Journal of Business Research 81 (2017) 96–106

Contents lists available at ScienceDirect

Journal of Business Research

journal homepage: www.elsevier.com/locate/jbusres

Consumer, brand, celebrity: Which congruency produces effective celebrity MARK

endorsements?
Noël Alberta,⁎, Laure Ambroiseb, Pierre Valette-Florencec,d
a
Kedge Business School, Domaine de Luminy BP 291, 13288 Marseille Cedex 9, France
b
COACTIS, Université de Lyon/Lyon 2, ISH, 14-16, avenue Berthelot, 69363 Lyon Cedex 07, France
c
Grenoble IAE, Université Grenoble Alpes, BP 47, 38040 Grenoble, France
d
CERGAM EA 4225, Aix-Marseille Université, France

A R T I C L E I N F O A B S T R A C T

Keywords: Celebrity endorsement research mainly focuses on celebrity–brand congruency, presented as a primary condition
Celebrity for the success of these marketing communications. In this view, the consumer functions as an outsider of the
Endorsement celebrity–brand dyad, even though endorsement efficiency depends on psychological processes by the consumer
Fit (i.e., identification, social influences). This study therefore proposes a model of celebrity endorsement to test the
Congruency
simultaneous effects of three congruencies among the brand, the celebrity, and the consumer. Data collected
Identification
from samples of real French women (mean age: 43.36 years) (N = 720) indicate that only the brand–consumer
Commitment
match exerts a significant impact on brand attitude, brand commitment, brand identification, and behavioral
intentions. Celebrity–brand and celebrity-consumer fits have a limited influence on the outcomes. These findings
challenge the exclusive prominence of celebrity–brand fit as a research and practical consideration and confirm
the consumer's active role in the endorsement.

1. Introduction Matthes (2017) detail the mixed impacts identified in celebrity en-
dorsement models on consumer perceptions, attitudes, and behavior.
Brands invest heavily to gain endorsements from celebrities. For These varying results might reflect a common limitation of classical
example, Nike signed a contract with Cristiano Ronaldo for £14.2 M per endorsement models: They tend to include consumers as passive spec-
year until 2020; LeBron James is the NBA's current endorsement king, tators of the celebrity-brand association and only measure those con-
with estimated off-court earnings of $48 million. However, despite this sumers' ultimate attitudes and behaviors toward the endorsed objects
infatuation with brands for celebrity endorsement, the association with (brands, products, services). In addition to celebrity-brand congruency,
a celebrity does not automatically benefit the brands (Zhou & Whitla, celebrity-consumer (Boon & Lomore, 2001; Choi & Rifon, 2012) and
2013), and a clear understanding of the mechanisms that underlie en- brand-consumer (Aaker, 1997; Fournier, 1998) congruencies could in-
dorsement effects remains an important goal for managers and re- fluence brand attitudes and relationships. Yet these three types of po-
searchers (Bergkvist & Zhou, 2016). tential fits or congruencies have never been investigated together in a
This research is interested in understanding endorsement effec- single empirical framework; such a standpoint could offer new con-
tiveness, which is defined here as an endorsement's positive influence tributions to theory and practice.
on consumer perceptions, attitudes, and behaviors toward the endorsed Such research also might provide answers to some pertinent ques-
brand. Previous research offers two main explanations for the influence tions. For example, are consumers just passive observers or do they
of an endorsement on consumers: celebrity characteristics, such as at- function as active, important contributors to endorsement outcomes?
tractiveness and credibility (Pornpitakpan, 2003; Till & Busler, 2000) Might the congruencies involving consumers be even more influential
and the proximity between the brand and the celebrity, as summarized than the one between the brand and the celebrity? To answer these
in the construct of congruency (Kamins & Gupta, 1994; Misra & Beatty, questions, this study proposes a global model with three “elements”
1990). These studies offer good insights, but existing endorsement associated with any endorsement (i.e., brand, celebrity, and consumer).
models cannot account for the entire system of influence associated Using a comparative approach, this study investigates the effects of the
with celebrity endorsements. In a recent meta-analysis, Knoll and congruencies across this triad on four outcomes of endorsement

⁎
Corresponding author.
E-mail addresses: [email protected] (N. Albert), [email protected] (L. Ambroise), pierre.valette-fl[email protected] (P. Valette-Florence).

http://dx.doi.org/10.1016/j.jbusres.2017.08.002

Available online 24 August 2017

Received 26 February 2015; Received in revised form 31 July 2017; Accepted 2 August 2017

0148-2963/ © 2017 Elsevier Inc. All rights reserved.

N. Albert et al. Journal of Business Research 81 (2017) 96–106

effectiveness: brand identification, attitude, commitment, and beha- association (Fleck, Korchia, & Le Roy, 2012). Such models of endorse-
vioral intentions. The results challenge prior literature, which relies on ment commonly focus on celebrity characteristics (attractiveness,
the celebrity-brand match-up as the primary explanation of the en- credibility) or brand proximity, leaving the consumer as a passive
dorsement's impact on consumer attitude. In turn, they reveal that spectator of the association. But studies in various contexts that simi-
brand-consumer congruency exerts the strongest impact on all out- larly include consumers, brands, and some third influential entity (e.g.,
comes. Extending Knoll and Matthes's (2017) findings, this study sug- another brand, a social cause, etc.) underline the notion that consumers
gests that researchers and managers must reorient their perceptions and always have an active role to play. For example, co-branding research
practices when it comes to endorsements; they should acknowledge that highlights three congruencies that inform attitudes toward brand alli-
celebrity endorsement involves three active entities and is not simply ances: between the brand and its partner, the brand and the consumer,
an alliance between a brand and a celebrity. and the consumer and the partner (Broderick, Jogi, & Garry, 2003). In
The next section details existing endorsement models and estab- cause-related marketing contexts, both cause-consumer and brand-
lishes a foundation for the argument that consumers may have a no- consumer congruencies influence the consumer's preference for a cause-
table impact on endorsement effectiveness, along with the brand and supporting brand (Goldsmith & Yimin, 2014).
the celebrity. After presenting the study methodology, this article de- However, endorsement literature does not offer any such triadic
tails the findings. The last section contains a discussion of the results, a framework. A few studies that include consumers largely base their
summary of the overall contributions, and some limitations and further arguments on balance theory (Mowen, 1980; Roy, Gammoh, & Koh,
research avenues. 2012). According to this reasoning, consumers prefer balance and
harmony in the elements that surround them (Mowen, 1980). Their
attitudes toward an endorsed brand thus depend not only on the ce-
2. Literature review lebrity-brand association but also on consumers' own connections with
the two partners. The lone empirical study that applies balance theory
Three criteria (also called endorsement models in the literature) de- to an endorsement context (Roy et al., 2012) is not sufficient to es-
fine a celebrity as an efficient endorser: attractiveness, credibility, and tablish the role and importance of the consumer in the endorsement
congruency with the brand (Erdogan, 1999). A more attractive celebrity effectiveness clearly. Furthermore, by focusing on harmony across the
delivers the message of the endorsed brand more efficiently three elements, balance theory cannot identify or compare the distinct
(Till & Busler, 2000). Celebrity attractiveness might refer to physical effects of each congruency (brand-consumer, celebrity-brand, or ce-
attributes or to other characteristics, such as intelligence or lifestyle lebrity-consumer) on attitudes toward the endorsed brand. Therefore,
(Erdogan, 1999), all of which can affect purchase intentions, brand this research aims to integrate, in a single model, the congruencies
recall, and attitudes toward the brand (Friedman & Friedman, 1979; across this triad of brand, celebrity, and consumer and then test and
Till & Busler, 2000). Celebrity credibility also appears central to the compare their effects on consumers.
endorsement's influence on consumers (Goldsmith, Lafferty, & Newell, Most previous studies focus on the influence of endorsements on
2000; Ohanian, 1991) and comprises various components (e.g., at- attitudes toward the brand or toward advertising and purchase inten-
tractiveness, dynamism, and objectivity). However, a general consensus tions. However, consumers also have more complex relationships with
focuses on expertise and trustworthiness dimensions (Goldsmith et al., brands, so attitudes may not be sufficient to predict their comprehen-
2000; Ohanian, 1991; Pornpitakpan, 2003). Trustworthiness reflects an sive behaviors (Fournier, 1998). Accordingly, this research determines
evaluation of the source's honesty and the objectivity of the information the measurement of endorsement effectiveness through four variables:
presented. Expertise corresponds to a consumer's assessment of the brand identification, brand attitude, behavioral intentions, and brand
celebrity's knowledge and the believability of his or her recommenda- commitment. Brand identification stems from information provided by
tions. Celebrity credibility then influences many indicators of en- the brand that applies to the consumer so that it reflects the degree of
dorsement effectiveness, including brand, product, and advertising at- overlap between their images (Tuškej, Golob, & Podnar, 2013). Brand
titudes, as well as purchase intentions (Friedman & Friedman, 1979; attitude implies a positive or negative disposition toward the brand
Pornpitakpan, 2003; Silvera & Austad, 2004). (Park, MacInnis, Priester, Eisingerich, & Iacobucci, 2010). Behavioral
However, some studies also note the limited impacts of attractive- intention refers to a consumer's willingness to perform difficult beha-
ness or credibility on consumer attitudes (Goldsmith et al., 2000) and viors that benefit the brand (e.g., promoting and defending it; Park
stress that these models are not universally applicable to all products et al., 2010). Finally, brand commitment is a consumer's willingness to
(Till & Busler, 2000). A more integrative approach, designed to over- maintain a relationship with the brand (Fournier, 1998).
come these limitations, emerges in the form of congruency models, also
referred to as fit, match-up, coherence, or similarity models. This ap- 3. Hypotheses
proach acknowledges that endorsements usually pursue harmony be-
tween the brand and the celebrity endorser (Kamins & Gupta, 1994). 3.1. Brand-consumer congruency hypotheses
That is, managers generally seek to ensure that the brand is consistent
with the image portrayed by the endorser, which can reduce perceived Brands have personal and social meaning (McCracken, 1989), which
risk (Erdogan, 1999) and also prompt positive perceptions by con- people use to create, enhance, or communicate their own identity (Belk,
sumers. In particular, a congruent endorsement should be more effi- 1988). Congruent images between the brand and consumers should
cient (match-up hypothesis), whereas the absence of congruency may lead to stronger brand identification. Brand-consumer congruency re-
imply that the celebrity has mainly financial motives for endorsing the fers to the perceived similarity between the consumer's self-image and
brand, which could produce negative evaluations (Kamins & Gupta, the brand's image (Tuškej et al., 2013). Brand identification is the de-
1994). gree to which the brand expresses the consumer's identity (Tuškej et al.,
Most research captures congruency in one of three forms: holistic 2013) and highlights the brand's integration into the consumer's self
appraisals (Kamins & Gupta, 1994), scales with attractiveness or ex- (Escalas & Bettman, 2003). The more the brand image will be perceived
pertise items (Till & Busler, 2000),1 or scales with items designed to as similar to the consumer's, the more he or she will identify with the
measure the relevance and expectancy of the celebrity-brand brand (Tuškej et al., 2013). Moreover, similar to other symbols, a brand
also earns greater appreciation when it strengthens users' self-concepts
1
These scales reflect, for example, the notion that congruency exists because an at-
(Belk, 1988). Thus a brand that is congruent with the consumer's own
tractive celebrity should be a more effective endorser for products that promise to en- self-image tends to generate more positive attitudes (Belk, 1988).
hance consumers' own attractiveness. According to self-consistency theory, consumers behave in ways

97
N. Albert et al. Journal of Business Research 81 (2017) 96–106

consistent with how they see themselves so that they feel motivated to 2012) so that responses to endorsements may be more positive when
buy and be loyal to brands that reinforce their self-perceptions (Sirgy, the consumer perceives that the celebrity image is congruent with the
Lee, Johar, & Tidwell, 2008; Tuškej et al., 2013). For example, Grubb image he or she would like to have. With a similar rationale, through
and Grathwohl (1967, p. 23) state that consumers' buying behavior is identification and internalization processes, celebrity-consumer con-
determined by “the interaction between the personality of the buyer gruency should affect consumers' relationships with the brand, enhan-
and the image of the bought product.” Thus, consumers are more likely cing their behavioral intentions and their commitment to the brand
to purchase repeatedly and to pursue relationships with brands that (Kelman, 1961). Thus, we offer the following hypothesis:
express or enhance their identities (Escalas & Bettman, 2003; Fournier,
H2. Celebrity-consumer congruency positively affects a consumer's (a)
1998; Tuškej et al., 2013). In turn, a brand-consumer match may con-
brand identification, (b) brand attitude, (c) behavioral intentions, and
tribute to brand behavior intentions and brand commitment. Therefore,
(d) brand commitment.
we present the following hypothesis:
H1. Brand-consumer congruency positively affects a consumer's (a)
brand identification, (b) brand attitude, (c) behavioral intentions, and 3.3. Celebrity-brand congruency hypotheses
(d) brand commitment.
Celebrity-brand congruency suggests that two partners share similar
images. According to cognitive consistency theory, people react more
3.2. Celebrity-consumer congruency hypotheses positively to stimuli that are congruent with their cognitive states. A
consumer who perceives the clear dedication of the celebrity to the
Celebrities function as reference groups for people so that some brand should develop favorable attitudes toward the endorsed brand.
consumers try to look like, search for information about, and make Moreover, according to transfer theory, the celebrity's image, status,
inferences about celebrities (Choi & Rifon, 2012). In turn, people may and attractiveness transfer to the brand through the endorsement
develop an attraction to a celebrity and invest time and energy to (McCracken, 1989). Perceptions and emotions generated by the celeb-
achieve a perceived relationship with him or her (Boon & Lomore, rities bind to the endorsed brands through recurring associations
2001). People also use celebrities to enhance or support their self- (Choi & Rifon, 2012), leading consumers to develop more positive at-
identities, because celebrities convey symbols through their media titudes and behaviors toward the endorsed brand (Erdogan, 1999;
presence and lifestyles (Boon & Lomore, 2001; Choi & Rifon, 2012). Misra & Beatty, 1990). Celebrity-brand congruency expressly facilitates
Such forms of involvement even can develop into a sense of intimacy the creation of an associative link that enhances the transfer of attri-
that increases people's perceptions of their congruency with the celeb- butes from the celebrity to the brand (Seno & Lukas, 2007). To be ef-
rity (Boon & Lomore, 2001). fective, the congruency must reflect the distinctive characteristics of the
In an endorsement context, celebrities express cultural meaning for celebrity and the salient attributes of the brand (Misra & Beatty, 1990).
consumers (e.g., lifestyle, personality), so consumers can construct or The more noticeable this congruency, the more the consumer accepts
enhance their identities by using celebrities who represent valued the celebrity's influence (Kamins & Gupta, 1994), which should lead to
characters (Thomson, 2006). Social influence theory (Kelman, 1961) identification with and positive attitudes toward the endorsed brand.
predicts that two processes—identification and inter- Furthermore, a consumer who adopts the celebrity's attitudes likely
nalization—explain how celebrities influence consumers and their be- buys and expresses similar commitment to an endorsed brand; there-
haviors. In an identification process, a person imitates the attitudes or fore, we put forth the following hypothesis:
behavior of another person in a specific context, mainly to look like that H3. Celebrity-brand congruency positively affects a consumer's (a)
other. Internalization instead implies that the focal person adopts the brand identification, (b) brand attitude, (c) behavioral intentions, and
postures of someone else, because those postures correspond to his or (d) brand commitment.
her value system. In an endorsement context, a consumer thus adopts a
celebrity's attitudes and behaviors because of a sense of similarity and Fig. 1 summarizes the overall research model, including the three
value closeness with that celebrity (Friedman & Friedman, 1979; congruencies and their effects on the four key outcomes of endorsement
Thomson, 2006). effectiveness.
In turn, the value of celebrities as endorsers is based on their pro-
pensity to activate social references, especially if the purchase of brands 4. Method
they endorse enables consumers to shape or modify their own self-
images (Erdogan, 1999). Thus celebrity-consumer congruency may af- To test the research model in Fig. 1 and its associated hypotheses,
fect consumers' attitudes toward an endorsed brand (Choi & Rifon, the authors conducted a two-phase study with 720 consumers. As a

Fig. 1. Research model.

98
N. Albert et al. Journal of Business Research 81 (2017) 96–106

measure of congruency, this study relies on the personality construct credibility as a brand endorser. A month later, the same participants
that can signal the fit among the three parties in an endorsement triad. reviewed an advertisement for the brand endorsed by the celebrity. In
Furthermore, to enhance validity, the research design features (1) two this second phase (N2 = 720), they evaluated the endorsed brand's
brands and four celebrities, which improves ecological validity; (2) a personality as well as their own personality. The items used to measure
randomly selected sample of actual consumers, obtained in collabora- identification, attitude, commitment, and behavioral intentions toward
tion with a dedicated, expert market research company, which im- the endorsed brand also appeared in this second phase. The final panel
proves external validity; and (3) data collected in two steps, which help of 720 complete observations across both phases served as the database
avoid halo effects of the three repeated measures of the personality for the analyses. All respondents were French women, with a mean age
construct (i.e., for the consumer, brand, and celebrity). Previous re- of 43.36 years. Furthermore, 15% earned monthly incomes lower than
search indicates gender influences on celebrity attractiveness assess- 1200€; 35% had monthly incomes between 1200€ and 2000€; 42%
ments and consumer responses to endorsements (Tzoumaka, between 2000€ and 4000€; and 8% > 4000€.
Tsiotsou, & Siomkos, 2016), so this study includes only female celeb-
rities and female respondents to limit the potential exogeneity of the 4.4. Measures
results.
Classical measures of congruency from endorsement literature (e.g.,
4.1. Product category and brand selections those based on attractiveness or credibility) seek to capture the har-
mony between the brand and the celebrity. In this sense, they may be
This study features product categories for which celebrity endorse- insufficient for capturing the psychological mechanisms associated with
ments are common (underwear and champagne) to reflect a realistic an endorsement, such as the consumer's identity formation due to a
situation. Dim (underwear) and Moët & Chandon (champagne) are well- sense of closeness with the brand and the celebrity. To capture con-
known in France (country of the data collection), and they also are gruency for this study, the required measure instead must be based on a
among the most consumed brands in their categories (according to construct that (1) can apply to all three parties (consumer, brand, ce-
SIMM2 2010, from TNS Media Intelligence ranking). All study partici- lebrity), (2) reflects their closeness, and (3) exerts some influence on
pants knew both brands, and 88.5% (63.3%) had bought Dim behavior. Thus, the personality construct has been selected to measure
(Moët & Chandon). the congruency among the three partners because it meets all these
requirements. Indeed, the brand literature has assumed for a long time
that consumers think of brands as if they were persons or celebrities
4.2. Celebrity selection
(Aaker, 1997; Fournier, 1998) and use personality adjective traits to
describe them (Berry, 1988). Moreover, brand personality scales are
The celebrity selection involved two steps. First, in a focus group
based on human personality inventories, and each original human trait
(N = 12), participants considered potential endorsers and proposed
has been evaluated to verify and confirm its applicability to brand
separate lists of celebrities for each brand. That is, the focus groups
(Aaker, 1997). Human personality inventories generally encompass
produced two extended lists of celebrity endorsers (one list per brand).
hundreds of traits; however, Aaker (1997) retained only 42 items
The focus group participants next reviewed each extended list and came
(structured in 15 facets and 5 dimensions) relevant in a branding
to a collective agreement about four celebrities per brand who seemed
context. Based on all these arguments, this research uses the Aaker
most relevant as endorsers. Second, a quantitative online survey
(1997) brand personality scale to measure the personality of the three
(N = 166 for Dim, N = 182 for Moët & Chandon) included these four
studied figures: the brand, the celebrity, and the consumer. Moreover,
endorsers for each brand, then asked respondents to indicate the most
extended research on human and brand personality proves the capacity
relevant endorser. The two most-cited celebrities were retained for each
of using such a construct to differentiate and discriminate people and
brand: Monica Belluci and Penelope Cruz emerged as most relevant to
brands, as well as to understand and predict human behavior
endorse Moët & Chandon, and Marion Cotillard and Eva Longoria
(Pervin & John, 1999)—specifically consumer behavior. Recent studies
emerged for Dim. To ensure that these selected celebrities possess the
even indicate that the personality concept is central to celebrity en-
qualities required to be good endorsers, the data collection for the main
dorsements and meaning transfer (Arsena, Silvera, & Pandelaere, 2014;
study tested their credibility (i.e., attractiveness, expertise, and relia-
Zamudio, 2016) and is highly relevant for addressing brand fit effi-
bility using semantic differential items from Ohanian, 1991). The
ciency (Van der Lans, Van den Bergh, & Dieleman, 2014). The person-
average scores confirm the high levels of credibility for each selected
ality of each element as it relates to an endorsement (consumer, brand,
celebrity (Appendix A).
celebrity) thus underlies the congruency measures for this study. With a
Using real advertisements from Dim and Moët & Chandon, the au-
view to parsimony, this study uses the 15 most salient adjective traits
thors added a picture of each selected celebrity (Appendix B). All the
from Aaker's (1997) scale (see Appendix C.1), in line with prior re-
created pairings were fictitious, but the use of real brands and celeb-
search (Batra, Lenk, & Wedel, 2010; Branaghan & Hildebrand, 2011;
rities in realistic ads increased the ecological validity of the study.
Eisend & Stokburger-Sauer, 2013; Kressmann et al., 2006).
The dependent variable scales measuring endorsement effectiveness
4.3. Participants and procedure all came from prior literature and were empirically validated. The
brand identification measure included four items from Escalas and
As suggested by Winer (1999), the data collection was conducted by Bettman (2003); the three semantic differential items for brand attitude
a dedicated market research company that has the means to issue were those proposed by Park et al. (2010); brand commitment relied on
random invitations to respondents to participate in the study, ensuring five items from Fullerton (2005); and the six items related to behavioral
better external validity. The selected respondents were randomly as- intentions came from Park et al. (2010) (see Appendix D).
signed to one advertisement each. Then to gauge their evaluations of
the personalities of the three entities, a two-step data collection process 5. Results
followed (see Fig. 2), which helped avoid halo effects. In the first phase
(N1 = 1304), participants rated the celebrity's personality and To test the research model and associated hypotheses, the estimates
of the three personality congruency factors functioned as the in-
2
SIMM - System of Information Media Market- is an annual report prepared by TNS
dependent variables in a multi-regression analysis. The following sec-
Sofres about a survey conducted on French consumers' behaviors and medias consump- tions present the congruency estimation procedures and detail the in-
tion. fluence of the three congruencies on different brand outputs.

99
N. Albert et al. Journal of Business Research 81 (2017) 96–106

Fig. 2. Research design.

Fig. 3. Brand-consumer congruency factor model.

5.1. Congruency estimation and measurement validity The normalized component scores then provided the formal latent
variables.
In line with prior research on the personality (Aaker, 1997), the Next, to calculate the congruency between each pair (brand, ce-
modeling of the three personality congruencies used a principal com- lebrity, consumer), another set of PCAs was applied to each personality
ponent analysis approach, with first-, second-, and third-order latent dimension and each dyad, using the saved normalized factors. This step
variables (first level for personality dimensions, second level for con- produced a congruency factor for each personality dimension for every
gruency per personality dimension, third level for global personality- dyad (brand-consumer, celebrity-consumer, celebrity-brand). Finally,
based congruency estimation; see Fig. 3). All analyses used XLSTAT the estimation of the global personality-based congruency factor for
(Version 19.03). A first set of principal component analyses (PCA) each dyad3 used the five personality dimension congruencies. A unique
served to compute the five personality dimensions suggested by Aaker
(1997) (excitement, sophistication, sincerity, competence, and rug-
gedness) for each element of the triad (brand, celebrity, and consumer). 3
Because the factor scores are centered, the closer the score is to 0, the weaker the
Appendix C.2 details the properties of these personality dimensions. congruency; inversely, the greater the absolute value is, the higher the congruency.

100
N. Albert et al. Journal of Business Research 81 (2017) 96–106

Table 1
Reliability, validity, and correlations of the dimensions of personality-based congruency.

Loadings Brand-consumer congruency Celebrity-brand congruency Celebrity-consumer congruency

Sincerity congruency 0.90 0.92 0.90

Excitement congruency 0.88 0.89 0.89
Competence congruency 0.91 0.93 0.92
Sophistication congruency 0.81 0.80 0.82
Ruggedness congruency 0.79 0.82 0.79

Global personality-based congruency Number of items α CR 1 2 3

1 Brand-consumer 5 0.91 0.93 (0.74)

2 Celebrity-brand 5 0.92 0.94 0.69a (0.76)
3 Celebrity-consumer 5 0.92 0.94 0.69 0.70 (0.75)

Notes: α = Cronbach's alpha, CR = composite reliability. All correlations are significant at p = 0.01. The average variances extracted (AVE) are along the diagonal.
a
All squared correlations are below the AVE (discriminant validity check).

factor emerged naturally in each case (i.e., one eigenvalue > 1, ex- Table 2
plaining more than 50% of the overall variance). Dependent variables, reliability, validity, and correlations.
Validity and reliability analyses checked each construct in the re-
Constructs Number of α CR 1 2 3 4
search model. The Cronbach's alpha and composite reliability (CR) items
values served to assess reliability, the average extracted variance (AVE)
provided the assessment of convergent validity, and Fornell and 1 Brand 4 0.94 0.96 (0.85)
Larcker's (1981) procedure was applied to estimate discriminant va- identification
2 Brand attitude 3 0.95 0.97 0.24a (0.91)
lidity (Henseler, Ringle, & Sinkovics, 2009). With regard to the three 3 Brand 5 0.93 0.95 0.80 0.32 (0.78)
personality-based congruencies,4 the indicators in Table 1 all reveal commitment
satisfactory convergent validity and reliability, with high Cronbach's 4 Brand behavioral 6 0.92 0.94 0.59 0.44 0.70 (0.71)
alpha and CR scores (above 0.91). The AVE for each congruency also is intentions
superior to its shared variance with any other congruencies (squared
Notes: α = Cronbach's alpha, CR = composite reliability. All correlations are significant
correlations) in support of discriminant validity. In addition, to confirm
at p = 0.01. The average variances extracted are along the diagonal.
this last point, the computation of the heterotrait-monotrait (HTMT) a
All squared correlations are below the AVE (discriminant validity check).
matrix confirms that each personality-based endorsement congruency
differs from the others, all computed values being below the 0.85 commitment β = 0.47, p < 0.00; behavioral intentions β = 0.65,
threshold (Appendix E). Finally, to assess the level of collinearity p < 0.00). The celebrity-consumer match-up has no significant influ-
among the three congruencies, the variance inflation factor (VIF) was ence on identification (p < 0.23) but slightly affects brand attitude
considered, as suggested by Hair and Hult (2016, p. 143). Because all (p < 0.10), commitment (p < 0.29), and behavioral intentions
VIF values are below the critical value of 5 (brand-consumer = 4.06, (p < 0.19), so these findings lead to the rejection of H2a and mar-
celebrity-brand = 4.18, celebrity-consumer = 4.12), no issue of mul- ginally support H2b–d. Finally, celebrity-brand congruency enhances
ticollinearity can be assumed among these variables. the attitude toward the brand (β = 0.26, p < 0.00) and brand com-
Among the dependent variables, each construct reveals satisfactory mitment (β = 0.14, p < 0.03), but it has no significant effect on brand
reliability scores (above 0.92). Moreover, all of the AVE levels are identification (p < 0.34) or behavioral intentions (p < 0.99). These
higher than the shared variance between constructs, providing em- results validate H3b and c but not H3a and d.
pirical support for discriminant validity, with the exception of identi-
fication and commitment (see Table 2). To further those results, we
calculated the HTMT ratio, which confirms the discriminant validity in 5.3. Test of a global personality-based endorsement congruency
every dependent variable (ratios from 0.51 to 0.80). Consequently,
because academics assert that identification and commitment differ As previously asserted, the primary objective of this study is to
theoretically (Fournier, 1998; Tuškej et al., 2013), the current study compare the influence of the three existing congruencies on endorse-
retains both constructs as dependent variables. ment effectiveness; however, the data collected also enable testing the
impact of a global triadic congruency that includes the brand, the ce-
lebrity, and the consumer all together to determine if this “global
5.2. Findings
congruency” exerts a stronger effect than each congruency dyad sepa-
rately. This new congruency, called the global personality-based en-
To test the hypotheses and assess the impact of the three con-
dorsement congruency (GPEC), is defined as the global assessment of
gruencies on brand relationship outcomes, the three global personality-
proximity in terms of personality existing among the three elements of
based congruency factors entered the multiple regressions as ex-
an endorsement. Conceptual and statistical arguments are in favor of
planatory variables. The results from a pooled database of 720 ob-
considering the three congruencies as defining characteristics of the
servations highlight the strong influence of the brand-consumer con-
GPEC and not as its manifestations (Jarvis, MacKenzie, & Podsakoff,
gruency but weak or no impact of the other two matches (Table 3).
2003). From a theoretical point of view, studies on the endorsement
Specifically, the brand-consumer match-up significantly and positively
using balance theory distinguish the nature of congruencies involving
affects all four outcomes, in support of H1a–d (respectively, brand
the consumer (i.e., brand-consumer and celebrity-consumer) from the
identification β = 0.36, p < 0.00; brand attitude β = 0.47, p < 0.00;
match-up in which he or she is not present (i.e., celebrity-brand)
(Mowen, 1980; Roy et al., 2012). The congruencies involving the
4
Appendix C2 contains detailed confirmation of the reliability and validity of each consumer are labeled sentiment relation, and a strong affective compo-
personality dimension. nent participates in the appreciation of the proximity between the

101
N. Albert et al. Journal of Business Research 81 (2017) 96–106

Table 3
Explanatory power of the three congruencies.

Dependent variables Un-normalized coefficients Normalized coefficients t p Value Adjusted R2 F p Value

B SE β

Brand identification
(Constant) −1.66 0.30 0.23 74.42 0.00
Brand-consumer congruency 0.69 0.13 0.36 5.53 0.00
Celebrity-brand congruency 0.12 0.12 0.06 0.96 0.34
Celebrity-consumer congruency 0.16 0.13 0.08 1.21 0.23

Brand attitude
(Constant) 0.89 0.20 0.38 147.58 0.00
Brand-consumer congruency 0.68 0.09 0.47 7.88 0.00
Celebrity-brand congruency 0.38 0.09 0.26 4.41 0.00
Celebrity-consumer congruency −0.15 0.09 − 0.10 − 1.65 0.10

Brand commitment
(Constant) −1.63 0.28 0.28 96.51 0.00
Brand-consumer congruency 0.89 0.12 0.47 7.43 0.00
Celebrity-brand congruency 0.26 0.12 0.14 2.19 0.03
Celebrity-consumer congruency −0.14 0.13 − 0.07 − 1.06 0.29

Brand behavioral intentions

(Constant) −1.53 0.28 0.34 123.83 0.00
Brand-consumer congruency 1.25 0.12 0.65 10.67 0.00
Celebrity-brand congruency 0.00 0.12 0.00 − 0.01 0.99
Celebrity-consumer congruency −0.16 0.12 − 0.08 − 1.31 0.19

Notes: Multiple regression results. The underlined betas are significant at p ≤ 0.01.

consumer and his or her partner. On the contrary, when the consumer is the GPEC on the four outcomes compared to the model integrating the
not directly involved, he or she assesses the proximity between the three congruencies (close R-square). Those results confirm again that
celebrity and the brand only through a cognitive and more global as- the endorsement effects do not come only from the association between
sessment of the similarities between the two (Mowen, 1980). From a the celebrity and the brand. The first model tested in the research has
statistical point of view, and in line with Jarvis et al. (2003), several confirmed the active participation of the consumer in the endorsement
indicators support the construction of the GPEC based on the three process through testing and comparing the influence of each of the
congruencies: The VIF values establish that there is no multicollinearity three pairings of congruencies. This second model also confirms the
among the three congruencies; the HTMT ratio of correlations asserts importance of the consumer and goes beyond in this demonstration by
the distinctiveness of the three personality-based congruencies; finally, highlighting that the congruency can be global, that is to say, include
a substantial part of the variance of these constructs is not shared. Thus all the endorsement's participants.
the GPEC was modeled as a formative variable and predicted scores
were used as explanatory variables in subsequent regressions. The three 6. Discussion, limitations, and further research
congruencies positively participate in the formation of the GPEC. All
path coefficients are significant and positive: celebrity-consumer con- Endorsement literature explains the impact of this practice on
gruency (0.354, p > 0.00), celebrity-brand congruency (0.358, consumer attitudes mainly by investigating the celebrity-brand link,
p > 0.00), and brand-consumer congruency (0.348, p > 0.00). relegating the consumer to a passive role. In contrast to this view, the
The results show that the GPEC positively influences brand identi- current research proposes that a consumer's proximity to the celebrity
fication (β = 0.48, p < 0.00), attitude (β = 0.60, p < 0.00), com- and the brand must be taken into account to understand the develop-
mitment (β = 0.51, p < 0.00), and behavioral intentions (β = 0.54, ment of attitudes toward an endorsed brand. The proposed, single, in-
p < 0.00) (Table 4). These results demonstrate a similar influence of tegrative model thus includes three likely congruencies among the

Table 4
Explanatory power of the global personality-based endorsement congruency (GPEC).

Dependent variables Un-normalized coefficients Normalized coefficients t p Value Adjusted R2 F p Value

B SE β

Brand identification
(Constant) − 1.69 0.30 0.23 212.02 0.00
GPEC 0.98 0.07 0.48 14.56 0.00

Brand attitude
(Constant) 0.76 0.21 0.35 396.16 0.00
GPEC 0.93 0.05 0.60 19.90 0.00

Brand commitment
(Constant) − 1.75 0.29 0.26 256.60 0.00
GPEC 1.04 0.07 0.51 16.02 0.00

Brand behavioral intentions

(Constant) − 1.64 0.29 0.29 288.95 0.00
GPEC 1.10 0.06 0.54 17.00 0.00

Notes: Simple regression results. The underlined betas are significant at p ≤ 0.01.

102
N. Albert et al. Journal of Business Research 81 (2017) 96–106

consumer, the brand, and the celebrity. The findings bring to light Fourth, from a methodological and theoretical perspective, this
notable differences in the impacts of the three tested congruencies. The study captures congruency by using the concept of personality, which is
celebrity-brand congruency significantly influences two outcomes of a novel approach to the proximity among partners in an endorsement
endorsement effectiveness, brand attitude, and commitment. context and thus differs somewhat from traditional measures. For
Surprisingly, it has no influence on brand identification, even though, consumers, an endorsement induces psychological mechanisms related
theoretically, it seemingly should underlie the influence of the brand to identity construction (McCracken, 1989), so from a theoretical
and the celebrity on consumer behavior, especially in an endorsement standpoint, personality offers an interesting means to capture proximity
context. For congruencies including the consumer, the results confirm across the three partners, in line with psychology literature, which in-
his or her active role in the endorsement: Brand-consumer congruency cludes personality as a frequently used concept for social comparison
significantly affects all four tested outcomes and the celebrity-consumer and identity construction (Pervin & John, 1999). The model's predictive
congruency has a moderate impact on three of the four outcomes power for brand attitudes, commitment, identification, and behavioral
(brand attitude, brand commitment, and behavioral intentions). intentions confirms the relevance of personality to the notion of con-
Through a comparison of their effects, this study offers a renewed and gruency. Therefore, academics might renew the criteria they use to
extended vision of endorsement effects and thereby establishes four assess congruency and include wider-ranging concepts and measures
main contributions to endorsement literature. that capture the complexity that arises in consumers' relationships with
First, this research challenges the hegemony of celebrity-brand celebrities and brands.
congruency as the sole explanation for endorsement effects on con- The varying results also produce some recommendations for man-
sumer attitudes and behaviors. It can enhance brand attitude and agers and practitioners. Similar to researchers, managers should di-
commitment, but it had no significant effects on brand identification or versify the criteria they use to assess and select celebrities; attractive-
behavioral intentions. For years, academics have considered this link ness and credibility may not be sufficient to capture psychological
fundamental to the effects on consumers (e.g., brand and advertising proximity. Personality can overcome this limitation, and managers al-
attitudes, brand affect, purchase intentions; Till & Busler, 2000). Yet the ready have operational tools to measure this construct (Aaker, 1997).
results of the current study align more with a recent meta-analysis that Personality notably offers managers an opportunity to focus on one or a
indicates a limited impact on consumers (Knoll & Matthes, 2017). This few dimensions—that is, those they consider most relevant for mea-
is not to suggest that celebrity-brand congruency is not essential; it suring the proximity of their brand to the consumer or the celebrity. For
facilitates the transfer of symbols from a celebrity to the brand example, a fragrance brand might be interested in measuring its
(McCracken, 1989) and makes the association credible proximity to consumers and celebrities on a sophistication dimension,
(Kamins & Gupta, 1994). But even as it contributes to the endorsement which is relevant in this product category, but a high-technology brand
effectiveness, through its effects on brand attitude and commitment, might prefer investigating the competence dimension.
celebrity-brand congruency may be not sufficient to exert substantial Another managerial recommendation comes from the mixed results
impacts on consumers' behavioral intentions. of the celebrity-brand congruency, which should encourage managers
Second, the results confirm the consumers' participation in en- to question their motivations in linking their brand to a celebrity. If
dorsements, which however vary with the partner. That is, brand- their primary objective is to influence consumer behaviors, they should
consumer congruency affects all four dependent variables, demon- prioritize their effort to increase the proximity between the brand and
strating the active role of the consumer through his or her proximity to the consumer, as this research reveals.
the endorsed brand. Because the endorsement causes symbols, images, Along with these contributions, this study has several limitations.
and meanings to transfer from the celebrity to the brand, an endorsed First, the sample characteristics may limit the generalizability of the
brand represents a material and accessible figure, full of meaning results, because the respondents all were women. Second, the use of
(McCracken, 1989). Congruency between an endorsed brand and the only female celebrities may have limited the influence of the tested
consumer enables the consumer to appropriate symbols and meanings congruencies. In their meta-analysis, Knoll and Matthes (2017) high-
in support of her or his identity construction (McCracken, 1989), which light that female celebrities generate weaker effects than male celeb-
likely explains the strong impact observed on attitudes and behaviors rities, because consumers tend to attribute less prestige and credibility
toward the brand. At the same time, though, celebrity-consumer con- to women. Third, all the celebrities in this research had famously en-
gruency has a moderate impact on brand attitude, brand commitment, dorsed other brands prior to the focal study. Marion Cotillard and
and behavioral intentions (but no influence on brand identification). Monica Belluci were the muses for, respectively, Dior and Dolce &
Therefore, this research confirms the active participation of the con- Gabana; Penelope Cruz has endorsed Mango and Schweppes; and Eva
sumer in the endorsement not only with the brand but also through his Longoria appeared in many L'Oréal marketing communications. Ac-
or her proximity with the celebrity. However, although some research cording to endorsement literature, a celebrity's effects on consumer
indicates that a consumer's proximity to a celebrity leads to positive attitudes are more important when that person endorses only one brand
affect toward the endorsed brand (Choi & Rifon, 2012; Fleck et al., or product (Mowen & Brown, 1981). Multiple endorsements for well-
2012), this study indicates that celebrity-consumer congruency is not known brands, which sometimes vary widely and differ from the brands
sufficient to enhance substantially consumer attitudes and behaviors used in this research, could have reduced the impact of congruencies
toward the endorsed brand. involving these celebrities in the current study. Fourth, similar to most
Third, congruency generally has been defined as a pairing of two endorsement studies, participants considered the advertisement that
entities; this research offers the first test of a global congruency that linked the brand to the celebrity only once, whereas in reality, attitudes
encompasses the brand, the celebrity, and the consumer. Indeed, the toward an endorsed brand result from a learning process, made possible
classical approach supports a comparison of the distinct effects of the through multiple exposures to marketing communications
three congruency dyads involved in an endorsement, but recognition of (Knoll & Matthes, 2017).
the consumer's active role suggests a new approach to congruency in These limitations, as well as this study's contributions, encourage
endorsement settings. The GPEC assessment provides an integrative further research on these topics. First, additional studies should confirm
approach that can capture the congruencies among the three partners' the results obtained in this study regarding the three congruencies. To
personalities with a single factor. Its positive impact on the four tested that purpose, researchers might replicate the proposed research model
outcomes affirms its empirical relevance. Furthermore, this approach to while varying certain conditions, such as the number of exposures to
congruency is not necessarily limited to an endorsement setting; it may the advertisement, celebrities' genders, or their skill domain (e.g., mo-
echo other research domains that rely on interactions among various vies, sports). Second, two of the congruencies tested in this research
entities, such as sponsorships, cause-related marketing, or co-branding. (celebrity-brand and celebrity-consumer) had no or only limited

103
N. Albert et al. Journal of Business Research 81 (2017) 96–106

impacts on consumers' attitudes and behaviors, which suggests that extension, further research could address the proximity among part-
researchers must investigate the notion of congruency more closely, ners. For this research, proximity was measured according to person-
especially in terms of the effects of a moderate match between en- ality, which is a cognitive concept (Pervin & John, 1999). However,
dorsement partners. Moderate incongruency offers the advantage of proximity also can be affective, as in the concept of attachment
stimulating consumers' attention so that they analyze the information (Thomson, 2006), or else combine cognitive and affective elements. Do
more deeply and develop more positive attitudes (Fleck et al., 2012). the current results change if the nature of the proximity among the
Further research thus could vary the degree of congruency among the partners shifts? Do the three congruencies exhibit the same influences
brand, the celebrity, and the consumer to determine if moderate in- and importance? Further studies should seek answers to such questions.
congruency might be preferable to a strong match. Third, in a related

Appendix A. Celebrity endorser characteristics

Celebrity endorser N Attractiveness Expertise Reliability Credibility

Mean SD Mean SD Mean SD Mean SD

Dim Eva Longoria 199 5.27 0.80 4.47 0.91 4.27 0.99 4.67 0.82
Marion Cotillard 201 5.05 0.79 4.69 0.84 4.60 0.95 4.78 0.79
Moët Monica Belluci 121 5.43 0.69 4.84 0.82 4.71 0.81 4.99 0.71
Penelope Cruz 199 5.31 0.81 4.60 0.95 4.47 0.92 4.79 0.81
a
Estimated means on six points and five items for each dimension; semantic differential.a

Appendix B. Two ads used in the research

Appendix C. Measurement scales of personalities

C.1. Loadings, means, and SD

Brand personality Consumer personality Celebrity personality

Loading Mean SD Loading Mean SD Loading Mean SD

Sincerity
Down-to-earth 0.82 4.31 1.08 0.77 4.87 1.01 0.84 4.15 1.14
Honest 0.85 4.39 1.12 0.80 5.15 0.91 0.87 4.30 1.08
Wholesome 0.80 4.39 1.20 0.80 4.72 1.05 0.87 4.40 1.12
Cheerful 0.75 4.45 1.16 0.78 4.71 1.06 0.75 4.74 1.03
Excitement
Daring 0.83 4.29 1.23 0.79 3.85 1.26 0.81 4.43 1.10
Spirited 0.75 3.87 1.32 0.76 3.56 1.37 0.78 4.23 1.18
Imaginative 0.79 4.17 1.27 0.74 4.30 1.24 0.79 4.14 1.11
Up-to-date 0.86 4.44 1.19 0.76 4.34 1.12 0.77 4.84 0.99
Competence
Reliable 0.76 4.71 1.10 0.80 4.98 0.99 0.86 4.22 1.14
Intelligent 0.85 4.35 1.12 0.87 4.46 1.02 0.90 4.53 1.08
Successful 0.85 4.35 1.17 0.77 4.02 1.19 0.82 4.74 1.02
Sophistication
Upper class 0.91 4.52 1.22 0.88 3.84 1.25 0.88 4.78 1.07
Charming 0.91 4.54 1.24 0.88 3.86 1.34 0.88 5.09 0.96

104
N. Albert et al. Journal of Business Research 81 (2017) 96–106

Ruggedness
Outdoorsy 0.83 3.87 1.31 0.84 4.58 1.30 0.89 3.97 1.24
Tough 0.83 4.4 1.20 0.84 4.07 1.17 0.89 3.8 1.20
Note: Each trait adjective was measured on six-point Likert scales (from “1—not appropriate at all” to “6—totally appropriate”).

C.2. Reliability, validity, and correlations

No. of α CR 1 3 2 4 5
items

Brand personality
1 Sincerity 4 0.82 0.88 (0.65)
2 Excitement 4 0.82 0.88 0.48 (0.66)
3 Competence 3 0.76 0.86 0.58 0.46 (0.68)
4 Sophistication 2 – 0.91 0.30 0.39 0.42 (0.83)
5 Ruggedness 2 – 0.82 0.49 0.35 0.39 0.20 (0.69)
Consumer personality
1 Sincerity 4 0.80 0.87 (0.62)
2 Excitement 4 0.76 0.85 0.35 (0.58)
3 Competence 3 0.74 0.85 0.50 0.41 (0.66)
4 Sophistication 2 – 0.87 0.22 0.49 0.36 (0.77)
5 Ruggedness 2 – 0.82 0.35 0.22 0.30 0.09 (0.70)
Celebrity personality
1 Sincerity 4 0.85 0.90 (0.70)
2 Excitement 4 0.79 0.87 0.54 (0.62)
3 Competence 3 0.82 0.90 0.72 0.59 (0.74)
4 Sophistication 2 – 0.87 0.34 0.44 0.46 (0.77)
5 Ruggedness 2 – 0.88 0.49 0.32 0.41 0.14 (0.79)
Notes: α = Cronbach's alpha, CR = composite reliability. All correlations are significant at p = 0.01. The average variances extracted are along the diagonal.
a
All squared correlations are below the AVE (discriminant validity check).a

Appendix D. Measurement scales of endorsement effectiveness

Loading Mean SD

Brand identification (four items from Escalas & Bettman, 2003)

α = 0.94; CR = 0.96; AVE = 0.86
Brand X reflects who I am. 0.95 2.64 1.46
I can identify with Brand X. 0.90 2.83 1.44
I use Brand X to communicate who I am to other people. 0.91 2.40 1.42
I think Brand X helps me become the type of person I want to be. 0.93 2.43 1.46
Brand attitude (three items from Park et al., 2010)
α = 0.95; CR = 0.97; AVE = 0.91
Good vs. bad 0.95 4.86 1.02
Positive vs. negative 0.96 4.82 1.07
Liked it vs. disliked it 0.96 4.73 1.09
Brand commitment (five items from Fullerton, 2005)
α = 0.93; CR = 0.95; AVE = 0.78
I feel emotionally attached to Brand X. 0.91 2.72 1.50
Brand X has a great deal of personal meaning for me. 0.86 3.21 1.43
It would be very hard for me to switch away from Brand X right now even if I wanted to. 0.86 2.89 1.54
My life would be disrupted if I switched away from Brand X. 0.87 2.34 1.47
It would be too costly for me to switch from Brand X right now. 0.91 2.84 1.54
Brand behavioral intention (six items from Park et al., 2010)
α = 0.92; CR = 0.94; AVE = 0.71
I usually wait to buy Brand X products rather than buying non-Brand X ones. 0.86 3.08 1.58
I defend Brand X when others speak poorly about it. 0.85 2.95 1.51

105
N. Albert et al. Journal of Business Research 81 (2017) 96–106

I am ready to pay more for Brand X products than for others. 0.87 3.12 1.58
I frequently use products with a Brand X logo. 0.83 2.80 1.63
I already have recommended Brand X products to other people. 0.86 3.30 1.61
I am used to buying Brand X products. 0.79 3.75 1.63
Notes: α = Cronbach's alpha, CR = composite reliability, AVE = average variance extracted. Items were measured on six-point Likert scales (from “1—not agree at all” to “6—totally
agree”).

Appendix E. HTMT matrix among the personality of each endorsement figure (brand, celebrity, consumer)

Celebrity personality Consumer personality Brand personality

Celebrity personality – –
Consumer personality 0.51 – –
Brand personality 0.59 0.58

References research. Journal of Consumer Research, 30(2), 199–218.

Kamins, M. A., & Gupta, K. (1994). Congruence between spokesperson and product type:
A matchup hypothesis perspective. Psychology and Marketing, 11(6), 569–586.
Aaker, J. L. (1997). Dimensions of brand personality. Journal of Marketing Research, Kelman, H. C. (1961). Processes of opinion change. Public Opinion Quarterly, 25(1), 57–78.
34(August), 347–356. Knoll, J., & Matthes, J. (2017). The effectiveness of celebrity endorsements: A meta-
Arsena, A., Silvera, D. H., & Pandelaere, M. (2014). Brand trait transference: When ce- analysis. Journal of the Academy of Marketing Science, 45(1), 55–75.
lebrity endorsers acquire brand personality traits. Journal of Business Research, 67(7), Kressmann, F., Sirgy, M. J., Herrmann, A., Huber, F., Huber, S., & Lee, D. J. (2006). Direct
1537–1543. and indirect effects of self-image congruence on brand loyalty. Journal of Business
Batra, R., Lenk, P., & Wedel, M. (2010). Brand extension strategy planning: Empirical Research, 59(9), 955–964.
estimation of brand-category personality fit and atypicality. Journal of Marketing McCracken, G. (1989). Who is the celebrity endorser? Cultural foundations of the en-
Research, 47(2), 335–347. dorsement process. Journal of Consumer Research, 16(3), 310–321.
Belk, R. W. (1988). Possessions and the extended self. Journal of Consumer Research, Misra, S., & Beatty, S. E. (1990). Celebrity spokesperson and brand congruence: An as-
15(2), 139–168. sessment of recall and affect. Journal of Business Research, 21(2), 159–173.
Bergkvist, L., & Zhou, K. Q. (2016). Celebrity endorsements: A literature review and re- Mowen, J. C. (1980). On product endorser effectiveness: A balance model approach.
search agenda. International Journal of Advertising, 35(4), 642–663. Current Issues and Research in Advertising, 3(1), 41–57.
Berry, N. C. (1988). Revitalizing brands. Journal of Consumer Marketing, 5(3), 15–20. Mowen, J. C., & Brown, S. W. (1981). On explaining and predicting the effectiveness of
Boon, S. D., & Lomore, C. D. (2001). Admirer-celebrity relationships among young adults. celebrity endorsers. Advances in Consumer Research, 8, 437–441.
Human Communication Research, 27(3), 432–465. Ohanian, R. (1991). The impact of celebrity spokespersons' perceived image on con-
Branaghan, R. J., & Hildebrand, E. A. (2011). Brand personality, self-congruity, and sumers' intention to purchase. Journal of Advertising Research, 46–54
preference: A knowledge structures approach. Journal of Consumer Behaviour, 10(5), (February–March).
304–312. Park, C. W., MacInnis, D. J., Priester, J., Eisingerich, A. B., & Iacobucci, D. (2010). Brand
Broderick, A., Jogi, A., & Garry, T. (2003). Tickled pink: The personal meaning of cause- attachment and brand attitude strength: Conceptual and empirical differentiation of
related marketing for customers. Journal of Marketing Management, 19(5–6), 583–610. two critical brand equity drivers. Journal of Marketing, 74(6), 1–17.
Choi, S. M., & Rifon, N. J. (2012). It is a match: The impact of congruence between Pervin, L. A., & John, O. P. (1999). Handbook of personality: Theory and research (2nd ed).
celebrity image and consumer ideal self on endorsement effectiveness. Psychology and New York: Guilford.
Marketing, 29(9), 639–650. Pornpitakpan, C. (2003). Validation of the celebrity endorsers' credibility scale: Evidence
Eisend, M., & Stokburger-Sauer, N. E. (2013). Brand personality: A meta-analytic review from Asians. Journal of Marketing Management, 19(1–2), 179–195.
of antecedents and consequences. Marketing Letters, 24(3), 205–216. Roy, S., Gammoh, B. S., & Koh, A. C. (2012). Predicting the effectiveness of celebrity
Erdogan, B. Z. (1999). Celebrity endorsement: A literature review. Journal of Marketing endorsements using the balance theory. Journal of Customer Behaviour, 11(1), 33–52.
Management, 15(4), 291–314. Seno, D., & Lukas, B. A. (2007). The equity effect of product endorsement by celebrities: A
Escalas, J. E., & Bettman, J. R. (2003). You are what they eat: The influence of reference conceptual framework from a co-branding perspective. European Journal of Marketing,
groups on consumers' connections to brands. Journal of Consumer Psychology, 13(3), 41(1/2), 121–134.
339–348. Silvera, D. H., & Austad, B. (2004). Factors predicting the effectiveness of celebrity en-
Fleck, N., Korchia, M., & Le Roy, I. (2012). Celebrities in advertising: Looking for con- dorsement advertisements. European Journal of Marketing, 38(11/12), 1509–1526.
gruence or likability? Psychology and Marketing, 29(9), 651–662. Sirgy, M. J., Lee, D. J., Johar, J. S., & Tidwell, J. (2008). Effect of self-congruity with
Fornell, C., & Larcker, D. F. (1981). Evaluating structural equation models with un- sponsorship on brand loyalty. Journal of Business Research, 61(10), 1091–1097.
observable variables and measurement error. Journal of Marketing Research, 18(1), Thomson, M. (2006). Human brands: Investigating antecedents to consumers' strong at-
39–50. tachments to celebrities. Journal of Marketing, 70(3), 104–119.
Fournier, S. (1998). Consumers and their brands: Developing relationship theory in Till, B. D., & Busler, M. (2000). The match-up hypothesis: Physical attractiveness, ex-
consumer research. Journal of Consumer Research, 24(4), 343–373. pertise, and the role of fit on brand attitude, purchase intent and brand beliefs.
Friedman, H. H., & Friedman, L. (1979). Endorser effectiveness by product type. Journal Journal of Advertising, 29(3), 1–13.
of Advertising Research, 19(5), 63–71. Tuškej, U., Golob, U., & Podnar, K. (2013). The role of consumer-brand identification in
Fullerton, G. (2005). The impact of brand commitment on loyalty to retail service brands. building brand relationships. Journal of Business Research, 66(1), 53–59.
Canadian Journal of Administrative Sciences, 22(2), 97–110. Tzoumaka, E., Tsiotsou, R. H., & Siomkos, G. (2016). Delineating the role of endorser's
Goldsmith, R. E., Lafferty, B. A., & Newell, S. J. (2000). The impact of corporate cred- perceived qualities and consumer characteristics on celebrity endorsement effec-
ibility and celebrity credibility on consumer reaction to advertisements and brands. tiveness. Journal of Marketing Communications, 22(3), 307–326.
Journal of Advertising, 29(3), 43–54. Van der Lans, R., Van den Bergh, B., & Dieleman, E. (2014). Partner selection in brand
Goldsmith, R. E., & Yimin, Z. (2014). The influences of brand-consumer and cause-con- alliances: An empirical investigation of the drivers of brand fit. Marketing Science,
gruence on consumer responses to cause-related marketing. Journal of Applied 33(4), 551–566.
Marketing Theory, 5(2), 74–95. Winer, R. S. (1999). Experimentation in the 21st century: The importance of external
Grubb, E. L., & Grathwohl, H. L. (1967). Consumer self-concept, symbolism and market validity. Journal of the Academy of Marketing Science, 27(3), 349–358.
behavior: A theoretical approach. Journal of Marketing, 31(4), 22–27. Zamudio, C. (2016). Matching with the stars: How brand personality determines celebrity
Hair, J. F., Jr., & Hult, G. T. M. (2016). A primer on partial least squares structural equation endorsement contract formation. International Journal of Research in Marketing, 33(2),
modeling (PLS-SEM). Thousand Oaks: Sage. 409–427.
Henseler, J., Ringle, C. M., & Sinkovics, R. R. (2009). The use of partial least squares path Zhou, L., & Whitla, P. (2013). How negative celebrity publicity influences consumer at-
modeling in international marketing. Advances in International Marketing, 20(1), titudes: The mediating role of moral reputation. Journal of Business Research, 66(8),
277–319. 1013–1020.
Jarvis, C. B., MacKenzie, S. B., & Podsakoff, P. M. (2003). A critical review of construct
indicators and measurement model misspecification in marketing and consumer

106