Content uploaded by Younghan Lee
Author content
All content in this area was uploaded by Younghan Lee on Jun 12, 2022
Content may be subject to copyright.
523
ARTICLE
Journal of Sport Management, 2015, 29, 523 -538
http://dx.doi.org/10.1123/jsm.2014-0195
© 2015 Human Kinetics, Inc.
Younghan Lee is with the Department of Kinesiology, Mis-
sissippi State University, Mississippi State, MS. Jakeum Koo
is with the Department of Sport Management, University of
Massachusetts, Amherst, MA. Address author correspondence
to Younghan Lee at yl690@msstate.edu.
Athlete Endorsement, Attitudes, and Purchase Intention:
The Interaction Effect Between Athlete Endorser-Product
Congruence and Endorser Credibility
Younghan Lee
Mississippi State University
Jakeun Koo
University of Massachusetts
The current study used a 2 × 2 analysis to explore the effect of athlete endorser-product congruence and
endorser credibility on consumer responses, such as attitude toward the advertisement, attitude toward the
brand, and purchase intention. Real people and actual brands were used as stimuli to enhance external valid-
ity and generalizability. Research results conrmed the interaction effects between athlete endorser-product
congruence and endorser credibility on three specic consumer responses. The research further examined and
identied the indirect path from attitude toward the advertisement and purchase intention, mediated by attitude
toward the brand. The ndings from the research ll gaps in the literature and extend the body of knowledge
in endorsement studies in general and sport celebrity-endorsement studies in particular.
Keywords: athlete endorsement, endorser credibility, the match-up hypothesis, endorser-product t
The existing literature indicates that scholars have
focused their research primarily on the match-up hypoth-
esis and endorser credibility in the domain of celebrity-
endorsement study. These two constructs have received
heightened attention for several decades because of their
positive effects on advertising outcome. The concept
of the match-up hypothesis suggests that endorsement
is most effective when congruence exists between the
image of an endorser and an endorsed product. (Kahle
& Homer, 1985; Kamins, 1990; Ohanian, 1991; Till &
Busler, 2000). Endorser credibility, on the other hand,
highlights the fact that a credible celebrity endorser is
more likely to generate positive consumer responses than
a less credible endorser (Goldberg & Hartwick, 1990;
Siemens, Smith, Fisher, & Jensen, 2008).
Furthermore, the construct of endorser credibility
is derived from source-credibility (Hovland & Weiss,
1951) and source-attractiveness models (McGuire,
1985). Although the source-credibility model argues that
a message is inuenced by the expertise and trustwor-
thiness of the source (Hovland, Janis, & Kelley, 1953),
the familiarity, likability, and similarity of the source
are the key elements of the source-attractiveness model
(McCracken, 1989; McGuire, 1985). Although the body
of knowledge in this particular aspect has been extended
through empirical examinations (Atkin & Block, 1983;
Friedman & Friedman, 1979), the full spectrum of the
endorser-credibility effect is still not fully explained
especially in the case when endorser-product congru-
ence is not considered jointly in the endorsement process
(McCracken, 1989).
McCracken (1989) specically claimed that the
source models are limited in terms of capturing all facets
of the celebrity-endorsement effects because congruence
between the endorser and the product cannot be measured
in these models. Further studies indicate that endorse-
ments based on a celebrity’s attractive or credible image
are persuasive only for certain products (Kim & Na,
2007). This nding additionally highlights the limitation
of the source models explaining the endorsement effect
without evaluating how a celebrity’s image interacts with
different products (McCracken, 1989) and inuences
consumer behaviors.
Despite the abundance in research pertaining to
these particular constructs, the match-up hypothesis and
endorser credibility, a majority of the studies were con-
ducted on separate research occasions without consider-
ing the two constructs’ interaction effects on advertising
outcome measures. This lack of research effort limits
524 Lee & Koo
JSM Vol. 29, No. 5, 2015
scholarly understanding. It further handicaps the full use
and implementation of the concepts in practice. Hence,
these gaps in the literature warrant further research in
this particular aspect.
Additional limitations in terms of generalizability
and realistic application of the match-up effect exist
because of the use of ctitious endorsers in previous
research settings. For instance, Koernig and Boyd
(2009), and Till and Busler (2000) used ctitious g-
ures as endorsers to examine the effectiveness of athlete
endorser-product congruence on consumers’ attitudes and
purchase intentions. Other researchers have conducted
similar studies in the context of sports using ctitious
gures (Cunningham, Fink, & Kenix, 2008). Scholars
have used ctitious gures in their research experiments
mainly because this approach increases internal validity
by minimizing confounds (Till & Busler, 2000). However,
not using a genuine celebrity may also affect research
outcomes by decreasing external validity in real situations
(Till & Busler, 2000; Cunningham, et al., 2008).
These two different approaches naturally generate
competing positive and negative results. Meanwhile, it
may be important at this stage in the celebrity-endorse-
ment domain to conduct research using both real-life
figures and actual product brands and compare the
results with those from previous research in which only
ctitious gures were used. This attempt may generate
signicant ndings that may add meaningful insights to
understanding the broader magnitude of sport celebrity-
endorsement studies.
The purpose of this study, therefore, is to examine
the interaction effect of endorser-product congruence and
endorser credibility on consumers’ attitude and purchase
behavior by using real athlete celebrities and actual
brands. The current study further investigates the indirect
and direct paths between consumers’ attitude toward the
advertisement and purchase intention, including the paths
associated with attitude toward the brand.
Literature Review
The Match-Up Hypothesis from Social
Adaptation Theory and Associative
Learning Theory
From the advertising practitioners’ perspective, there
should be a tight t between the celebrity endorser and
the endorsed brand. In academia, the match-up hypothesis
has been discussed as one of the more important topics
in celebrity-endorsement studies. To conceptualize the
match-up hypothesis, researchers have cited two theories:
social adaptation theory (Kahle & Homer, 1985; Kamins,
1990) and associative learning theory (Till & Busler, 2000).
Kahle and Homer (1985) indicate that the match-
up hypothesis of celebrity selection ts well with social
adaptation theory. This theory posits that the adaptive
signicance of information will determine its impact
and assumes that an attractive spokesperson in the adver-
tisement generates in the consumers’ mind an effective
source of information for an attractiveness-related prod-
uct. From this perspective, Kamins (1990) argues that
physical attractiveness of a celebrity endorser positively
affects consumers’ evaluations for a product used to
enhance one’s attractiveness.
Another concept used to develop the match-up
hypothesis in celebrity endorsement is associative learn-
ing theory, which explains that an associative network
structure in a person’s memory contains a certain pattern
of connected concepts (Collins & Loftus, 1975). In the
context of celebrity endorsements, consumer experi-
ence is related to products and spokespeople make links
between them in memory. According to Till and Busler
(2000), belongingness, relatedness, t, and similarity are
important factors that are easily established between asso-
ciative connection and the endorser. That is, a connection
between a product and a spokesperson is perceived as an
associative link in the consumers’ mind.
The Match-Up Hypothesis in the
Perspectives of Attractiveness and
Expertise
To demonstrate the effectiveness of the endorser-product
congruence on consumer responses, some researchers
have examined whether congruence between attractive
image of an endorser and an attractive-related product
has a positive impact (Kahle & Homer, 1985; Kamins,
1990; Misra & Beatty, 1990). Others have sought to
demonstrate the use of an endorser with a product that
is in line with the endorser’s expertise (Koernig & Boyd,
2009; Ohanian, 1991; Till & Busler, 2000).
Several researchers offered a contingency access
to studying source attractiveness and its impact on both
products and advertisement evaluations before the match-
up hypothesis was conceptualized (Joseph, 1982). The
proposition is that a celebrity endorser’s physical attrac-
tiveness could enhance the assessments on the basis of
both brand and advertisement if the attributes of the prod-
uct matched the image being delivered by the endorser.
Likewise, match-up research has focused on examining
the effectiveness of congruence between a physically
attractive endorser and an attractive product (Kahle &
Homer, 1985; Kamins, 1990; Misra & Beatty, 1990).
Kahle and Homer examined how an endorser’s physical
attractiveness affects consumers’ responses when reacting
to an attractiveness-related product. The authors found
that, when dealing with beauty-related products, endors-
ers’ physical attractiveness positively affected consumers’
purchase intention, brand recall, and attitude toward the
brand. However, the authors did not consider products
that are not associated with attractiveness.
Meanwhile, Kamins (1990) examined both products
related to attractiveness and unrelated to attractive-
ness. Research ndings indicated that endorser-product
attractiveness match enhanced the spokesperson’s cred-
ibility and attitude toward the advertisement. However,
Athlete-Product Congruence and Endorser Credibility 525
JSM Vol. 29, No. 5, 2015
a signicant endorser-product congruence effect was
statistically insignicant in terms of attitude toward the
brand and purchase intention. Misra and Beatty (1990),
on the other hand, examined the congruence between a
celebrity’s perceived image and the endorsed product
in which both brand recall and affect were signicantly
increased during congruent conditions between endorser
and the brand (Misra & Beatty, 1990).
Ensuing studies have examined the role of expertise
leading up to the match-up effect. Ohanian (1991) indi-
cated that expertise is more useful than attractiveness with
respect to the spokesperson-product congruence effect.
Till and Busler (2000) further examined the effectiveness
of the endorser-product match, considering the aspect of
the spokesperson’s expertise by using an athlete endorser
as the research stimulus. The authors compared attractive-
ness with expertise as a match-up factor in two experi-
ments. Expertise was conrmed to be a more signicant
dimension than physical attractiveness when measuring
the effectiveness of the endorser-product congruence.
Specically, a signicant Product × Endorser interaction
was reported for attitude toward the brand but not for
purchase intention when the spokesperson’s expertise was
used as a congruence factor. Meanwhile, when physical
attractiveness was used as a match-up dimension, there
was no signicant interaction in terms of purchase inten-
tion or attitude toward the brand.
Endorser Credibility Based on
Source-Credibility Model and Source-
Attractiveness Model
According to the source-credibility model, individuals
tend to be persuaded when they perceive the source of
information as credible (Hovland et al., 1953; Ohanian,
1991; Petty & Cacioppo, 1986). Highly credible sources
generate attitudes that are more positive as well as more
behavioral compliance than less credible sources (Hov-
land & Weiss, 1951; Ohanian, 1991). Source-credibility
models focus on dening key components of source
credibility (Ohanian, 1990) such as expertise and trust-
worthiness (Hovland et al., 1953; Hovland & Weiss, 1951;
McCracken, 1989; Ohanian, 1991). For this reason, a
great number of studies related to source credibility have
been produced. For example, consumers’ trustworthiness
of the source of information declines when advertisers
use misleading practices to share product information.
As a result, consumers are less likely to purchase the
advertised product (Romani, 2006). Studies have also pro-
vided evidence that both trustworthiness and expertise are
fundamental elements that affect viewers’ perception and
attitudes (Harmon & Coney, 1982; Wu & Shaffer, 1987).
Expertise can be dened as an individual’s experi-
ence or knowledge of the truth with regard to a specic
topic, including endorser’s skills, knowledge, or special
experience (Hovland et al., 1953). Furthermore, expert
sources affect the audience’s perceptions of the brand’s
quality. A source perceived to have more expertise is
more persuasive and generates intentions that are more
positive among the audience to purchase the product
(Ohanian, 1991).
Trustworthiness relates to the individual’s incentive
to tell the truth regarding the topic (Hovland et al., 1953).
According to Smith (1973), consumers consider untrust-
worthy spokespersons to be doubtful sources regard-
less of other qualities they may have. In the meantime,
advertisers exploit trustworthiness by using endorsers
who are widely perceived to be honest, believable, and
dependable (Shimp, 1997). Arguing that trustworthiness
is the primary determinant of source credibility, Fried-
man, Santeramo, and Traina (1978) sought to determine
which source attributes correlate with trust. They found
likeability to be the most important attribute of trust.
Hence, a well-known and well-liked celebrity is neces-
sary when choosing someone to participate in advertising
endorsements.
The source attractiveness model is also derived from
social psychological studies. According to McGuire’s
(1985) model, a message’s effectiveness stems primarily
from familiarity, likability, and similarity of the source.
McGuire’s model indicates that when consumers know,
like, and are similar to the source, the source is recognized
as attractive and, consequently, persuasive. The stronger
these attributes, the more attractive he or she will be to
the consumer and the easier it will be for the consumer
to accept the message (Tellis, 2003). Several empirical
studies have demonstrated that source attractiveness
generates positive consumer responses. The effective-
ness of physically attractive source was supported with
respect to changing consumers’ attitudes (Caballero &
Pride, 1984; Debevec & Kernan, 1984) and enhancing
purchase intention (Petroshius & Crocker, 1989; Petty &
Cacioppo, 1981; Tingchi Liu, Huang, & Minghua, 2007).
To identify the effectiveness of endorser credibility,
Ohanian (1990) constructed scales reecting the attributes
of the three components: expertise, trustworthiness, and
attractiveness. As indicated previously, expertise relates
to the endorser’s perceived knowledge, skills, or experi-
ence that enables him or her to offer correct information
(Hovland et al., 1953), whereas trustworthiness relates to
consumers’ belief that the individual can give information
in an unbiased and honest manner. Finally, attractiveness
is determined by whether the consumers nd the indi-
vidual to be physically attractive/likable (Ohanian, 1991).
Thus, knowledgeable, sincere, and physically attractive
endorsers are perceived to be credible and, consequently,
promote consumers’ positive attitudinal and behavioral
responses (Ohanian, 1991).
Previous studies have discovered that credible
endorsers positively affect consumers’ attitudes toward
the advertisement and the brand (Atkin & Block, 1983;
Goldberg & Hartwick, 1990). Additional studies have
found that attitude toward the advertisement, attitude
toward the brand, and purchase intention (Lafferty &
Goldsmith, 1999; Siemens et al., 2008) are inuenced
by endorser credibility in advertisements.
526 Lee & Koo
JSM Vol. 29, No. 5, 2015
Hypotheses
The Match-Up and Credibility Effects of
Athlete Endorsement
According to Fink, Cunningham, and Kensicki (2004),
athlete attractiveness and expertise are positively related to
athlete-event t, and the effects of athlete expertise on per-
ceptions of athlete-event t were stronger than the effects
of athlete attractiveness. Likewise, athletes as spokespeople
have frequently been used to examine the role of expertise
in the match-up effect in advertising because they can
be well matched with products related to sports such as
sport drinks, energy bars, and running shoes. These athlete
endorsers are perceived as more credible sources than are
nonathlete endorsers when dealing with a sport-related
product (Nataraajan & Chawla, 1997).
Koernig and Boyd (2009) examined endorser (i.e.,
athlete endorser versus nonathlete endorser) by product
(i.e., sport brand vs. nonsport brand) interaction, nding
evidence to support the effectiveness of athlete endorser-
product congruence on attitude toward the advertise-
ment, attitude toward the brand, and purchase intention.
Consistent with previous studies (Ohanian, 1991; Till &
Busler, 2000), the present research focuses on the role
of expertise in enhancing the perceived t between a
spokesperson and a product to measure the match-up
effect, using athlete endorsers in the stimuli.
Several studies that measured the effectiveness of
endorser credibility were examined in the domain of sport
management because of advertisers’ frequent use of sport
stars to endorse their products in practice. A number of
studies have used athletes as the stimuli to demonstrate
the endorser-credibility effect on consumer responses.
To verify the effectiveness of endorser credibility, Laf-
ferty and Goldsmith (1999) used an athlete (Florence
Grifth-Joyner) as a high- credibility endorser and an
actor (Roseanne Barr) as a low-credibility endorser for
athletic shoes. Meanwhile, Lafferty, Goldsmith, and
Newell (2002) used an athlete (Tiger Woods) as the most
credible spokesperson and an actor (Wayne Knight) as the
least credible endorser for running shoes. Although both
studies veried the credibility effect of athlete endorse-
ments on consumer responses, they did not consider the
impact of a low-credibility athlete endorser.
The Relationship Between Athlete
Endorser-Product Congruence and
Endorser Credibility
Both academic research and practical ndings have
demonstrated the effectiveness of celebrity endorsements.
However, certain factors might decrease the effectiveness
of celebrity endorsements in advertising. One factor
that may reduce the impact of the celebrity effect is low
congruence between celebrity and the product. A second
factor is the lack of celebrity’s credibility. Likewise,
McCracken (1989) suggested that the effectiveness of
source credibility in celebrity endorsement would interact
with celebrity-product congruence.
Despite the importance of these two variables (i.e.,
endorser-product congruence and endorser credibility),
previous studies have failed to focus on the joint effects.
Kim and Na (2007) examined the effectiveness of athlete
endorsers’ advertisement in the role of compatibility and
credibility. Consumer responses were measured in a high-
compatibility condition (i.e., an athlete endorsed sports
shoes) and a low-compatibility condition (i.e., an athlete
endorsed perfume). The research results indicated that
credibility effects of athlete celebrity on product attitude
generated statistically signicant results when the image
of the athlete celebrity and the product matched well.
However, the effect of credibility on product attitude was
statistically insignicant when incongruence occurred.
These results suggest the likelihood of interaction effects
between endorser-product congruence and endorser cred-
ibility on advertising outcome measures such as attitude
toward the advertisement, attitude toward the brand, and
purchase intention. In light of this, endorser-credibility
effect on consumers’ reaction will be based on endorser-
product congruence.
Hypothesis 1: There will be an interaction between
endorser-product congruence and endorser credibil-
ity on attitude toward the advertisement.
Hypothesis 2: There will be an interaction between
endorser-product congruence and endorser cred-
ibility on attitude toward the brand.
Hypothesis 3: There will be an interaction between
endorser-product congruence and endorser cred-
ibility on purchase intention.
Mediation Effect of Attitude
Toward the Brand
Traditionally, in the domain of marketing and advertising
research, attitudes toward the advertisement, attitudes
toward the brand, and purchase intention have been
measured to test the effectiveness of advertising. Attitude
toward the advertisement and brand attitude have been
identied, in particular, as important constructs mediat-
ing the advertising effects on purchase intention. In this
context, various studies have indicated the indirect path
from attitude toward the advertisement to brand attitude,
subsequently inuencing purchase intention (Goldsmith,
Lafferty, & Newell, 2000; La Ferle & Choi, 2005;
MacKenzie & Lutz, 1989). Therefore, we put forward
the following hypotheses.
Hypothesis 4: Positive attitude toward the advertise-
ment positively inuences attitude toward the brand.
Hypothesis 5: Positive attitude toward the brand
positively inuences purchase intention.
Hypothesis 6: Attitude toward the brand mediates the
relationship between attitude toward the advertise-
ment and purchase intention.
Athlete-Product Congruence and Endorser Credibility 527
JSM Vol. 29, No. 5, 2015
Method
Participants and Study Design
A total of 143 college students (46% male, 54% female;
71.3% White; mean age = 21.17 years, SD = 2.79; age
range = 17–32 years) participated in the experiment
voluntarily. A random sampling approach was used in
the subject selection process. Our analysis used a 2 ×
2 between-subjects factorial design: (a) congruence
between the image of a celebrity and the image of a prod-
uct (congruence and incongruence) and (b) credibility of
celebrity endorser (high and low). Participants received
one of four advertisement manipulations: high credibility,
high congruence (n = 36); high credibility, low congru-
ence (n = 35); low credibility, high congruence (n = 37);
and low credibility, low congruence (n = 35).
Pretests
A preliminary analysis was performed before the main
experiment to identify the appropriate endorsers pertain-
ing to high-credibility and low-credibility conditions. The
current research used the Harris Poll 2014 as the reference
to create a shortlist of celebrity athlete endorsers to be
examined for the high-credibility condition. According
to this poll, LeBron James was selected as the favorite
sport star in North America, followed by Michael Jordan
and Derek Jeter. Although Michael Jordan was ranked
higher than Derek Jeter on the list, he was excluded from
the pretest experiment because the current study consid-
ered only athletes currently playing in their respective
sports. Hence, Lebron James and Derek Jeter were nally
selected for the pretest.
As for low-credibility measures, Michael Vick and
Tiger Woods were selected. Michael Vick was believed
to have a persistent negative image that was caused by an
incident for which he was eventually sentenced to several
months in prison. Tiger Woods, on the other hand, was
involved in a sex scandal that naturally created a negative
perception about his personal image.
The participants rated the level of congruence
between celebrities and product types on a 7-point
semantic differential scale on which 1 was the low or
negative end and 7 was the positive or high end. The
three items were not compatible/compatible, not a good
t/a good t, and not congruent/congruent. This scale
was originally developed by Rifon, Choi, Trimble, and Li
(2004). Participants also rated credibility using the same
7-point scale. The nine items were attractive/unattractive,
beautiful/ugly, sexy/not sexy, honest/dishonest, sincere/
insincere, trustworthy/untrustworthy, expert/not an
expert, experienced/inexperienced, and knowledgeable/
unknowledgeable. The work by Ohanian (1990) provided
the basis for this particular scale. Finally, eight different
combinations (i.e., four celebrities × two product types)
were given to the respondents to determine the appropri-
ate match between the athlete celebrities and the products.
The pretest results identied LeBron James as the
most credible gure and congruent source within each
product category, whereas Michel Vick was rated as
the least credible and most incongruent source with the
products. Therefore, LeBron James and Michel Vick
were selected as the stimuli for the high-credibility and
low-credibility conditions, respectively, in the main study.
Detailed results of the pretest are indicated in Table 1.
The current experiment considered two real brands
within the product categories of sport drinks and tooth-
pastes to maximize external validity. Gatorade was chosen
for the high-congruence condition, and Colgate was chosen
for the low-congruence condition; they were considered
appropriate brands for the stimuli because each is one of
the best-selling brands in its respective product category.
It was also expected that the two brands were not signi-
cantly different in terms of familiarity and favorability. For
additional parsimonious support, pretest was conducted to
check the two brands’ familiarity and favorability.
Twenty-one participants rated brand familiarity of
Gatorade and Colgate on a 7-point semantic differential
scale. The two items were not known to me/known to
me and not familiar/familiar (developed by Hirschman,
1986). Meanwhile, favorability of each brand was
measured using the same 7-point scale anchored with
unfavorable/favorable. According to the results of one-
way analyses of variance (ANOVAs), the familiarity of
Gatorade (M = 6.95) was not signicantly different from
that of Colgate (M = 6.81), F(1, 40) = 2.69, p > .05.
Likewise, Gatorade’s favorability (M = 5.10) was not
signicantly different from that of Colgate (M = 5.43),
F(1, 40) = 0.53, p > .05. Thus, Gatorade and Colgate
Table 1 Pretest Results
Celebrity
Endorser-product congruence Endorser credibility
Sport drinks Toothpastes Sport drinks Toothpastes
M SD M SD M SD M SD
LeBron James 6.67 0.44 2.16 1.17 5.70 0.77 3.20 0.96
Derek Jeter 5.98 0.94 1.78 0.75 5.18 0.83 2.96 0.68
Tiger Woods 5.64 1.57 1.97 1.16 4.63 1.39 2.38 0.89
Michael Vick 4.48 1.57 1.50 0.71 3.53 1.29 1.91 0.87
Note. M = average of three 7-point scale items; SD = standard deviation.
528 Lee & Koo
JSM Vol. 29, No. 5, 2015
were determined to be appropriate brands for use in the
main experiment.
Stimulus Materials
Full-page color print advertisements for each athlete were
created to embody two different endorser-product congru-
ence levels: Gatorade endorsed by LeBron James, Gatorade
endorsed by Michael Vick, Colgate endorsed by LeBron
James, and Colgate endorsed by Michael Vick. The print
materials included the athlete celebrity’s picture holding
either a basketball (Lebron James) or football (Michael
Vick), a short reference to the athlete (e.g., “LeBron
James, NBA player, spokesperson for Gatorade”), and the
product’s brand logo. The print materials were designed
exactly the same in spacing and layouts, with a large pic-
ture of the athlete’s upper body in which he was wearing
a regular tank-top shirt. The picture was in the center of
the advertisement, and the reference was centered in the
bottom portion. The product’s brand logo appeared on the
upper right side of the advertisement. Any visual objects
or letters, including images of the actual products, signify-
ing teams or afliations associated with the athletes were
omitted from the advertisements to eliminate preference
bias toward certain teams or products.
Procedures
The experiment was conducted in a university classroom.
After introducing the experiment, stimulus materials (i.e.,
print advertisements) and questionnaires were provided to
the participants. The four different types of advertisements
were randomly distributed to the subjects, who were asked
to view the advertisements and complete the questionnaires.
The questionnaires consisted of six sections: (a) attitude
toward the advertisement, (b) attitude toward the brand,
(c) purchase intention, (d) opinions regarding the endorser-
product congruence, (e) opinions regarding the endorser
credibility, and (f) demographic information. Participants
had 10–15 min to complete the questionnaires. Thereafter,
the responses were coded into SPSS for statistical analysis.
Measurement Items and Variables
Endorser-product congruence. Participants were
asked to use a 7-point semantic differential scale to rate
the following items: compatible/not compatible, a good
t/not a good t, and congruent/not congruent. The
original measure was developed by Rifon et al. (2004)
and empirically validated by White, Goddard, and Wilbur
(2009) to measure the congruence between the image of
the endorser and the product.
Endorser credibility. Participants used a 7-point
semantic differential scale to rate the following items:
attractive/unattractive, classy/not classy, beautiful/
ugly, elegant/plain, sexy/not sexy, dependable/
undependable, honest/dishonest, reliable/unreliable,
sincere/insincere, trustworthy/untrustworthy, expert/not
an expert, experienced/inexperienced, knowledgeable/
unknowledgeable, qualied/unqualied, and skilled/
unskilled. The measure was developed by Ohanian
(1990) to assess endorser credibility. Previous studies
have empirically validated this particular scale on several
research occasions (Goldsmith, Lafferty, & Newell, 2000;
La Ferle & Choi, 2005; Siemens et al., 2008).
Attitude toward the advertisement. To measure
attitude toward the advertisement, we asked participants
to use a 7-point semantic differential scale to rate the
following items: good/bad, favorable/unfavorable,
and pleasant/unpleasant (MacKenzie & Lutz, 1989).
This particular scale had been empirically validated
in previous research efforts (Goldsmith et al., 2000;
Lafferty & Goldsmith, 1999).
Attitude toward the brand. To measure attitude
toward the brand, we asked participants to use a 7-point
semantic differential scale to rate the following items:
good/bad, favorable/unfavorable, and positive/negative.
The current study implemented the scale previously
developed by MacKenzie and Lutz (1989).
Purchase intention. To measure purchase intention, we
asked participants to use a 7-point semantic differential
scale to rate the following items: very likely/very unlikely,
probable/improbable, and possible/impossible. The
current study implemented the scale previously developed
by Yi (1990).
Results
Scale Reliability
Cronbach’s α test generated values of .892 to .982 for
all measurement items (see Table 2 for complete data).
These values indicated acceptable internal reliability
based on the standard (.70) suggested by Murphy and
Davidshofer (2001).
Manipulation Check
One-way ANOVAs were conducted to demonstrate whether
the two operationalized endorser-product congruence levels
differed signicantly from each other; and whether the two
manipulated conditions of endorser credibility differed
signicantly from each other. The average rating for high
congruence (i.e., Gatorade) was 4.80, whereas for low
congruence (i.e., Colgate), the average rating was 1.85. This
difference was statistically signicant, F(1, 141) = 129.95, p
< .001. The average rating for high endorser credibility (i.e.,
LeBron James) was 4.33, whereas low endorser credibility
(i.e., Michael Vick) generated a rating of 3.52. The analysis
results indicated a statistically signicant difference, F(1,
141) = 17.01, p < .001.
Preliminary Analysis
Data for the four experimental conditions are shown
in Table 3. Bivariate correlations are shown in Table
4. These correlations were statistically signicant at p
= .01 level, indicating linear relationships among the
dependent variables.
Athlete-Product Congruence and Endorser Credibility 529
JSM Vol. 29, No. 5, 2015
Table 2 Reliability Analysis for Independent and
Dependent Measures
Measures Number of
items Cronbach’s
α
Independent measures
Endorser-product congruence 3 .982
Endorser credibility 15 .948
Dependent measures
Attitude toward the advertisement 3 .892
Attitude toward the brand 3 .955
Purchase intention 3 .916
Table 3 Distribution, Means, and Standard Deviations for Four
Experimental Conditions: Ad Attitude, Brand Attitude, and Purchase
Intention
Ad attitude Brand
attitude Purchase
intention
n M SD M SD M SD
High congruence, high credibility 36 4.72 0.95 4.76 0.89 4.31 1.43
High congruence, low credibility 37 3.84 1.21 4.00 1.16 2.90 1.39
Low congruence, high credibility 35 3.10 1.03 3.66 1.05 3.12 1.45
Low congruence, low credibility 35 2.96 1.10 3.67 1.39 3.03 1.38
Note. M = average of three 7-point scale items; SD = standard deviation.
Table 4 Means, Standard Deviations, and
Bivariate Correlations
Variable 1 2 3
1. Attitude toward the advertisement — — —
2. Attitude toward the brand .69** — —
3. Purchase intention .55** .61** —
M3.67 4.02 3.34
SD 1.28 1.21 1.51
Note. **p < .01.
Multivariate Analysis of Variance
The three dependent variables were entered into a
multivariate ANOVA (MANOVA) with endorser cred-
ibility and endorser-product congruence as independent
categorical variables. Levene’s tests for each dependent
variable were statistically insignicant (all p > .05),
indicating equality of error variances across the treatment
groups on each dependent variable. In addition, Box’s
M test was statistically insignicant (p > .05), indicating
equality of variance/covariance matrices of the multiple
dependent variables across the treatment groups. The
two-way MANOVA interaction between endorser cred-
ibility and endorser-product congruence was statistically
signicant, Wilks’s λ = .942, F = 2.79, p < .05, allowing
separate use of ANOVAs for the three dependent variables
with the protection of the Type 1 error rate.
Hypothesis 1 predicted that the interaction of
endorser credibility and endorser-product congruence
would have an effect on attitude toward the advertise-
ment. The analysis generated statistically signicant
results, F(1, 139) = 4.52, p < .05, supporting Hypothesis
1. In addition, the endorser-credibility effect was more
signicant on attitude toward the advertisement within the
high endorser-product-congruence group than in the low
endorser-product-congruence group, F(1, 71) = 12.07,
530 Lee & Koo
JSM Vol. 29, No. 5, 2015
p < .01. Credibility, in the meantime, had no signicant
effect on attitude toward the advertisement under the
low-congruence condition, F(1, 68) = 0.20, p > .05. The
interaction effect on attitude toward the advertisement is
plotted in Figure 1.
Hypothesis 2 predicted that the interaction of
endorser credibility and endorser-product congruence
would have an effect on attitude toward the brand. The
interaction effect of the two independent variables on
brand attitude was statistically signicant, F(1, 139) =
4.01, p < .05. Therefore, Hypothesis 2 was supported.
Further analysis indicated that endorser credibility had a
signicantly stronger effect on attitude toward the brand
for subjects in the high endorser-product-congruence
group, F(1, 71) = 9.82, p < .01, than in the low endorser-
product-congruence group. Meanwhile, credibility had no
effect on attitude toward the brand under the low-congru-
ence condition, F(1, 68) = 0.01, p > .05. The interaction
effect on attitude toward the brand is plotted in Figure 2.
Hypothesis 3 predicted that the interaction of
endorser credibility and endorser-product congruence
would have an effect on purchase intention. The analy-
sis result indicated a statistically signicant interaction
effect, F(1, 139) = 7.75, p < .01. Therefore, Hypothesis 3
was supported. The research subjects who were exposed
to the high-credibility-endorser condition displayed a
signicantly higher level of purchase intention than those
who were exposed to the low-credibility-endorser con-
dition in the high endorser-product-congruence setting,
F(1, 71) = 18.34, p < .001]. The analysis, however, did
not discover a statistically signicant credibility effect on
purchase intention in terms of the low endorser-product-
congruence condition, F(1, 68) = 0.08, p > .05. The inter-
action effect on purchase intention is plotted in Figure 3.
Figure 1 — Attitude toward the advertisement. Attitude toward the advertisement is graphed as a function of endorser credibility,
separately for high and low product congruence.
Figure 2 — Attitude toward the brand. Attitude toward the brand is graphed as a function of endorser credibility, separately for
high and low product congruence.
Athlete-Product Congruence and Endorser Credibility 531
JSM Vol. 29, No. 5, 2015
Regression Analysis
Hypothesis 4 predicted that positive attitude toward the
advertisement would increase attitude toward the brand.
Hypothesis 5 predicted that positive attitude toward the
brand would enhance purchase intention. Before analyz-
ing the data jointly under these research conditions in a 2
× 2 factorial design, we referred to the guidelines regard-
ing the hierarchical regression model provided by Weaver
and Wuensch (2013), which was based on the original
work by Potthoff (1966). This particular method was used
to analyze the data to examine whether the predictions
varied among the different conditions.
In the rst step, a predictor variable is included;
thereafter, indicator variables and the products of these
indicators with the predictor variable are added in the
second step. Through this analysis, the null hypotheses
that all intercepts and all slopes are the same can be
identied.
Table 5 shows the summary for four simple linear
regression models with attitude toward the advertisement
regressed on attitude toward the brand. The results indi-
cated that attitude toward the advertisement positively
inuenced attitude toward the brand for all conditions (all
p values < .001). Potthoff analysis was conducted to check
whether the four regression models were signicantly dif-
ferent. In the rst step, attitude toward the advertisement
was included as a predictor variable; thereafter, in the
second step, indicator variables and the products of these
indicators with attitude toward the advertisement were
added. Tables 6 and 7 indicate the ANOVA test results
for this particular model and the parameter estimates,
respectively. The R2 values for Steps 1 and 2 of the hier-
archical regression model did not change signicantly,
Figure 3 — Purchase intention. Purchase intention is graphed as a function of endorser credibility, separately for high and low
product congruence.
Table 5 Parameter Estimates for Four Simple Linear Regression Models with
Ad Attitude Regressed on Brand Attitude
Condition Independent
variable
b SE βt p
Congruence level Credibility level
High High Constant 1.76 0.57 — 3.09 .00
Ad attitude 0.63 0.12 0.68 5.35 .00
High Low Constant 1.71 0.51 — 3.34 .00
Ad attitude 0.60 0.13 0.62 4.68 .00
Low High Constant 1.71 0.46 — 3.71 .00
Ad attitude 0.63 0.14 0.61 4.47 .00
Low Low Constant 1.27 0.52 — 2.46 .02
Ad attitude 0.80 0.16 0.65 4.93 .00
Note. SE = standard error.
532 Lee & Koo
JSM Vol. 29, No. 5, 2015
Table 6 ANOVA Summary Table for the Hierarchical Regression Model
with Ad Attitude Entered on Step 1 and Three Condition Indicators and
their Products with Ad Attitude Entered on Step 2
Step
df F p
1 Regression 1 130.75 .00a
Residual 141 — —
Total 142 — —
2 Subset test
Aad 1 35.06 .00b
G1 indicator, G2 indicator, G3 indicator 3 0.24 .87b
Aad × G1, Aad × G2, Aad × G3 3 0.47 .71b
Regression 7 18.61 .00c
Residual 135 — —
Total 142 — —
Note. ANOVA = analysis of variance; dependent variable = attitude toward the brand; Aad = attitude toward
the advertisement. G1 = high congruence/high credibility; G2 = high congruence/low credibility; G3 = low
congruence/high credibility; G4 = low congruence/low credibility.
aPredictors: Constant, Aad; bTested against the full model; cPredictors in the full model: Constant, Aad, G1
indicator, G2 indicator, G3 indicator, Aad × G1, Aad × G2, Aad × G3.
Table 7 Parameter Estimates for the Hierarchical Regression Model with Ad Attitude Entered on
Step 1 and Three Condition Indicators and their Products with Ad Attitude Entered on Step 2
Step Variable
b SE βt p
1 Constant 1.61 0.22 — 7.20 .00
Aad 0.66 0.06 0.69 11.43 .00
2 Constant 1.27 0.43 — 2.95 .00
Aad 0.80 0.14 0.84 5.92 .00
G1 indicator 0.49 0.87 0.18 0.57 .57
G2 indicator 0.44 0.65 0.16 0.68 .50
G3 indicator 0.44 0.65 0.16 0.68 .50
Aad × G1 –0.16 0.21 –0.29 –0.79 .43
Aad × G2 –0.20 0.18 –0.30 –1.11 .27
Aad × G3 –0.17 0.20 –0.20 –0.85 .40
Note. SE = standard error; dependent variable = attitude toward the brand; Aad = attitude toward the advertisement.
F(6, 135) = 0.44, p > .05. Thus, the null hypothesis of
coincident regression lines was not rejected.
The separate analyses for the intercepts and slopes
indicated the homogeneity of the intercepts, F(3, 135) =
0.24, p > .05, and the homogeneity of the slopes, F(3,
135) = 0.47, p > .05 (see Table 6). All pairwise com-
parisons of each condition with the reference condition
(i.e., low congruence, low credibility) were statistically
insignicant (all p values > .05) in terms of intercepts
and slopes (see Table 7). Therefore, the four conditions
can be estimated jointly to test the effect of the attitude
toward the advertisement on brand attitude.
The analysis results for the four simple linear regres-
sion models with attitude toward the brand regressed on
purchase intention is presented in Table 8. The results
indicate that attitude toward the brand positively inu-
enced purchase intention under all conditions (all p values
< .01). Potthoff analysis was further conducted to identify
Athlete-Product Congruence and Endorser Credibility 533
JSM Vol. 29, No. 5, 2015
whether the four regression models were signicantly
different. Attitude toward the brand was included as a
predictor variable in the rst step; thereafter, indicator
variables and the products of these indicators with attitude
toward the brand were added in the second step. Table 9
indicates the results of an ANOVA for this specic model,
and Table 10 gives the parameter estimates.
The R2 change between Steps 1 and 2 of the hier-
archical regression model was statistically insignicant,
F(6, 135) = 2.11, p > .05. Thus, the null hypothesis of
coincident regression lines was not rejected. The sepa-
rate analyses for the intercepts and slopes conrmed the
homogeneity of the intercepts, F(3, 135) = 0.39, p > .05,
and the homogeneity of the slopes, F(3, 135) = 0.94, p
> .05] (see Table 9). The pairwise comparisons showed
that each condition was signicantly indifferent from the
reference condition (i.e., low congruence, low credibility;
all p values > .05) in terms of intercepts and slopes (see
Table 10). Therefore, the four conditions can be estimated
jointly to identify the relationship between attitude toward
the brand and purchase intention.
To test the mediation effect of brand attitude between
attitude toward the advertisement and purchase intention,
we performed four steps based on Baron and Kenny’s
Table 8 Parameter Estimates for Four Simple Linear Regression Models with
Brand Attitude Regressed on Purchase Intention
Condition
b SE βt p
Congruence level Credibility level
High High Constant –0.61 1.02 — –0.60 .56
Brand attitude 1.03 0.21 0.64 4.91 .00
High Low Constant 0.58 0.74 — 0.78 .44
Brand attitude 0.58 0.18 0.48 3.27 .00
Low High Constant 0.74 0.80 — 0.92 .37
Brand attitude 0.65 0.21 0.47 3.09 .00
Low Low Constant 0.57 0.51 — 1.12 .27
Brand attitude 0.67 0.13 0.67 5.15 .00
Note. SE = standard error.
Table 9 ANOVA Summary Table for the Hierarchical Regression Model
with Brand Attitude Entered on Step 1 and Three Condition Indicators
and their Products with Brand Attitude Entered on Step 2
Step
df F p
1 Regression 1 82.69 .00a
Residual 141 — —
Total 142 — —
2 Subset test
Ab1 21.04 .00b
G1 indicator, G2 indicator, G3 indicator 3 0.39 .76b
Ab × G1, Ab × G2, Ab × G3 3 0.94 .42b
Regression 7 14.18 .00c
Residual 135 — —
Total 142 — —
Note. ANOVA = analysis of variance; dependent variable = purchase intention; Ab = attitude toward the
brand; G1 = high congruence/high credibility; G2 = high congruence/low credibility; G3 = low congruence/
high credibility; G4 = low congruence/low credibility.
aPredictors: Constant, Ab; bTested against the full model; cPredictors in the full model: (Constant), Ab, G1
indicator, G2 indicator, G3 indicator, Ab × G1, Ab × G2, Ab × G3.
534 Lee & Koo
JSM Vol. 29, No. 5, 2015
(1986) guideline. The rst regression tested the signi-
cance of the relationship between the initial independent
variable (i.e., attitude toward the advertisement) to the
mediator (i.e., attitude toward the brand). Attitude toward
the advertisement accounted for 48% of variance, β =
0.69, p < .001; thus the rst step of mediation (Hypothesis
4) was supported.
The second regression analysis tested the signi-
cance of the relationship between the initial independent
variable and dependent variable (i.e., purchase intention).
Attitude toward the advertisement explained 30% of
variance, β = 0.55 p < .001; hence the second step was
supported.
The third regression analysis tested the signicance
of the relationship between the mediator and dependent
variable. Attitude toward the brand accounted for 37%
of variance, β = 0.61, p < .001. Therefore, the third step
entails that Hypothesis 5 was supported.
In the last step, after controlling for attitude toward
the advertisement, the effect of brand attitude on purchase
intention was statistically signicant, β = 0.43, p < .001.
Meanwhile, after controlling for brand attitude, the effect
of advertisement attitude on purchase intention was still
signicant, β = 0.25, p < .01. According to Sobel’s (1982)
test, however, the effect of advertisement attitude on
purchase intention was signicantly reduced, z = 7.12,
p < .001, after controlling for attitude toward the brand.
Therefore, Hypothesis 6 was supported.
Discussion
The present study explored how endorser-product congru-
ence and endorser credibility affect consumer responses,
such as attitude toward the advertisement, attitude toward
the brand, and purchase intention. The study examined
and discovered the interaction effects of endorser-product
congruence and endorser credibility on three specic con-
sumer responses. In particular, the endorser-credibility
effects on consumer responses are more signicant under
the high endorser-product-congruence condition than in
the low endorser-product-congruence condition. Analyses
of the study results further indicated that brand attitude
played a signicant mediating role in the association
between attitude toward the advertisement and purchase
intention. Finally, this study used real athletes and actual
brands (products) in the experiment. This effort is rarely
found in extant studies; in this case, external validity for
the current results is increased. The specic ndings of
the current study entail several important academic and
managerial implications and add further depth to our
knowledge of endorsement studies in general and sport
celebrity-endorsement studies in particular.
First, prior studies in the domain of advertising, mar-
keting, and sport management have supported the effects
of endorser-product congruence and endorser credibility
on consumer responses with traditional advertising
measures. The majority of these studies have primarily
focused on the main effect of endorser-product t or
endorser credibility without considering the interaction
effects of the two variables. The current study, in the
meantime, conrmed all hypotheses (1–3) indicating the
interaction effects between endorser-product congruence
and endorser credibility on attitude toward the advertise-
ment, attitude toward the brand, and purchase intention.
Thus, the current study suggests that consumer responses
are maximized when the two variables interact with each
other rather than when each predictor separately affects
the dependent variables.
This particular research finding complements
McCracken’s (1989) conceptual study in which the exis-
Table 10 Parameter Estimates for the Hierarchical Regression Model with Brand Attitude
Entered on Step 1 and Three Condition Indicators and their Products with
Brand Attitude Entered on Step 2
Step Variable
b SE βt p
1 Constant 0.28 0.35 — 0.80 .42
Ab0.76 0.08 0.61 9.09 .00
2 Constant 0.57 0.57 — 1.00 .32
Ab0.67 0.15 0.54 4.59 .00
G1 indicator –1.18 1.22 –0.34 –0.96 .34
G2 indicator 0.01 0.91 0.01 0.01 .99
G3 indicator 0.16 0.93 0.05 0.18 .86
Ab × G1 0.36 0.27 0.51 1.36 .18
Ab × G2 –0.09 0.22 –0.11 –0.41 .68
Ab × G3 –0.02 0.24 –0.02 –0.08 .94
Note. SE = standard error; dependent variable = purchase intention; Ab = attitude toward the brand.
Athlete-Product Congruence and Endorser Credibility 535
JSM Vol. 29, No. 5, 2015
tence of such interaction effect was argued. McCracken
pointed out that it would be difcult to explain the appeal
of a certain celebrity on the basis of the source models
alone because the endorser was persuasive only for certain
products. Although McCracken claimed that it was dif-
cult to assess how a certain celebrity’s credible image
interacted with different products because of the scope
of the research, the current study successfully provides
empirical evidence that the match-up effect strengthened
the endorser-credibility effect on consumer responses.
McCracken (1989) further argued that source models
prevented researchers from identifying the matches
and mismatches. This phenomenon may have occurred
because the celebrity endorser was, in fact, an actor. The
present study, on the other hand, used athlete celebrity
endorsers who t very well with sport-related products. In
comparison with athlete endorsers, it is difcult to nd a
general celebrity gure (e.g., actor, actress, or singer) who
is well matched with the endorsed products. This may
have been the main reason for the discrepancy in research
results between the two studies. Thus, it may be possible
to discover the interaction effects in a nonathlete celebrity
setting if a signicantly strong link between the image
of the celebrity and the endorsed product is conrmed.
In addition, the results of the interaction effects are
consistent with results from Kim and Na’s (2007) study
to a certain extent, which indicates that endorser cred-
ibility is more important when the t between the athlete
endorser and the endorsed product is congruent in terms
of attitude toward the product or brand. In their research,
attitude toward the product was the only dependent mea-
sure. The current research, however, not only ensured
the interaction effects between endorser credibility and
endorser-product congruence but also expanded the
dependent measures involved in the process by includ-
ing variables such as attitude toward the advertisement
and purchase intention. The ndings from this research
may help scholars and managers understand better that
an image of a credible celebrity is not easily transferred
to the endorsed brand image without the presence of
endorser-product congruence.
Second, the regression analysis result supported
the partial mediating effect of brand attitude between
attitude toward the advertisement and purchase intention
(Hypotheses 4–6). The path structure of the three tradi-
tional advertising measures can be found in various types
of marketing and advertising research (Goldsmith et al.,
2000; La Ferle & Choi, 2005; MacKenzie & Lutz, 1989).
Although Goldsmith et al. (2000) and La Ferle & Choi
(2005) suggest the indirect path from attitude toward the
advertisement to brand attitude, subsequently inuencing
purchase intention in terms of the endorser-credibility
effect, those studies differ from the present research
because they explain the endorser-credibility effects on
consumer responses without considering the interaction
effects. From this perspective, the current study lls the
gaps in the literature on celebrity-endorsement studies
and extends the body of knowledge to sport celebrity-
endorsement studies in the context of sport management.
Third, this study supplemented several limitations
inherent in methodological approaches of the prior
research. Unlike previous studies, which used ctitious
celebrities to verify the effectiveness of endorser-product
congruence (Koernig & Boyd, 2009; Till & Busler, 1998,
2000), the current study used real celebrities and actual
brands; thus enhancing external validity. Meanwhile,
previous research efforts have also focused on the differ-
ent endorser-credibility effects that occur when athlete
spokespersons rather than nonathlete celebrities endorse
sport-related products (Lafferty & Goldsmith, 1999;
Lafferty et al., 2002). The present research, however,
examined the effectiveness of endorser credibility and
endorser-product congruence by using athlete celebri-
ties endorsing both sport-related and non-sport-related
products. These efforts add additional meaning to the
current research outcome.
An important implication this study provides for
marketing management is that they need to consider
endorser-product congruence as well as endorser cred-
ibility. This may explain why some advertisers use chief
executive ofcers (CEOs) to endorse products, because
CEOs’ expert image is well-matched with their product’s
image and in turn induce customers to purchase the prod-
uct (Ohanian, 1991). Likewise, when a celebrity with a
credible image endorses a product, marketers expect the
positive image will be transferred to the endorsed product,
and thus create positive advertisement outcomes. This
idea induced managers to use celebrity athletes on certain
occasions to endorse products even when the brand image
is not congruent with the athlete’s image. This approach,
however, might not be as effective as it should be if they
use celebrity athlete endorsements relying only on the
athlete’s level of credibility or popularity. The ndings
from the current study recommend that managers addi-
tionally need to identify whether the athlete’s image ts
well with the image of the endorsed brand before making
a decision to use the athlete as an endorser.
Likewise, Stone, Joseph, and Jones (2003) note
that managers draw on famous athletes’ star power and
use celebrity endorsement to create an emotional tie
between the consumer and the athlete to enhance both
product awareness and purchase intention among target
consumers. This approach is thought to be effective to
sport consumers who have extremely high emotional
attachment toward the sport celebrity even if the celeb-
rity’s image does not t well with the endorsed product.
However, when congruence between the endorser and the
endorsed product is absent, the endorser-credibility effect
on general consumers who are not highly attached to the
sport celebrity or mainly persuaded by rational appeal
can be mitigated. These consumers in particular may
require strong image congruence between the product
and the endorser for the use of a credible endorser to take
full effect. Thus, the endorser-credibility effect can be
maximized in general by considering endorser-product
congruence in an advertisement campaign.
In addition, regression analysis results indicated that
attitude toward the advertisement is positively correlated
536 Lee & Koo
JSM Vol. 29, No. 5, 2015
with attitude toward the brand, and attitude toward the
brand is a predictor of purchase intention. These nd-
ings entail that promoting positive attitude toward the
advertisement will positively affect brand attitude, and
enhancement in brand attitude will increase consumers’
willingness to purchase the advertised product.
The major nding, however, is the partial mediat-
ing role of attitude toward the brand in the association
between attitude toward the advertisement and purchase
intention. This nding provides an idea of how consum-
ers’ attitudes are established leading up to purchase
behavior. In essence, consumers’ willingness to purchase
a product is signicantly inuenced by attitude toward
the advertisement via attitude toward the brand. Manag-
ers may apply this nding to their business strategy by
effectively promoting attitude toward the advertisement
via brand attitude to maximize consumers’ purchasing
efforts.
Limitations and Future Research
The present study is not without limitations. To begin
with, the participants were undergraduate students at a
university. Thus, the study results might not be applicable
to older adults in general. In addition, the Cronbach’s α
for the measures (except for attitude toward the ad) are
somewhat high as to be of concern. When Cronbach’s
α is high, item redundancy is recommended rather than
item complementarity. Future research should carefully
consider this issue and put effort to generate additional
parsimonious support.
Sponsorship could be an important issue in the
domain of match-up or endorser-credibility effects as
sport marketing and sponsorship have become more
popular. Indeed, many studies regarding event-sponsor
t stemmed from the match-up hypothesis in endorse-
ment. Previous studies found a path from event-sponsor
t to purchase intention mediated by sponsor attitude
(Gwinner & Bennett, 2008; Lee & Cho, 2009). In terms
of the credibility effect, Rifon et al. (2004) indicated that
sponsor credibility mediates the association between
sponsor-cause congruence and sponsor attitude. In addi-
tion, Chanavat, Martinent, and Ferrand (2010) showed the
causal relationship between image components in which
brand image dimensions of a sponsor inuenced those
of a sponsee. In this perspective, match-up effect and
endorser credibility relevant to sponsorship studies would
generate meaningful results. For instance, the interaction
effects between endorser credibility and endorser-product
congruence that were identied in the current study could
be extended to the interaction effects between sponsor
credibility and event-sponsor t in future researches.
Although this study offers further insights regard-
ing the effects of celebrity endorsement, it is difcult
to conclude that the results can be applied to other
advertisements in which general celebrities endorse a
product because only athlete endorsers were used in
the experiment. Future researchers are encouraged to
examine celebrity-product congruence by considering
issues beyond consumers’ responses to advertisements
and brands. For example, studies may examine consum-
ers’ perceptions of product quality and their ability to
recall advertisements. In addition, future research should
incorporate the use of celebrities from various areas as
well as different types of products. Such efforts would
promote the external validity of the results.
The current study used perceptually isolated
product categories (i.e., sport drinks and toothpastes)
in manipulating the degree of congruence. Although
other variables that had the potential to confound the
results were removed as much as possible, it is likely
that a certain amount of quirks within dissimilar product
categories create noise related to the perceived treat-
ment effects (e.g., price). Conducting an experiment
regarding endorser-product congruence with the same
product category is challenging. However, it is possible to
manipulate the levels of product involvement. According
to the previous studies, the levels of product involvement
play a moderating role in the impact of endorser-product
congruence (Bower & Landreth, 2001; Lee & Thorson,
2008; Till & Busler, 2000) and in the endorser-credibility
effect on consumer responses (Goldsmith et al., 2000;
Lafferty & Goldsmith, 1999, 2004; La Ferle & Choi,
2005). Thus, additional research including involvement
as a moderator may induce meaningful results.
Conclusion
The current research conrmed the interaction effects
between endorser credibility and endorser-product con-
gruence on consumers’ attitudes and purchase intention
both directly and indirectly. Advertisers cannot maximize
the effectiveness of celebrity advertising if they consider
only one of the two important factors. Furthermore,
endorser-product congruence plays a signicant role in
the association between endorser credibility and con-
sumer responses. Thus, endorser credibility should be
considered an important element to enhance consumers’
attitude toward the advertisement, attitude toward the
brand, and purchase intention. Finally, endorser-product
congruence should be carefully considered while devis-
ing an effective advertisement campaign to maximize the
effectiveness of endorser credibility.
References
Atkin, C., & Block, M. (1983). Effectiveness of celebrity
endorsers. Journal of Advertising Research, 23, 57–61.
Baron, R.M., & Kenny, D.A. (1986). The moderator–mediator
variable distinction in social psychological research: Con-
ceptual, strategic, and statistical considerations. Journal
of Personality and Social Psychology, 51, 1173–1182.
Bower, A.B., & Landreth, S. (2001). Is beauty best? Highly
versus normally attractive models in advertising. Journal
of Advertising, 30, 1–12. doi:10.1080/00913367.2001.1
0673627
Athlete-Product Congruence and Endorser Credibility 537
JSM Vol. 29, No. 5, 2015
Caballero, M.J., & Pride, W.M. (1984). Selected effects of sales-
person sex and attractiveness in direct mail advertisements.
Journal of Marketing, 48, 94–100. doi:10.2307/1251315
Chanavat, N., Martinent, G., & Ferrand, A. (2010). Brand
images causal relationships in a multiple sport event
sponsorship context: Developing brand value through
association with sponsees. European Sport Management
Quarterly, 10, 49–74. doi:10.1080/16184740903554090
Collins, A.M., & Loftus, E.F. (1975). A spreading-activation
theory of semantic processing. Psychological Review, 82,
407–428. doi:10.1037/0033-295X.82.6.407
Cunningham, G.B., Fink, J.S., & Kenix, L.J. (2008). Choosing
an endorser for a women’s sporting event: The interaction
of attractiveness and expertise. Sex Roles, 58, 371–378.
doi:10.1007/s11199-007-9340-z
Debevec, K., & Kernan, J.B. (1984). More evidence on the
effects of a presenter’s attractiveness: Some cognitive,
affective, and behavioral consequences. Advances in
Consumer Research, 11, 127–132.
Fink, J.S., Cunningham, G.B., & Kensicki, L.J. (2004). Using
athletes as endorsers to sell women’s sport: attractiveness
vs. expertise. Journal of Sport Management, 18, 350–367.
Friedman, H.H., & Friedman, L. (1979). Endorser effective-
ness by product type. Journal of Advertising Research,
19, 63–71.
Friedman, H.H., Santeramo, M.J., & Traina, A. (1978). Cor-
relates of trustworthiness for celebrities. Journal of the
Academy of Marketing Science, 6, 291–299. doi:10.1007/
BF02732313
Goldberg, M.E., & Hartwick, J. (1990). The effects of advertiser
reputation and extremity of advertising claim on advertis-
ing effectiveness. The Journal of Consumer Research, 17,
172–179. doi:10.1086/208547
Goldsmith, R.E., Lafferty, B.A., & Newell, S.J. (2000). The
impact of corporate credibility and celebrity credibility on
consumer reaction to advertisements and brands. Journal
of Advertising, 26, 43–54. doi:10.1080/00913367.2000.
10673616
Gwinner, K., & Bennett, G. (2008). The impact of brand cohe-
siveness and sport identication on brand t in a sponsor-
ship context. Journal of Sport Management, 22, 410–426.
Harmon, R.R., & Coney, K.A. (1982). The persuasive effects
of source credibility in buy and lease situations. Journal
of Marketing Research, 19, 255–260.
Hirschman, E.C. (1986). The effect of verbal and pictorial
advertising stimuli on aesthetic, utilitarian and familiarity
perceptions. Journal of Advertising, 15, 27–34. doi:10.10
80/00913367.1986.10673002
Hovland, C.I., Janis, I.L., & Kelley, H.H. (1953). Communi-
cation and persuasion. New Haven, CT: Yale University
Press.
Hovland, C.I., & Weiss, W. (1951). The inuence of source cred-
ibility on communication effectiveness. Public Opinion
Quarterly, 15, 635–650. doi:10.1086/266350
Joseph, W.B. (1982). The credibility of physically attractive
communicators: A review. Journal of Advertising, 11,
15–24. doi:10.1080/00913367.1982.10672807
Kahle, L.R., & Homer, P.M. (1985). Physical attractiveness
of the celebrity endorser: A social adaptation perspec-
tive. The Journal of Consumer Research, 11, 954–961.
doi:10.1086/209029
Kamins, M.A. (1990). An investigation into the “match-up”
hypothesis in celebrity advertising: When beauty may be
only skin deep. Journal of Advertising, 19, 4–13. doi:10.
1080/00913367.1990.10673175
Kim, Y.J., & Na, J.H. (2007). Effects of celebrity athlete
endorsement on attitude toward the product: The role of
credibility, attractiveness and the concept of congruence.
International Journal of Sports Marketing & Sponsorship,
8, 310–320.
Koernig, S.K., & Boyd, T.C. (2009). To catch a tiger or let him
go: The match-up effect and athlete endorsers for sport and
non-sport brands. Sport Marketing Quarterly, 18, 15–37.
La Ferle, C., & Choi, S.M. (2005). The importance of perceived
endorser credibility in South Korean advertising. Journal
of Current Issues and Research in Advertising, 27, 67–81.
doi:10.1080/10641734.2005.10505182
Lafferty, B.A., & Goldsmith, R.E. (1999). Corporate credibil-
ity’s role in consumers’ attitudes and purchase intentions
when a high versus a low credibility endorser is used
in the ad. Journal of Business Research, 44, 109–116.
doi:10.1016/S0148-2963(98)00002-2
Lafferty, B.A., & Goldsmith, R.E. (2004). How inuential are
corporate credibility and endorser attractiveness when
innovators react to advertisements for a new high-tech-
nology product? Corporate Reputation Review, 7, 24–36.
doi:10.1057/palgrave.crr.1540209
Lafferty, B.A., Goldsmith, R.E., & Newell, S.J. (2002). The dual
credibility model: The inuence of corporate and endorser
credibility on attitudes and purchase intentions. Journal of
Marketing Theory and Practice, 10, 1–12.
Lee, H.S., & Cho, C.H. (2009). The matching effect of brand
and sporting event personality: Sponsorship implications.
Journal of Sport Management, 23, 41–64.
Lee, J., & Thorson, E. (2008). The impact of celebrity-product
incongruence on the effectiveness of product endorse-
ment. Journal of Advertising Research, 48, 433–449.
doi:10.2501/S0021849908080446
MacKenzie, S.B., & Lutz, R.J. (1989). An empirical examina-
tion of the structural antecedents of attitude toward the ad
in an advertising pretesting context. Journal of Marketing,
53, 48–65. doi:10.2307/1251413
McCracken, G. (1989). Who is the celebrity endorser? Cultural
foundations of the endorsement process. The Journal of
Consumer Research, 16, 310–321. doi:10.1086/209217
McGuire, W.J. (1985). Attitudes and attitude change. In G.
Lindzey & E. Aronson (Eds.), Handbook of social psy-
chology (Vol. 2, 3rd ed., pp. 233–346). New York, NY:
Random House.
Misra, S., & Beatty, S.E. (1990). Celebrity spokesperson
and brand congruence: An assessment of recall and
affect. Journal of Business Research, 21, 159–173.
doi:10.1016/0148-2963(90)90050-N
Murphy, K.R., & Davidshofer, C.O. (2001). Validity of measure-
ment: Content and construct-oriented validation strategies.
538 Lee & Koo
JSM Vol. 29, No. 5, 2015
In L. Pearson (Ed.), Psychological testing: Principles and
applications (5th ed., pp. 153–177). Upper Saddle River,
NJ: Prentice Hall.
Nataraajan, R., & Chawla, S.K. (1997). “Fitness” marketing:
Celebrity or non-celebrity endorsement? Journal of Pro-
fessional Services Marketing, 15, 119–129. doi:10.1300/
J090v15n02_09
Ohanian, R. (1990). Construction and validation of a scale to
measure celebrity endorsers’ perceived expertise, trust-
worthiness, and attractiveness. Journal of Advertising, 19,
39–52. doi:10.1080/00913367.1990.10673191
Ohanian, R. (1991). The impact of celebrity spokespersons’
perceived image on consumers’ intention to purchase.
Journal of Advertising Research, 31, 46–54.
Petroshius, S.M., & Crocker, K.E. (1989). An empirical analy-
sis of spokesperson characteristics on advertisement and
product evaluations. Journal of the Academy of Marketing
Science, 17, 217–225.
Petty, R.E., & Cacioppo, J.T. (1981). Issue involvement as a
moderator of the effects on attitude of advertising content
and context. Advances in Consumer Research, 8, 20–24.
Petty, R.E., & Cacioppo, J.T. (1986). Communication and
persuasion: Central and peripheral routes to attitude
change. New York: Springer-Verlag. doi:10.1007/978-1-
4612-4964-1
Potthoff, R.F. (1966). Statistical aspects of the problem of
biases in psychological tests. (Institute of Statistics Mimeo
Series No. 479). Chapel Hill: University of North Carolina,
Department of Statistics.
Rifon, N.J., Choi, S.M., Trimble, C.S., & Li, H. (2004). Congru-
ence effects in sponsorship: The mediating role of sponsor
credibility and consumer attributions of sponsor motive.
Journal of Advertising, 33, 30–42. doi:10.1080/0091336
7.2004.10639151
Romani, S. (2006). Price misleading advertising: Effects on
trustworthiness toward the source of information and
willingness to buy. Journal of Product and Brand Man-
agement, 15, 130–138. doi:10.1108/10610420610658965
Shimp, T.A. (1997). Advertising, promotion, and supplemental
aspects of integrated marketing communications (4th ed.).
Fort Worth, TX: Dryden.
Siemens, J.C., Smith, S., Fisher, D., & Jensen, T.D. (2008).
Product expertise versus professional expertise: Congru-
ency between an endorser’s chosen profession and the
endorsed product. Journal of Targeting, Measurement
and Analysis for Marketing, 16, 159–168. doi:10.1057/
jt.2008.8
Smith, R.G. (1973). Source credibility context effects. Com-
munication Monographs, 40, 303–309.
Sobel, M.E. (1982). Asymptotic condence intervals for indirect
effects in structural equation models. Sociological Meth-
odology, 13, 290–312. doi:10.2307/270723
Stone, G., Joseph, M., & Jones, M. (2003). An exploratory study
on the use of sports celebrities in advertising: A content
analysis. Sport Marketing Quarterly, 12, 94–102.
Tellis, G.J. (2003). Effective advertising: Understanding when,
how, and why advertising works. London, UK: Sage
Publications.
Tingchi Liu, M., Huang, Y., & Minghua, J. (2007). Relations
among attractiveness of endorsers, match-up, and purchase
intention in sport marketing in china. Journal of Consumer
Marketing, 24, 358–365. doi:10.1108/07363760710822945
Till, B.D., & Busler, M. (1998). Matching products with endors-
ers: Attractiveness versus expertise. Journal of Consumer
Marketing, 15, 576–586. doi:10.1108/07363769810241445
Till, B.D., & Busler, M. (2000). The match-up hypothesis:
Physical attractiveness, expertise, and the role of t on
brand attitude, purchase intent and brand beliefs. Journal
of Advertising, 29, 1–13. doi:10.1080/00913367.2000.1
0673613
Weaver, B., & Wuensch, K.L. (2013). SPSS and SAS programs
for comparing Pearson correlations and OLS regression
coefcients. Behavior Research Methods, 45, 880–895.
doi:10.3758/s13428-012-0289-7
White, D.W., Goddard, L., & Wilbur, N. (2009). The effects
of negative information transference in the celebrity
endorsement relationship. International Journal of
Retail & Distribution Management, 37, 322–335.
doi:10.1108/09590550910948556
Wu, C., & Shaffer, D.R. (1987). Susceptibility to persuasive
appeals as a function of source credibility and prior
experience with the attitude object. Journal of Personality
and Social Psychology, 52, 677–688. doi:10.1037/0022-
3514.52.4.677
Yi, Y. (1990). Cognitive and affective priming effects of the
context for print advertisements. Journal of Advertising,
19, 40–48. doi:10.1080/00913367.1990.10673186
