, not a definite purchase intention, as a
consumer will most likely not make a purchase decision based on an advertisement,
especially with only one exposure to the promotion. Therefore, it may be beneficial to
inquire into a consumer' s intention to examine or search out middle to high-end products
(e.g., digital cameras, automobiles, season tickets) as a result of an advertisement.
Focusing on an individual's "consideration of purchase" may provide marketers with a
greater understanding of the actual impact of athlete endorsers on consumer' s purchase
intentions, especially in a highly competitive market where consumers are looking for a
way to differentiate among brands.
The Einal structural model included 42 items within 5 factors: Athlete
Identification, Sport Identifieation, Match-Up, Perceived Value (Quality, Price, Social),
and Purchase Intention. The model (Figure 3) fit adequately well, explained 54% of the
variance in Purchase Intention, and examined four main relationships: Identifieation with
both Sport and Athlete to Match-Up, Identifieation with Athlete to Purchase Intention,
Match-Up to Perceived Value, and Perceived Value to Purchase Intention. Identification
had relatively no influence on any aspect of the model. Previous investigations did not
include this relationship, so it should be reexamined in future studies. The relationship
between Athlete Identifieation and Purchase Intention explained less than one percent of
the variance in the model as well. However, this is not reflective of previous findings
regarding identification in sport. Previous studies have found that identification does
influence purchase intention (e.g., Kwon et al., 2005; Robinson & Trail, 2005; Robinson
et al., 2004; Trail et al., 2003). The finding may be due to the specific athlete used in the
study, the way the questions were phrased, and/or there is no overall influence of
identification with an athlete on an individual's purchase intentions regarding an
endorsed product. Although examining identification in this context is new to academia,
practitioners have used a criterion similar to identification when selecting athlete
endorsers for use in campaigns (Tenser, 2004). Therefore, it would be beneficial to
continue examining the importance of this factor in the future.
The amount of variance that Sport Identification and Athlete Identification
combined to explain in Match-Up was extremely low (7%). This was surprising, as
previous research did find that a consumer tends to react to his or her perception of the
talents or traits of an athlete endorser (Braunstein & Zhang, 2005a), which was part of the
rationale behind this relationship. Athlete (Player) Identification was one of the Points of
Attachment examined by previous researchers (Fink et al., 2002; Kwon et al., 2005;
Robinson & Trail, 2005; Robinson et al., 2004; Trail et al., 2003), who found that these
factors did influence consumer purchase (attendance) intentions. Identification has not
been involved in studies similar to the research design of the current investigation.
Findings of the current study indicated that identification with the sport in general may
not influence a consumer' s perception of the congruency between the endorser and the
product. The small amount of variance explained by the relationships may mean that
identification is not an integral aspect of the decision-making process, or it could be the
result of a number of different factors. For example, the athlete endorser used in this
study may not have elicited the same level of identification as other endorsers. It is
important to use an endorser that a specific market relates to (Rein, Kotler, & Stoller,
1997; Stevens et al., 2003; Stotlar, Veltri, & Viswanthan, 1998). Perhaps the endorser
used in this advertisement was not appropriate for the Generation Y market. The only
way to determine the relevance of identification in this relationship is through continued
The model did explain 3 8% of the variance in the relationship between Match-Up
and Perceived Value. While it has been said that an individual's attitude is not easily
altered, marketing practices have been found to moderately influence a potential
consumer' s perception of a product (MacKenzie, Lutz, & Belch, 1986). Although the
current Eindings of the current study supported this relationship, the amount of variance
explained may be inflated due to similar items being included in both scales. After further
examination of the Trustworthiness sub scale of the SASP CP and the Quality sub scale
of the PERVAL scale, there are a number of items that may have caused the exaggeration
of the importance of the relationship (e.g., Trustworthiness: quality, dependability,
reliability; Quality: would perform consistently, is high-quality, is well-made). Therefore,
the items will need to be reexamined if the scales are to be used in conjunction with one
another in future studies. The aforementioned alterations to the congruency and brand
perception constructs may also strengthen these Eindings in the future.
The largest amount of variance explained by the model was in the relationship
between Perceived Value and Purchase Intention (52%). That is, the participant' s
perception of the value of the product likely influenced his/her intentions of purchasing
the product. The Eindings of this relationship coincided with previous findings, as a
consumer' s perception of the value of the product has been found to influence purchase
intentions (e.g., Lee et al., 2005; Sweeney & Soutar, 2001). The Emotion subscale,
perceived as an important aspect of a consumer' s perception of a product (Lee et al.,
2005; Sheth, Newman, & Gross, 1991; Sweeney & Soutar, 2001), was eliminated from
the structural model. However, future studies may find that this factor plays a key role in
mediating the relationship between Perceived Value and Purchase Intention (e.g.,
Lavidge & Steiner, 1961; Palda, 1966; Ray, 1973). Very importantly, the model
explained a large amount of variance in Purchase Intention (54%), which provides
evidence that the factors included in the model were integral to the decision-making
The model provided presented a potential foundation for marketers to use when
evaluating the use of athlete endorsers in the future. This initial investigation, that
included identification as a potentially influential factor, may provide practitioners with a
basis for the selection of athlete endorsers. While academicians have not included
identification (athlete and/or sport) in their examination of the effectiveness of athlete
endorsers, practitioners have noted that these factors play an important role in their
selection of endorsers. Marketers have used selection criterion including: "is hottest new
star in his/her sport," "has been playing or played the sport for a long time,"
"plays/played for one of my favorite teams," and "plays/played sport I follow" (Tenser,
2004). They have also examined factors such as the sport that the athlete is involved in
(sport that your target market likes) and the individual's level of success (success on the
field and post-sports career) when making final selections (Tenser). While practitioners
have begun to look into factors that lean toward identification, they have not truly
examined it as a part of this process. Therefore, a model that further examines
identification's role in the overall decision-making process may help practitioners to
determine if identification is integral in positively altering a consumer' s perception of the
product, ultimately ending in a purchase intention. In this study, the participants were not
highly identified with the athlete or sport involved in the campaign, and the athlete's
appearance in the advertisement did not make an impact on the participant' s purchase
intentions. If the sample was the company's target market, perhaps the athlete was not the
appropriate endorser to select for their campaign. However, if future studies show that
identification does impact purchase intentions, it may benefit an organization if their
executives take this into consideration when designing a marketing campaign. For
example, an expanded use of focus groups to determine levels of identification (athlete
and sport) and belief of product-endorser congruency may provide a better platform for
an organization interested in selecting an endorser that has the best potential to reach a
Because of the purpose of the study, the composition of the sample, the format of
the questionnaire, the advertisement (i.e., only one athlete and one non-sport product),
and the exploratory nature of the study, the findings can only be viewed as part of the
continuing development and validation process for the model. Although the model
provided an adequate fit to the data, continuous improvement is necessary to obtain the
fit indices that are desired. According to Baumgartner and Jackson (1999), in addition to
examining the content and construct validity of an instrument, criterion and predictive
validity should also be evaluated in future studies. Therefore, future studies with other
samples representing different populations are necessary to further validate the model.
When studying the influence of athlete endorsers in the future, it would be beneficial to
reexamine the relationships proposed in this model. Although the sample size was
adequate for the number of parameters in the model, it would be beneficial to examine
the relationships using experimental design with different types of advertisements (e.g.,
informative, humorous), athlete endorsers (e.g., high-profile, low-profile), and products
(e.g., low cost, high cost / low-involvement, high-involvement / sport, non-sport). Using
experimental design with a hypothetical product or a product that is new to the market
will provide researchers with the opportunity to examine participant' s perception of the
product before and after exposure to the advertisement. It may also be beneficial to
eliminate the screening questions in order to observe all individuals that may have
exposure to the advertisement, and not just those that already know the endorser and/or
the product. Additionally, a study that examines the consumer' s attitude toward the
advertisement, product/brand, and athlete in a very simple way (like/dislike) in order to
observe the participant' s basic feelings regarding the effectiveness of athlete endorsers
could prove useful in reevaluating the model.
Pre-advertisement attitude toward the product influences both initial purchase
intention and post-advertisement attitude toward the product. However, the latter
relationship could be mediated by congruency between the endorser and the product.
Both the initial purchase intention and the post-advertisement attitude toward the product
would be related to post-advertisement purchase intention. This was not examined in the
current study. In future studies, researchers should consider using experimental design to
test attitude change as well as within and between group differences. Once the model has
been reevaluated, a multigroup structural equation modeling technique could be used to
examine the differences between male and female athlete endorsers and the differences
between athlete endorsers as a result of their level of athletic star power. This would
allow for the use of a control group to determine if the relationships are as influential
with the advertisement as without it.
It may also be beneficial to reexamine the model with attitude presented in a
hierarchical structure (i.e., cognition, affect, and conation; e.g., Lavidge & Steiner, 1961;
Palda, 1966; Ray, 1973). As previously stated, some of the Perceived Value factors
(Quality, Price, and Social) may examine cognitive dimensions, while the fourth factor
(Emotion) looks at an affective dimension. Restructuring the model in order to examine a
potential hierarchical structure may provide a solution for the lack of discriminant
validity between Emotion and Quality (see Figure 4) as well as the problems that the
PERVAL scale has faced in the past (e.g., Lee et al., 2005).
In order to obtain a greater understanding of the impact of athlete endorsers on
consumer' s intention to purchase endorsed products, future studies should also consider
an alternative way to examine the Match-Up Hypothesis (e.g., Kahle & Homer, 1985;
Kamins, 1989; Kamins, 1990). Even though the factors were not distinct in this study,
previous studies have found unique factors that they believe to be an integral aspect of
the endorser concept. In addition to the reanalysis of product-endorser congruency,
identification must be further explored as well. Even though the model did not explain a
large amount of variance regarding the relationships between Identification (Athlete and
Sport) and Match-Up, identification has been proven to increase consumer purchase
intentions in the past (e.g., Kwon et al., 2005; Robinson & Trail, 2005; Robinson et al.,
2004; Trail et al., 2003). While these studies have focused on purchase intentions
regarding event attendance, marketers are using a number of the same variables when
selecting athlete endorsers (e.g., Tenser, 2004). With the purpose behind the use of
athlete endorsers stemming from an interest in increasing brand recognition, consumer's
identification with the athlete or sport may prove to be an integral part of the relationship
leading to purchase intentions. Although this was not the case in this study, the
importance of identification must be further examined to determine its relevance in the
selection of endorsers. With the continued use of athlete endorsers in advertisements and
promotional tools, additional research regarding both Match-Up and Identification will be
beneficial to practitioners in the selection of appropriate endorsers for their products.
In summary, the model in this analysis provided an initial examination into some of
the factors that may influence the use of athlete endorsers. Although alterations to the
model are recommended, the findings do provide a framework (based on identity theory
and attitude theory) for additional analyses. Both practitioners and academicians may
adapt and adopt the model to assess the influence of these factors on consumer purchase
intentions regarding endorsed products. Mailed, telephone, in-person, or online
quantitative and qualitative questionnaires may be used to further assess these
relationships and their effectiveness with various types of endorsers and products. This
may provide further evidence for or against the use of athlete endorsers to influence
consumer purchase intentions.
Figure 4. Recommended Model for Future Research
ITEM CORRELATIONS FOR THE MEASUREMENT MODEL
Al S1 A2 S2 S3 A4 S4 El T1 Il E2 T2 Il E3 T3 IS
0.258 0.161 1.000
0.199 0.108 0.521
0.168 -0.048 0.402
0.106 0.038 0.310
0.149 0.064 0.347
0.125 -0.047 0.284
0.125 0.113 0.409
0.113 0.076 0.416
0.149 -0.017 0.269
0.148 0.072 0.364
0.050 0.044 0.369
0.065 0.015 0.279
0.181 0.101 0.370
0.111 0.105 0.385
0.242 0.056 0.316
0.205 0.097 0.425
0.178 0.089 0.412
0.166 0.116 0.332
0.154 0.053 0.303
0.143 0.019 0.329
0.084 -0.006 0.106
0.041 0.002 0.210
0.088 0.067 0.130
0.179 0.099 0.218
0.153 0.055 0.231
0.150 0.034 0.237
0.134 0.105 0.237
0.142 0.042 0.275
0.180 0.099 0.205
0.202 0.031 0.269
0.152 0.052 0.247
0.124 0.074 0.169
0.077 0.020 0.168
0.138 0.061 0.218
KEY: A =Athlete, S =Sport; E =Expertise; T =Trustivorthiness; I= Image, EM= Emotion; Q =Quality; P =Price; S =Social; PI =Purchase Intention