What We Know About Celebrity Endorsement in Advertising

Enhancing Brand Credibility

Via Celebrity Endorsement
Trustworthiness Trumps
Attractiveness and Expertise

Stephen W. Wang Building on source-credibility theory, the authors tested a structural model for advertisers
National Taiwan Ocean
and scholars to explain brand outcomes of celebrity endorsement. The empirical context is
University
stephen@mail.ntou.edu.tw the global airline industry, with a fitting sample surveyed at an international airport
(N = 637). Results of structural equation analyses show that consumers’ perception of a
Angeline Close
Scheinbaum celebrity endorser’s attractiveness and trustworthiness brought a lift in brand attitude,
University of Texas brand credibility, and purchase intention toward endorsed brands. The contribution
at Austin
to source-credibility theory is the finding that endorser trustworthiness was the only
angeline@
austin.utexas.edu component of source credibility that was important to low-involvement consumers. The
takeaway for advertisers in this industry is to use attractive celebrity endorsers with a
global appeal who are trustworthy to enhance brand credibility.

INTRODUCTION necessarily enhance advertisement likability (Tay-

In advertising, it has been said that “nothing sells lor, 2016; Tomkovik, Yelkur, and Christians, 2001).
like celebrity,” but esteemed advertising scholars
1
In the last five years of Super Bowl advertising,
recently have warned that including a celebrity advertisements with celebrities underperformed
endorser in advertising is “not necessarily a recipe slightly (Taylor, 2016).
for success” (Taylor, 2016, p. 167). Celebrities do not Despite mixed results, there remains a sustained
interest in celebrity endorsement in the advertis-
1
J. Creswell. “Nothing Sells Like Celebrity.” The New York Times, ing industry. Approximately 20 percent of all tel-
June 22, 2008. Retrieved August 22, 2017, from http://www.nytimes. evision commercials incorporate a celebrity, which
com/2008/06/22/business/media/22celeb.html?mcubz=0.

• In airline-industry advertising, consumer trustworthiness of a celebrity endorser influences brand

credibility for the brand the celebrity endorsed.

• Endorser trustworthiness is the only component of source credibility theory that is important to
airline consumers who are not highly involved with airline travel.

• Endorser attractiveness is an important factor to consumers who are involved highly.

Received (in revised form) October
• Regardless of involvement, airline consumers’ perception of a celebrity endorser’s attractiveness
29, 2016; accepted November 4,
and trustworthiness lifts attitude and perception of credibility toward the endorsed brand, which
2016; published online November
leads to purchase intention.
13, 2017

16 JOURNAL OF ADVERTISING RESEARCH March 2018DOI: 10.2501/JAR-2017-042

 Enhancing Brand Credibility Via Celebrity Endorsement thearf.org

traditionally has equated to 10 percent of equates to more successful marketing in this journal (Erdogan, Baker, and Tagg,
all advertising expenditures going to pay communications. Although it has been 2001) took a practitioner’s perspective on
the celebrity endorser (Agrawal and Kam- established that celebrity sources optimize celebrity endorsement, and with almost
akura, 1995). To the authors’ knowledge, consumer information processing, a celeb- 500 citations, it suggested a strong interest
there is no confirmed research suggest- rity source does not necessarily contribute in both scholarship and practice for this
ing that the occurrence of this practice has to the message or a positive brand outcome. topic. An intended contribution to practice
changed much since that study. The 2016 It can be risky for a sponsor to be associated of this article therefore is actionable mana-
Super Bowl had such a notable rise in with celebrities; modern celebrities’ lives gerial implications.
celebrity endorsers in the advertisements are public and more global with the pro-
that the media termed it the “Celeb Bowl” liferation of social media. For international LITERATURE REVIEW AND CONCEPTUAL
(Poggi, 2016; Taylor, 2016). brands, celebrity endorsers can have more FRAMEWORK
One reason for the sustained interest in international exposure. Source-Credibility Theory
celebrity endorsement in the advertising The source-credibility model represents
industry is that a celebrity source for the Objective and Intended Contribution an established theory of source credibility
message can help consumers process the The authors applied source-credibility the- that can help explain or predict message
information. It is important to optimize ory to explain how the celebrity endorser’s efficacy. It captures the three most influ-
the processability of a message (Carrillat, source characteristics affect outcomes for ential source effects on purchase intent,
d’Astous, and Charette Couture, 2015), the endorsed brand. The objective was to brand attitudes, and attitude toward an
and one way to do so is to include a celeb- examine the impact of celebrity endorsers’ advertisement (Amos, Holmes, and Strut-
rity. Celebrities likely will help attract source characteristics, such as trustwor- ton, 2008). The source-credibility model
and maintain attention in advertising thiness, expertise, and attractiveness on suggests that the efficacy of the message
(Kamins, 1989); they also may have a role consumers’ brand attitude, brand credibil- communicated by an endorser depends
in how consumers perceive an endorsed ity, and purchase intention for endorsed on consumers’ perception of the brand
brand’s credibility. brands. A secondary objective was to endorser ’s expertise, trustworthiness,
A recent meta-analysis of almost 50 examine how consumer involvement mod- and attractiveness (e.g. Amos et al., 2008;
studies on celebrity-endorsement effi- erates these relationships. Friedman, Termini, and Washington, 1976;
cacy (Knoll and Matthes, 2017) rejected The intended contribution is relevant to Hovland and Weiss, 1951; McGuire, 1969;
many average effects in areas such as both scholars and advertising or market- Ohanian, 1991). The impact of deploying
enhancement of awareness, endorsement ing professionals. Whereas extant studies a credible spokesperson (Goldsmith, Laf-
explicitness, and endorsement frequency. often have focused on one or two axioms ferty, and Newell, 2000; Mathur, Mather,
The meta-analysis found that endorser of the theory, the authors of this study and Rangan, 1997) improves the persua-
match, the consumer’s familiarity with the tested three components of source cred- siveness of the message.
endorser, and attitude toward the endorsed ibility—expertise, trustworthiness, and In the current study, the authors exam-
object each were significant (Knoll and attractiveness. The authors also examined ined three components of source cred-
Matthes, 2017). Consumers’ evaluations the indirect effect that consumer involve- ibility—expertise, trustworthiness, and
of celebrity source characteristics might ment has on established relationships attractiveness—while also testing the role
reflect on the endorsed brand. It, therefore, in the source-credibility model. Another of consumer involvement. This resulted
is important to examine aspects of source aspect of the contribution is that extant in a theory-based conceptual framework
credibility—namely, a celebrity endors- scholarship on brand credibility (Erdem (See Figure 1). The framework is based on
er’s attractiveness, trustworthiness, and and Swait, 2004; Erdem, Swait, and Valen- expertise, trustworthiness, and attractive-
expertise—from a source-credibility theo- zuela, 2006; Rao, Qu, and Ruekert, 1999; ness of a celebrity endorser as predictors
retical lens (Erdogan, 1999; McCracken, Rao and Ruekert, 1994; Wernerfelt, 1988) of brand attitude, brand credibility, pur-
1989; McGuire, 1969; Ohanian, 1991; Petty, has not been examined as a function of chase intention, and involvement. An over-
Cacioppo, and Schumann, 1983). celebrity endorsement in advertising. view of the source-credibility components
Research on celebrity endorsement in The model was tested in an industry and brand outcomes follows, along with
advertising is important because advertisers that has yet to be examined with respect ensuing hypotheses, as supported by the
cannot assume that celebrity endorsement to celebrity endorsement. Past scholarship celebrity-endorsement literature.

March 2018 JOURNAL OF ADVERTISING RESEARCH 17

What We Know About Celebrity Endorsement in Advertising

H5
Attractiveness

H2 Brand
Credibility H9
H6
Purchase
Trustworthiness H8 H11 Intention
H3
H10
Brand
H1 Attitude
H7 H12
H4
Expertise

Involvement

Direct effects

Moderating effects

Figure 1 Source Credibility Theory-Based Model of Celebrity Endorsement-Based Brand Outcomes

Expertise trustworthiness is a substantial component and Bailey, 2015). A previous study called
The first theoretical axiom in the model of source-credibility theory. this process a “meaning transfer” in celeb-
of source credibility is expertise. “Exper- rity endorsement (Roy, 2016).
tise” is defined as the degree of perceived Trustworthiness Scholars have established a link from
understanding, skills, and knowledge that Trustworthiness, or how worthy someone perceived expertise to trustworthiness
the endorser has (Hovland, Janis, and Kel- is of trust, is a crucial construct to examine (Erdem and Swait, 2004). In product cate-
ley, 1953). An endorser’s expertise is akin theoretically and managerially. “Perceived gories generally studied, such as packaged
to the source’s qualification, which directly trustworthiness” is defined as the attrib- goods, fashion, and sports equipment,
influences the level of conviction to per- utes of dignity, believability, and honesty main effects between trustworthiness and
suade consumers to purchase that which possessed by the endorser and observed by expertise constructs are statistically sig-
is endorsed. consumers (Friedman et al., 1976). When a nificant; this indicates that higher exper-
Expertise has a positive influence on consumer believes that a source is trust- tise has led to stronger trustworthiness in
both brand attitude and purchase inten- worthy, he or she also assumes that the such other product categories (Erdem and
tion (Till and Busler, 2000). When a con- communicated message is highly believ- Swait, 2004). The authors thus predicted
sumer perceives that a celebrity endorser able (Hovland and Weiss, 1951). An inte- that the more a consumer perceived that
has a high level of expertise, he or she grated brand-promotion strategy could be a celebrity endorser has expertise in a par-
more likely will be persuaded by the mes- associating with celebrity endorsers whom ticular area, the greater would be the trust-
sage in the advertisement (Amos et al., consumers perceive as honest, believable, worthiness the consumer perceived in the
2008; Ohanian, 1991; Speck, Schumann, and dependable. In addition, efficacy of endorser—regardless of product category
and Thompson, 1988). A celebrity with a celebrity endorser may transfer from a or industry.
high expertise is assumed to be more per- greater perceived trustworthiness of the
suasive (Erdogan, 1999; Ohanian, 1991) celebrity by the consumer (Choo, 1964; H1: A celebrity endorser’s expertise
than a celebrity with low levels of per- Horai, Naccari, and Fatoullah, 1974; Hov- positively influences his or her
ceived expertise. In addition to expertise, land and Weiss, 1951; Mishra, Subhadip, perceived trustworthiness.

18 JOURNAL OF ADVERTISING RESEARCH March 2018

 Enhancing Brand Credibility Via Celebrity Endorsement thearf.org

H5
0.325
(2.140**)
Attractiveness
H9
H2 0.198
Brand (4.427**)
H6 0.205 Credibility
0.340 (3.002**)
(2.208**) H8
Purchase
Trustworthiness H3 0.150 H11**
(3.083**)
Intention
0.593
(3.490**) H10
H1 H7 Brand 0.771
0.905 0.225 Attitude (15.732**)
(16.397**) (1.516) H4 H12
–0.162
(–0.853)
Expertise

Involvement

Direct effects

Moderating effects

Figure 2 Results of Conceptual Framework

Attractiveness Likability is the liking for the endorser Brand Attitude

A third axiom of the source-credibility that the consumer inculcated from the The components of the source-credibility
model is attractiveness. In addition to qualification, outward appearance, and model—trustworthiness, expertise, and
both trustworthiness and expertise, a conduct of the endorser (McGuire, 1969). attractiveness—may be associated with
celebrity endorser ’s attractiveness, as Together, these can enhance perceptions a lift in a consumer’s attitude toward the
perceived by consumers, potentially is of attractiveness. brand endorsed by a celebrity. “Brand
important. Source attractiveness directly Consequently, a consumer more attitude” refers to consumers’ attitudes
influences the effectiveness of a commu- likely may accept information given by as the sum of the products of beliefs mul-
nication message (McGuire, 1969). Attrac- an attractive source (Kelman, 1961). A tiplied by the evaluations (Bagozzi and
tiveness entails one’s outward physical consumer’s purchase intention also is Silk, 1983). These brand associations are
appearance—that is, someone who is affected when the source is attractive the attributes, benefits, and attitudes per-
perceived as beautiful, elegant, or classy (Erdogan, 1999; Joseph, 1982; Petroshius ceived by the consumer concerning the
(Amos et al., 2008; McCracken, 1989; Oha- and Crocker, 1989; Petty et al., 1983). An brand (Keller, 1993). Attributes are the
nian, 1990). attractive celebrity has a positive effect on features that describe and characterize
Researchers have found attractiveness brand attitude and purchase intentions the brand, whereas benefits are the self-
to be a function of how similar, familiar, (Kahle and Homer, 1985). On the basis of identified values that the product can
and likable someone is (McGuire, 1969). source-credibility theory and the litera- fulfill for the consumer. Attitudes are the
Similarity represents the resemblance ture, the authors predicted that regardless overall evaluations of the brand from the
that was perceived between the endorser of the product category endorsed, how a consumer’s perspective.
and the consumer. Familiarity is a con- consumer perceived the attractiveness Previous studies have shown that
sumer ’s knowledge about the source of a celebrity endorser would enhance there is a positive impact on a product
that he or she gained through repeated the consumer’s attitude for the endorsed and brand’s attributes through a celeb-
coverage of the celebrity in the media. brand, or brand attitude. rity endorsement (Atkin and Block, 1983;

March 2018 JOURNAL OF ADVERTISING RESEARCH 19

What We Know About Celebrity Endorsement in Advertising

Friedman and Friedman, 1979; Lang- largely by the quality of the information the endorsed product (Bagozzi, Tybout,
meyer and Walker, 1991; McCracken, 1989; conveyed through the marketing strate- Craig, and Sternthal, 1979; Daneshwary
Mowen and Brown, 1981; Till and Busler, gies associated with a brand (Erdem and and Schwer, 2000; Ostrom, 1969).
2000; Till and Shimp, 1998). Hence, the Swait, 1998). Although endorsement com- Finally, as synthesized in the literature
authors tested the following three hypoth- prises some of the signals of the endorsed (Batra and Ray, 1986; MacKenzie, Lutz,
eses in efforts to confirm past research: brand, it seems likely that the credibil- and Belch, 1986; MacKenzie and Spreng,
ity of an endorser subsequently trans- 1992), the authors tested the extent to
H2: A celebrity endorser’s attrac- fers to the brand. Hence, a high extent which brand attitudes and purchase
tiveness positively influences of endorser credibility should lead to intentions relate in the context of celeb-
consumers’ attitude toward the higher brand credibility for the endorsed rity endorsement. They formulated the
endorsed brand. brand: following hypotheses:

H3: A celebrity endorser’s trustwor- H5: A celebrity endorser’s attractive- H9: The credibility of the endorsed
thiness positively influences ness positively influences brand brand positively influences con-
consumers’ attitude toward the credibility for the endorsed sumers’ purchase intention for
endorsed brand. brand. the endorsed brand.

H4: A celebrity endorser’s expertise H6: A celebrity endorser ’s trust- H10: Consumers’ attitude toward the
positively influences consumers’ worthiness positively influ- endorsed brand positively influ-
attitude toward the endorsed ences brand credibility for the ences purchase intention for the
brand. endorsed brand. endorsed brand.

Brand Credibility H7: A celebrity endorser’s expertise Involvement

Just as it is important to examine brand positively influences brand cred- A consumer’s involvement with the prod-
attitude, it is important to note that brand ibility for the endorsed brand. uct or service category is an important
credibility is a related but distinct brand consideration, because it indirectly may
outcome. “Brand credibility” is defined H8: Consumers’ attitude toward the influence various relationships in source
as the believability of the information endorsed brand positively influ- credibility and brand outcomes in celeb-
conveyed by a brand, which requires that ences brand credibility for the rity endorsement. “Involvement” refers to
consumers perceive that the brand has endorsed brand. the perceived importance of specific prod-
the ability and willingness to continu- ucts or services on the basis of customers’
ously deliver what has been promised Purchase Intention needs, values, and interests (Mittal, 1995).
(Erdem and Swait, 1998, 2004; Erdem In addition to brand credibility, it is impor- Customers may have varying levels of
et al., 2006). Previous literature has stud- tant to examine purchase intention for the involvement on dissimilar product cat-
ied brand credibility (Erdem and Swait, endorsed product or service as related to egories (Bloch and Richins, 1983). Highly
2004; Erdem et al., 2006; Rao et al., 1999; celebrity endorsement. Whereas brand involved customers are inclined to display
Rao and Ruekert, 1994; Wernerfelt, 1988); attitudes are summary evaluations, inten- more loyal buying behavior (Flynn and
however, brand credibility generally is tions represent the consumer’s plan to Goldsmith, 1993).
not examined as a function of celebrity carry out a given behavior, such as mak- The authors thus considered indirect
endorsement in advertising. ing a purchase (Eagly and Chaiken, 1993). effects due to the consumer’s involvement
Outside of a celebrity-endorsement Purchase intentions are a consumer’s plan in the industry. Involvement may moder-
context, brand credibility not only affects to make an effort to purchase a brand. ate the effects of brand attitude and brand
consumers’ brand choice (Erdem and Intent to purchase is a kind of decision that credibility on purchase intention for the
Swait, 2004) but moderates consumers’ investigates why a consumer purchases a endorsed product or service. Research has
price sensitivity (Erdem, Swait, and Lou- particular brand. There is an effect of per- yet to examine, however, whether and to
viere, 2002). A company can work toward ceived brand credibility and brand attitude what extent consumer involvement indi-
building brand credibility, which is driven on a consumer’s willingness to purchase rectly affects the link from credibility to

20 JOURNAL OF ADVERTISING RESEARCH March 2018

 Enhancing Brand Credibility Via Celebrity Endorsement thearf.org

purchase intention and the link from atti- A second reason the authors chose Procedure
tude toward the brand to purchase inten- this industry is because there is a need An author administered a self-report sur-
tion. The authors thus formulated the to test these established relationships in vey in person with international airline
following hypotheses: such a service industry, where trust may passengers. The questions were based on
be especially important. Trustworthi- a review of the literature and specific air-
H11: The relationship between cred- ness of the endorser is a focal concept in line service contexts, and the questionnaire
ibility toward a brand and pur- this research. Airlines are a service, and was pretested and revised. The survey was
chase intention is intensified generally a utilitarian one—getting a con- distributed at a large international airport
when the consumer is highly sumer from point A to point B—at that. based on a convenience sampling method
involved. Celebrity endorsement in general often in May 2014.
is seen for products and hedonic goods; A relatively high 637 out of a total of 700
H12: The relationship between atti- trust in advertising may not be quite as questionnaires were deemed usable, yield-
tude toward a brand and pur- important in the clothing, soft-drink, per- ing a 91 percent response rate from those
chase intention is intensified fume, or make-up industries as it is in the who agreed to participate (See Table 2).
when the consumer is highly airline industry. Two reasons this response rate was high

involved. Most of the highest paying brand are the in-person sampling procedure and
endorsements activated in advertising the fact that many people sitting at an air-
are products, not services. Some of the port are waiting and might have time to
METHODS
most successful celebrity endorsements help. The main reason, however, that the
Empirical Context
of all time include Michael Jordan (Nike, fieldwork was done at an international air-
The empirical context selected to examine
Hanes), David Beckham (H&M), Nicole port is because it is a match with the con-
the source-credibility–based model was
Kidman (Chanel), Tiger Woods (Rolex), text, the international airline industry, in
the global airline industry. This empirical
Taylor Swift (Coca-Cola), Britney Spears which the authors examined theory.
context was selected for several reasons.
(Pepsi), Ellen DeGeneres (Cover Girl),
First, the prevalence of celebrity endorse-
and Jessica Simpson (Proactiv).2 Catherine Measures
ment for international airline brands sug-
Zeta-Jones’s endorsement of T-Mobile is an All measures used a 5-point Likert-type
gests a need for scholarship focused in this
exception: Although T-Mobile provides a scale. Because the original scales were in
economically significant sector. The airline
service, her endorsement was also one of English, the questionnaire was translated
industry is witnessing a sustained trend in
the most successful. Testing some estab- in a tripartite process that included lan-
celebrity endorsements that is deserving of
lished relationships in a service industry, guage, back-translation, and a third-party
scholarly focus.
such as the international airline industry, retranslation to confirm the extent to which
Consider the following list of celebrities
can give some important insights. the translation was a conceptually consist-
endorsing an international airline brand:
A third reason for the need to test a more ent attempt to measure each scale item.
Jennifer Aniston, Nicole Kidman, Kobe established model in the airline sector is To measure expertise, trustworthiness,
Bryant, Miranda Kerr, Hugh Jackman, its international nature. Many of the afore- and attractiveness of a celebrity endorser,
Pele, Kevin Costner, Richard Simmons, mentioned celebrity endorsements are for the authors adopted four items from pre-
John Travolta, and Lionel Messi. Celeb- international brands or with international vious research (McCracken, 1989; Oha-
rity endorsement in the airline industry is celebrities. It thus is paramount to examine nian, 1990). Brand attitude had three items
prevalent internationally; examples include source credibility of the celebrity endorser adopted from previous research (Brett,
Virgin Atlantic, Qantas, Singapore Air, and in a different and broader context—service Wentzel, and Tomczak, 2008; Mitchell and
Turkish Air (See Table 1). In an example branding and a more international scope. Olson, 1981). The authors adopted seven
of advertisements activating the celebrity items of the brand credibility construct
endorsement, Turkish Air’s campaign “Feel from prior research (Erdem and Swait,
Like a Star” features celebrity Kevin Cost- 2
A. Greenman. (2014, September 10). “10 Of The Most 2004) and modified them to fit the con-
ner in its advertisements to help position Successful Celeb Endorsements of All Time.” Retrieved text. To measure purchase intention, the
October 26, 2016, from TheRichest: http://www.therichest.
the airline brand as one with “extraordi- authors adopted two items from other
com/rich-list/most-influential/10-of-the-most-successful-
nary service” (See the Appendix). celeb-endorsements-of-all-time/. research (Jamieson and Bass, 1989; Putrevu

March 2018 JOURNAL OF ADVERTISING RESEARCH 21

What We Know About Celebrity Endorsement in Advertising

TABLE 1 Examples of Celebrities Featured in International misspecification and complexity (Hair,

Airline Advertising Anderson, Tatham, and Black, 2006). There
were no univariate or multivariate outliers.
Celebrity Featured in Airline Advertising Airline Name Country of Airline
The value of kurtosis ranged from −.671 to
Actors:
1.645, and skewness ranged from −.197 to
Kevin Costner (American actor) Turkish Airlines Turkey
−.976. These results satisfy evaluation cri-
Nicole Kidman (Australian actress) Etihad Airways United Arab Emirates teria (ranging from −2 to 2; Mardia, 1985).
Katrina Kaif (Bollywood actress) (UAE)
Jennifer Aniston (American actress) Emirates United Arab Emirates Measurement Model, Reliability,
(UAE) And Validity
John Travolta (American actor) Qantas Australia The measurement models were estimated
Miranda Kerr (Australian actress) with LISREL 8.80 (Jöreskog and Sörbom,
Hugh Jackman (Australian actor) 1989). The chi-square statistics were sig-
Mark Webber (American actor) nificant at the .05 level, as expected with
Adam Goodes (Australian actor) a large sample (Doney and Cannon, 1997).
Richard Simmons (American actor/Fitness Air New Zealand New Zealand The values for the comparative fit index,
personality) nonnormed fit index, root-mean-square
Orlando Bloom (English actor) British Airways U.K. error of approximation, and standardized
Helen Mirren (English actress) Virgin Air U.K. root-mean residual were acceptable for the
Daniel Craig (English actor) model, on the basis of the criteria suggested
Lin Chi-Ling (Taiwanese actress) China Airline R.O.C. by previous researchers (Hu and Bentler,
Lin Huai-min (Taiwanese actor) 1999): .95 for the comparative fit index
and nonnormed fit index, .06 for root-
Takeshi Kaneshiro (Taiwanese actor) EVA Airline R.O.C.
mean-square error of approximation, and
Athletes:
.08 for standardized root-mean residual.
Lionel Messi (Argentinean soccer player) Turkish Airlines Turkey
Given that each goodness-of-fit index (See
Kobe Bryant (American basketball player)
Table 3) met the criteria and that the model
Didier Drogba (Ivorian soccer player)
was developed on a theoretical base, there
Nicol David (Squash) KLM Netherlands
were no further model specifications.
Cristiano Ronaldo (Portuguese soccer player) Emirates United Arab Emirates After testing the measurement model,
Pele (Brazilian soccer player) the authors assessed unidimensional-
Kim Yu-Na (South Korean figure skater) Korean Air South Korea ity, reliability, convergent validity, and
Chen Wei-Yin (Taiwanese MLB player) China Airline R.O.C. discriminant validity. The authors first
Daniel Ricciardo (Australian race car driver) Qantas Australia assessed unidimensionality on the basis
Greg Norman (American golfer) of principal-components analyses on all
Kelly Cartwright (Australian Paralympian) items. All items loaded .65 or higher on
the factors, and no cross-loading was
identified. This shows unidimensionality
and Kenneth, 1994). Finally, the authors framework. First, they used exploratory for each of the constructs. With regard to
adopted nine involvement items from factor analysis to check any deviation composite reliability, all Cronbach’s alpha
previous research (Kapferer and Laurent, from the structure of the constructs. Then, values exceeded the suggested .6 bench-
1985a, 1985b; Laurent and Kapferer, 1985; confirmatory factor analysis (CFA) helped mark (Bagozzi and Yi, 1988).
Rodgers and Schneider, 1993). evaluate construct validity regarding con- Next, in a CFA setting, the authors
vergent and discriminant validity before assessed convergent validity (i.e., the
Exploratory Factor Analysis structural path analysis. degree of association between measures
The authors tested the theory-based model The sample size (N = 637) was large of a construct) by examining t statistics
with a two-stage structural equation enough to compensate for model related to the factor loadings. The result

22 JOURNAL OF ADVERTISING RESEARCH March 2018

 Enhancing Brand Credibility Via Celebrity Endorsement thearf.org

Table 2 Sample Characteristics

Characteristics Category Full Sample High Low t
involvement involvement (group
Times % Times % Times % differences)

Gender Male 344 54.0 84 52.8 93 58.5

Female 293 46.0 75 47.2 66 41.5
Age < 20 Years 20   3.1   3   1.9   8   5.0 ***p ≤ 0.01
21–30 129 20.3 45 28.3 26 16.4
31–40 187 29.4 50 31.4 36 22.6
41–50 119 18.7 29 18.2 36 22.6
> 51 182 28.6 32 20.1 53 33.3
Education Elementary   3    .5   1    .6   0   0 ***p ≤ 0.01
completed Junior high   8   1.3   2   1.3   3   1.9
Senior high 155 24.3 26 16.4 52 32.7
College 397 62.3 107 67.3 98 61.6
Graduate school 74 11.6 23 14.5   6   3.8
No. trips abroad 0 180 28.3 31 19.5 77 48.4 ***p ≤ 0.01
last year 1 214 33.6 43 27.0 43 27.0
2 129 20.3 46 28.9 21 13.2
3 58   9.1 24 15.1   7   4.4
4 19   3.0   8   5.0   3   1.9
5 37   5.8   7   4.4   8   5.0
Type of trip abroad Group traveling 431 67.7 101 63.5 106 66.7
DIY tour 121 19.0 38 23.9 29 18.2
Semi DIY tour 55   8.6 13   8.2 15   9.4
Abroad for official business 30   4.7   7   4.4   9   5.7
Expense per abroad <$600 USD 110 17.3 18 11.3 42 26.4 ***p ≤ 0.01
$600–$1,500 USD 332 52.1 78 49.1 84 52.8
$1,500–$2,500 USD 142 22.3 49 30.8 22 13.8
$2,500–$3,000 USD 34   5.3   7   4.4   9   5.7
>$3,000 USD 19   3.0   7   4.4   2   1.3
Seen the endorsed advertisement Yes 500 78.5 131 82.4 105 66.0 ***p ≤ 0.01
No 137 21.5 28 17.6 54 34.0
Been to the location in the Yes 274 43.0 87 54.7 39 24.5 ***p ≤ 0.01
advertisement No 363 57.0 72 45.3 120 75.5
Know the location in the Yes 354 55.6 108 67.9 55 34.6 ***p ≤ 0.01
advertisement No 283 44.4 51 32.1 104 65.4
Sub Total 637 100.0 159 100.0 159 100.0

March 2018 JOURNAL OF ADVERTISING RESEARCH 23

What We Know About Celebrity Endorsement in Advertising

Table 3 Measurement Scales and Summary Statistics

Constructs/ Mean/ Reliability/ Average Variance Explained Scale Source
Attractiveness (M = 3.82; (μHigh Inv. = 4.19; μLow Inv. = 3.47), reliability = 0.849, AVE. = 0.789)
“I prefer to watch advertisements with physically attractive (beautiful, elegant, classy) endorsers.” McCracken (1989),
“I think that attractiveness is an important characteristic for celebrity endorsers.” Ohanian (1990)
“I feel that a physically attractive endorser influences my purchase intention towards a celebrity endorsed brand.”
“I remember a brand that is endorsed by an attractive celebrity more.”
Expertise (M = 3.84 (μHigh Inv. = 4.20; μLow Inv. = 3.45), reliability = 0.813, AVE. = 0.686)
“I think an advertisement with a celebrity endorser who has expertise (skilled, qualified, knowledgeable, experienced) McCracken (1989),
is more respectable.” Ohanian (1990)
“I pay more attention to advertisements using a celebrity with expertise.”
“I will buy a product if the celebrity endorsing it is an expert.”
“I think a brand being endorsed by a celebrity with expertise is more trustable.”
Trustworthiness (M = 3.72 (μHigh Inv. = 4.09; μLow Inv. = 3.40), reliability = 0.820, AVE. = 0.714)
“I think the advertisement with a trustworthy (dependable, honest, sincere, reliable) endorser receives less negative McCracken (1989),
recalls.” Ohanian (1990)
“I feel that an advertisement with a trustworthy endorser pushes me to remember that advertisement and the product
that is being endorsed.”
“I prefer to buy a product if the celebrity endorser is a trustworthy person.”
“I think that a brand endorsed by a trustworthy celebrity is more respectable and desirable.”
Brand Attitude (M = 3.33 (μHigh Inv. = 3.74, μLow Inv. = 2.89), reliability = 0.778; AVE = 0.790)
“I have bought products due to the influence of celebrity endorsements.” Brett et al. (2008),
“I keep using a brand only because of the endorsing celebrity.” Mitchell and Olson
“Celebrities help me to remember a brand.” (1981)
Purchase Intention (M = 3.38 (μHigh Inv. = 3.84, μLow Inv. = 2.89), reliability = 0.671; AVE = 0.743)
“I will buy the product if the celebrity I like starts endorsing it.” Jamieson and Bass
“I would continue to buy the same products from the market irrespective of advertising the same product through any (1989), Putrevu and
specific celebrity.” Kenneth (1994)
Brand Credibility (M = 3.70 (μHigh Inv. = 4.07, μLow Inv. = 3.36) reliability = 0.886; AVE = 0.699)
“This brand reminds me of someone who is competent and knows what he/she is doing.” Erdem and Swait
“This brand has the ability to deliver what it promises.” (2004)
“This brand delivers what it promises.”
“This brand’s product claims are believable.”
“Over time, my experiences with this brand have led me to expect it to keep its promises, no more and no less.”
“This brand has a name you can trust.”
“This brand doesn’t pretend to be something it isn’t.”
Involvement (M = 3.71 (μHigh Inv. = 4.47, μLow Inv. = 2.89), reliability = 0.894; AVE = 0.783)
“I attach great importance to flying in an airplane.” Kapferer and
“Flying in an airplane interests me a lot.” Laurent
“Flying in an airplane leaves me totally indifferent.” (1985a, 1985b),
“It would give me pleasure to purchase airline transportation for myself.” Laurent and
Kapferer (1985),
“When you buy an airline ticket, it is a bit like giving a gift to yourself.”
Rodgers and
“Having a chance to take an airplane is a pleasure for me.”
Schneider (1993)
“You can tell something about a person by the airline transportation service (s)he picks out.”
“The airline transportation service you buy tells a little bit about you.”
“The airline transportation service I buy shows what type of man/woman I am.”

24 JOURNAL OF ADVERTISING RESEARCH March 2018

 Enhancing Brand Credibility Via Celebrity Endorsement thearf.org

that all t statistics were statistically signifi- one-factor test, exploratory factor analy- they had been to the location in the adver-
cant at the .05 level suggests that all indica- sis generated four factors; no factors had tisement, and knowledge of the location
tor variables provided sound measures to a factor loading larger than 50 percent in the advertisement.
their construct, offering evidence of con- (Podsakoff and Organ, 1986). Finally,
vergent validity (Hoyle and Panter, 1995). when the authors compared the difference Model Fit
Average variance extracted values related of CFA from a single factor and multifac- The authors used simultaneous maximum-
to all constructs at or higher than .50, offer- tor structure (Lindell and Whitney, 2001), likelihood-estimation procedures to
ing evidence for convergent validity (For- Δχ = 388.092, Δdf = 24, p = .0 (Δχ /Δdf =
2 2
examine relationships among expertise,
nell and Larcker, 1981). 16.1705). CMV therefore did not affect trustworthiness, attractiveness, brand
Finally, the authors verified discri- validity (Carson, 2007). attitude, brand credibility, purchase inten-
minant validity using the procedures In addition, the authors examined any tion, and involvement. The overall fit of
recommended by previous research- nonresponse bias. As suggested in earlier the model was acceptable, on the basis of
ers (Anderson, 1987; Bagozzi and Phil- studies, the authors assessed nonresponse many fit statistics (See Table 4).
lips, 1982). They performed a series of bias by comparing early and late respond-
chi-square difference tests on the nested ents across all variables in the study RESULTS
models to assess whether the chi-square (Armstrong and Overton, 1977). No sig- Main Effects
values were significantly lower for the nificant difference existed across the early Results support eight of the ten direct
unconstrained models where the phi coef- and late responders. path hypotheses. Two of the three source-
ficient was constrained to unity (Ander- credibility axioms examined—trust-
son, 1987). The critical values related to Multigroup Analysis worthiness and attractiveness—had
the chi-square difference (at the .05 sig- The sample size (N = 637) was consid- significant positive effects on both brand
nificance level) were higher than 3.84 in ered sufficiently large for model specifica- attitude and brand credibility in the air-
all possible pairs of constructs; this out- tion and multigroup analysis (Chen and line industry context, which supports
come gives support to discriminant valid- Chang, 2008; Sichtmann, 2007). To test any Hypotheses 2, 3, 5, and 6. Expertise was
ity. The constructs thus met reliability and moderating effects of involvement with air still important, however, and had a sig-
validity standards. travel, the authors performed chi-square nificant positive effect on trustworthiness,
difference tests on the nested models to which supports H1. Positive brand atti-
Common Method Variance Testing assess whether the chi-square values were tude significantly influenced both brand
Common method variance (CMV) is a significantly lower for the unconstrained credibility and purchase intention, which
potential problem in behavioral research models with the phi coefficient constrained supports Hypotheses 8 and 10, respec-
(Bagozzi and Yi, 1990; Cote and Buckley, to unity (Anderson, 1987). The authors tively. The effect of brand credibility on
1987, 1988; Williams, Buckley, and Cote, tested moderating effects of involvement purchase intention was also significant,
1989). The authors adopted procedural by median splitting the sample into high-, which supports H9.
remedies, including temporal, proximal, medium-, and low-involvement groups
psychological, or methodological separa- (Aiken and West, 1996). They then ran Moderated Effects
tion of measurement, to eliminate CMV structural equation modeling with two To examine the moderating effect of
(Podsakoff, MacKenzie, Lee, and Podsa- reduced groups (high involvement, n = involvement (Hypotheses 11 and 12), the
koff, 2003). The authors also protected 159; low involvement, n = 159). authors divided the sample using the
respondent anonymity and reduced evalu- Before hypothesis testing, the authors quartiles of the level of these involve-
ation apprehension while balancing ques- tested the difference between these two ment items (Chen and Chang, 2008). The
tion order (Carson, 2007). groups. Results of t tests revealed that first and last quartiles were defined as
This study came to an insignificant (p highly involved consumers were signifi- high-involvement (μ = 4.47) and low-
= .95, p > .05) conclusion for CMV test cantly different from low-involvement involvement (μ = 2.89) groups and subse-
by unmeasured latent method construct consumers in many items, including age, quently were used for testing the causal
(Richardson, Simmering, and Sturman, education, yearly trips abroad, expense relationships. For the high-involvement
2009; Williams et al., 1989). When the per trip abroad, whether they had seen group, the effect of attractiveness on brand
authors tested CMV with Harman’s the endorsed advertisement, whether attitude (H2), the effect of trustworthiness

March 2018 JOURNAL OF ADVERTISING RESEARCH 25

What We Know About Celebrity Endorsement in Advertising

Table 4 Structural Parameter Estimates and Goodness-of-Fit Indices

Direct Paths Hypothesis Full-Sample Moderating Effects of Involvement (H11–H12)
Model High Inv. Low Inv.
Expertise→Trustworthiness H1 0.905(16.397**) 0.961(7.025**) 0.903(6.901**)
Attractiveness→Brand Attitude H2 0.205(3.002**) 0.245(1.001) 0.128(1.073)
Trustworthiness→Brand Attitude H3 0.593(3.490**) 0.774(3.357**) 0.398(2.348**)
Expertise→Brand Attitude H4 -0.162(-0.853) -0.15(-0.452) 0.017(0.015)
Attractiveness→Brand Credibility H5 0.325(2.140**) 0.284(2.084**) 0.075(0.82)
Trustworthiness→Brand Credibility H6 0.340(2.208**) 0.138(1.010) 0.162(0.184)
Expertise→Brand Credibility H7 0.255(1.516) 0.217(1.224) 0.317(0.371)
Brand Attitude→Brand Credibility H8 0.150(3.083**) 0.411(2.841**) 0.062(0.743)
Brand Credibility→Purchase Intention H9 0.198(4.427**) 0.339(2.713**) 0.134(1.009)
Brand Attitude→Purchase Intention H10 0.771(15.732**) 0.906(9.211**) 0.778(5.964**)
Fit statistics χ =835.693
2
μHigh Inv. =4.47 μLow Inv. =2.89
(P=.00, df=232)
χ2/df = 3.602,
GFI=0.893,
CFI=0.928,
RMSEA=0.048
Moderated path in the chi-square Chi-square difference test between low
difference test and high involvement for moderated path
Brand Credibility→Purchase Intention H11 Δχ2=4.73, df=1, H11 Supported
p<.05
Brand Attitude→Purchase Intention H12 Δχ2=1.85, df=1, H12 Not supported
p>.05
Sample size n=637 n=159 n=159

Note. Estimates are standardized. Values in parentheses are t values. Inv. = involvement; GFI = goodness-of-fit index; CFI = comparative fit index; RMSEA = root-mean-square error of
approximation. **p ≤ .05.

on brand credibility (H6), the effect of Hypothesis 11 was supported—which validity. The constructs thus met reliability
expertise on brand attitude (H4), and the means that consumer involvement and validity standards.
effect of expertise on brand credibility strengthened the associations among 1982).
(H7) were not significant (See Table 4). In They performed a series of chi-square dif- Common Method Variance Testing
the highly involved group, therefore, there ference tests on the nested models to assess Common method variance (CMV) is a
was support for Hypotheses 1, 3, 5, and whether the chi-square values were signifi- potential problem in behavioral research
8–10. For the low-involvement group, the cantly lower for the unconstrained models (Bagozzi and Yi, 1990; Cote and Buckley,
data support the paths from expertise to where the phi coefficient was constrained 1987, 1988; Williams, Buckley, and Cote,
trustworthiness (H1), trustworthiness to to unity (Anderson, 1987). The critical val- 1989). The authors adopted procedural
brand attitude (H3), and brand attitude to ues related to the chi-square difference (at remedies, including temporal, proxi-
purchase intention (H10). the .05 significance level) were higher than mal, psychological, or methodological
Multigroup structural equation mod- 3.84 in all possible pairs of constructs; this separation of measurement, to elimi-
eling structural path results indicate that outcome gives support to discriminant nate CMV (Podsakoff, MacKenzie, Lee,

26 JOURNAL OF ADVERTISING RESEARCH March 2018

 Enhancing Brand Credibility Via Celebrity Endorsement thearf.org

and Podsakoff, 2003). The authors also reduced groups (high involvement, n = authors divided the sample using the
protected respondent anonymity and 159; low involvement, n = 159). quartiles of the level of these involve-
reduced evaluation apprehension while Before hypothesis testing, the authors ment items (Chen and Chang, 2008). The
balancing question order (Carson, 2007). tested the difference between these two first and last quartiles were defined as
This study came to an insignificant (p groups. Results of t tests revealed that high-involvement (μ = 4.47) and low-
= .95, p > .05) conclusion for CMV test highly involved consumers were signifi- involvement (μ = 2.89) groups and subse-
by unmeasured latent method construct cantly different from low-involvement quently were used for testing the causal
(Richardson, Simmering, and Sturman, consumers in many items, including age, relationships. For the high-involvement
2009; Williams et al., 1989). When the education, yearly trips abroad, expense group, the effect of attractiveness on brand
authors tested CMV with Harman’s one- per trip abroad, whether they had seen attitude (H2), the effect of trustworthi-
factor test, exploratory factor analysis the endorsed advertisement, whether ness on brand credibility (H6), the effect
generated four factors; no factors had a they had been to the location in the adver- of expertise on brand attitude (H4), and
factor loading larger than 50 percent (Pod- tisement, and knowledge of the location the effect of expertise on brand credibility
sakoff and Organ, 1986). Finally, when in the advertisement. (H7) were not significant (See Table 4). In

the authors compared the difference of the highly involved group, therefore, there
Model Fit was support for Hypotheses 1, 3, 5, and
CFA from a single factor and multifac-
The authors used simultaneous maximum- 8–10. For the low-involvement group, the
tor structure (Lindell and Whitney, 2001),
likelihood-estimation procedures to data support the paths from expertise to
Δχ2 = 388.092, Δdf = 24, p = .0 (Δχ2/Δdf =
examine relationships among expertise, trustworthiness (H1), trustworthiness to
16.1705). CMV therefore did not affect
trustworthiness, attractiveness, brand brand attitude (H3), and brand attitude to
validity (Carson, 2007).
attitude, brand credibility, purchase inten- purchase intention (H10).
In addition, the authors examined any
tion, and involvement. The overall fit of Multigroup structural equation mod-
nonresponse bias. As suggested in earlier
the model was acceptable, on the basis of eling structural path results indicate that
studies, the authors assessed nonresponse
many fit statistics (See Table 4). Hypothesis 11 was supported—which
bias by comparing early and late respond-
means that consumer involvement
ents across all variables in the study
RESULTS strengthened the associations among
(Armstrong and Overton, 1977). No sig-
Main Effects attractiveness and brand credibility, brand
nificant difference existed across the early
Results support eight of the ten direct attitude and brand credibility, and brand
and late responders.
path hypotheses. Two of the three source- credibility and purchase intention.
credibility axioms examined—trustworthi-
Multigroup Analysis
ness and attractiveness—had significant DISCUSSION
The sample size (N = 637) was consid-
positive effects on both brand attitude and While some paths are replicative in nature,
ered sufficiently large for model specifica-
brand credibility in the airline industry the industry focus and sample add an inter-
tion and multigroup analysis (Chen and context, which supports Hypotheses 2, 3, 5, national dimension to the endorsement lit-
Chang, 2008; Sichtmann, 2007). To test any and 6. Expertise was still important, how- erature. The results of a established model
moderating effects of involvement with air ever, and had a significant positive effect of source credibility tested in the empiri-
travel, the authors performed chi-square on trustworthiness, which supports H1. cal context of the airline industry demon-
difference tests on the nested models to Positive brand attitude significantly influ- strate how a celebrity endorser’s expertise,
assess whether the chi-square values were enced both brand credibility and purchase trustworthiness, and attractiveness affect
significantly lower for the unconstrained intention, which supports Hypotheses 8 brand attitude, brand credibility, and pur-
models with the phi coefficient constrained and 10, respectively. The effect of brand chase intention. The results also show
to unity (Anderson, 1987). The authors credibility on purchase intention was also the moderating effects of involvement in
tested moderating effects of involvement significant, which supports H9. an industry such as air travel, which is a
by median splitting the sample into high-, more distinct contribution with respect
medium-, and low-involvement groups Moderated Effects to more established work on source cred-
(Aiken and West, 1996). They then ran To examine the moderating effect of ibility in advertising. Highly involved cus-
structural equation modeling with two involvement (Hypotheses 11 and 12), the tomers focused on the attractiveness and

March 2018 JOURNAL OF ADVERTISING RESEARCH 27

What We Know About Celebrity Endorsement in Advertising

trustworthiness of the endorser, whereas basis of the trustworthiness of the celebrity (See Table 1), it seems that international
low-involvement customers focused only endorser. These findings reveal the impor- airline brands tend to have endorsements
on trustworthiness. tance of involvement and trustworthi- by actors and, to a lesser extent, athletes.
This takeaway relates to the crucial role ness in air travel service. Trustworthiness Given the importance of the finding that a
of trustworthiness in celebrity endorse- trumps beauty attractiveness and credibil- celebrity endorser’s perceived trustworthi-
ment in the airline industry. Airline con- ity when it comes to celebrity-endorsement ness resonates even with low-involvement
sumers who were low involvement were source characteristics in the airline indus- consumers, future research can examine
not affected as much by the airline celebrity try—where trust is especially paramount. different types of celebrity endorsers—
endorser’s expertise and attractiveness; There is a clear implication for advertis-
actors, athletes, musicians, politicians,
they focused on the celebrity’s trustworthi- ers. These results suggest that advertisers
business leaders, activists—to see whether
ness. Trustworthy celebrity endorsers are looking to enhance their brand credibil-
certain types of celebrities resonate espe-
important in advertising because they res- ity should invest in a celebrity endorser
cially high on trustworthiness.
onate with low-involvement consumers. whom, above all, consumers perceive
Of the source characteristics, trustwor- as being trustworthy. Advertisers using
thiness most aptly explained consumers’ celebrity endorsement should feature an
ABOUT THE AUTHORS
brand attitude toward the endorsed air- attractive celebrity who is perceived to be
line brand by celebrities among custom- trustworthy to best influence both their Stephen W. Wang is professor of marketing at

ers who had low involvement with air low- and their high-involvement con- the Department of Shipping and Transportation

travel. Whereas a celebrity endorser’s sumer base. In a global industry such as Management, College of Maritime Science and

expertise led to trustworthiness, brand airlines, such a celebrity should be inter- Management, National Taiwan Ocean University. He

attitude influenced purchase intention for nationally recognized. develops and tests theory-based models of customer

low-involvement consumers. A celebrity relationship management (CRM), celebrity endorsement,

endorser’s expertise and attractiveness did Limitations and Future Research and consumer behavior in airline, shipping, and logistics
not influence brand attitude or brand cred- There are limitations to this study. First, sectors. His research stream is grounded in a profound
ibility toward the endorsed brand for these this research does not address the long- interest in explaining and predicting linkages among
consumers. Trustworthiness of the celeb- term result of advertising, as suggested by CRM, brand, consumer attitude, affect, cognition,
rity endorser thus was especially impor- previous researchers (Wood and Poltrack, behavioral intent, and consumer behavior. He has
tant for low-involvement consumers. 2015). Second, although the industry exam- published in, among others, the Journal of Air Transport
ined has commonalities with some other Management, the Services Industries Journal, Journal of
Contributions to Source-Credibility Theory relatively utilitarian service sectors, such Service Theory and Practice, Current Issues in Tourism,
And Implications for Advertising as car rentals, the airline industry might
Computers and Human Behavior, and International
There are contributions in this research for not be generalizable to other industries.
Journal of Banking Marketing.
both theory and advertising practice. This The airline industry is more risk oriented,
contribution expands on the model built on given the nature of its service. Interna-
Angeline Close Scheinbaum is associate professor at
source-credibility theory (Erdogan, 1999; tional airlines may be especially sensitive
the University of Texas at Austin at the Stan Richards
McCracken, 1989; McGuire, 1969; Mishra to risk, because of longer or potentially
School of Advertising & Public Relations. Her research
et al., 2015; Ohanian, 1991; Petty et al., more dangerous flights. In light of this,
explains and predicts links among consumer attitude,
1983). The results contribute to the source- however, the finding of the importance of
credibility literature as it empirically delin- endorser trustworthiness has face validity. affect, cognition, behavioral intent, and consumer

eates the importance of trustworthiness The nature of celebrity endorsement in behavior. Close Scheinbaum’s work is published more

above other source characteristics with the airline industry necessitates an endorser than 20 peer-reviewed publications in rigorous journals,

respect to consumer involvement. who has an international presence. In this such as the Journal of Advertising Research, Journal

In contrast to previous studies, this industry context, it seems that international of Academy of Marketing Science, Journal of Business

study classified the participants according airline brands only feature certain types of Research, Journal of Advertising, and European Journal

to their involvement with air travel. Low- celebrities—actors with an international of Marketing. She is coauthor of Advertising & Integrated

involvement consumers tended to evalu- exposure. From the examples of airline Brand Promotion (Cengage, 2017), an advertising
ate the air transportation service on the endorsers in international airline brands textbook focused on branding.

28 JOURNAL OF ADVERTISING RESEARCH March 2018

 Enhancing Brand Credibility Via Celebrity Endorsement thearf.org

REFERENCES Bagozzi, R. P., and Y. Yi.“Assessing Method Vari- Cote, J. A., and M. R. Buckley. “Measure-
ance in Multitrait–Multimethod Matrices: The ment Error and Theory Testing in Consumer
Agrawal, J., and W. A. Kamakura. “The Eco- Case of Self-Reported Affect and Perceptions at Research: An Illustration of the Importance
nomic Worth of Celebrity Endorsers: An Event Work.” Journal of Applied Psychology 75, 5 (1990): of Construct Validation.” Journal of Consumer
Study Analysis.” Journal of Marketing 59, 3 547–560. Research 14, 4 (1988): 579–582.
(1995): 56–63.
Bartra, R., and M. L. Ray. “Affective Responses Daneshwary, R., and K. R. Schwer.“The Associa-
Aiken, L., and S. G. West. Multiple Regression: Mediating Acceptance of Advertising.” Journal of tion Endorsement and Consumers’ Intention to
Testing and Interpreting Interactions. Thousand Consumer Research 13, 2 (1986): 234–249. Purchase.” Journal of Consumer Marketing 17, 3
Oaks, CA: Sage, 1996. (2000): 203–213.
Bloch, P. H., and M. L. Richins. “A Theoretical

Amos, C., G. Holmes, and D. Strutton.“Explor- Model for the Study of Product Importance
Doney, P. M., and J. P. Cannon.“An Examination
ing the Relationship between Celebrity Endorser Perceptions.” Journal of Marketing 47, 3 (1983):
of the Nature of Trust in Buyer–Seller Relation-
Effects and Advertising Effectiveness: A Quanti- 69–81.
ships.” Journal of Marketing 61, 2 (1997): 35–51.
tative Synthesis of Effect Size.” International Jour-
Brett, M., D. Wentzel, and T. Tomczak.“Effects
nal of Advertising 27, 2 (2008): 209–234. Eagly, A. H., and S. Chaiken. The Psychology of
of Susceptibility to Normative Influence and
Attitudes. San Diego, CA: Harcourt Brace Jova-
Type of Testimonial on Attitudes toward Print
Anderson, J. C.“An Approach for Confirmatory novich, 1993.
Advertising.” Journal of Advertising 37, 1 (2008):
Measurement and Structural Equation Modeling
29–43.
of Organizational Properties.” Management Sci- Erdem, T., and J. Swait. “Brand Equity as a Sig-
ence 33, 4 (1987): 525–541. naling Phenomenon.” Journal of Consumer Psy-
Carrillat, F. A., d’Astous, A., and Charette
chology 7, 2 (1998): 131–157.
Couture, M. P. “How Corporate Sponsors Can
Armstrong, S. J., and T. S. Overton.“Estimating
Optimize the Impact of Their Message Content.
Nonresponse Bias in Mail Surveys.” Journal of Erdem, T., and J. Swait.“Brand Credibility, Brand
Mastering the Message: Improving the Process-
Marketing Research 14, 3 (1977): 396–402. Consideration and Choice.” Journal of Consumer
ability And Effectiveness of Sponsorship Activa-
Research 31, 1 (2004): 191–199.
tion.” Journal of Advertising Research 55, 3 (2015):
Atkin, C., and M. Block.“Effectiveness of Celeb-
255–269.
rity Endorsers.” Journal of Advertising Research Erdem, T., J. Swait, and J. Louviere. “The Impact
23, 1 (1983): 57–61. of Brand Credibility on Consumer Price Sensi-
Carson, S. J.“When to Give Up Control of Out-
tivity.” International Journal of Research in Market-
sourced New Product Development.” Journal of
Bagozzi, R. P., and L. W. Phillips.“Representing ing 19, 1 (2002): 1–19.
Marketing 71, 1 (2007): 49–66.
and Testing Organizational Theories: A Holistic
Construal.” Administrative Science Quarterly 27, Chen, C. F., and Y. Y. Chang. “Airline Brand Erdem, T., J. Swait, and A. Valenzuela. “Brands
3 (1982): 459–489. Equity, Brand Preference, and Purchase Inten- as Signals: A Cross-Country Validation Study.”

tions—The Moderating Effects of Switching Journal of Marketing 70, 1 (2006): 34–49.

Bagozzi, R. P., and A. J. Silk. “Recognition and Costs.” Journal of Air Transport Management 14,
the Measurement of Memory for Print Adver- Erdogan, B. Z.“Celebrity Endorsement: A Liter-
1 (2008): 40–42.
tisements.” Marketing Science 2, 2 (1983): 95–134. ature Review.” Journal of Marketing Management
Choo, T.-H. “Communicator Credibility and 15, 4 (1999): 291–314.
Bagozzi, R. P., A. M. Tybout, C. S. Craig, and B. Communication Discrepancy as Determinants
Sternthal. “The Construct Validity of the Tri- of Opinion.” Journal of Social Psychology 64, 1 Erdogan, B. Z., M. J. Baker, and S. Tagg.“Select-

partite Classification of Attitudes.” Journal of (1964): 65–76. ing Celebrity Endorsers: The Practitioner’s Per-

Marketing Research 16, 1 (1979): 88–95. spective.” Journal of Advertising Research 41, 3
Cote, J. A., and M. R. Buckley. “Estimating (2001): 39–48.
Bagozzi, R. P., and Y. Yi. “On the Evaluation of Trait, Method, and Error Variance: Generalizing
Structural Equation Models.” Journal of the Acad- across 70 Construct Validation Studies.” Journal Flynn, L. R., and R. E. Goldsmith. “Applica-
emy of Marketing Science 16, 1 (1988): 74–94. of Marketing Research 24, 3 (1987): 315–318. tion of the Personal Involvement Inventory

March 2018 JOURNAL OF ADVERTISING RESEARCH 29

What We Know About Celebrity Endorsement in Advertising

in Marketing.” Psychology and Marketing 10, 4 Hu, L., and P. M. Bentler.“Cutoff Criteria for Fit Journal of the Academy of Marketing Science 45, 1
(1993): 357–366. Indexes in Covariance Structure Analysis: Con- (2017): 55–75.
ventional Criteria versus New Alternatives.”
Fornell, C., and D. F. Larcker. “Evaluating Structural Equation Modeling 6, 1 (1999): 1–55. Langmeyer, L., and M. Walker. “A First Step
Structural Equation Models with Unobservable to Identify the Meaning in Celebrity Endors-
Variables and Measurement Error.” Journal of Jamieson, L. F., and F. M. Bass.“Adjusting Stated ers.” Advances in Consumer Research 18, 1 (1991):
Marketing Research 18, 1 (1981): 39–50. Intention Measures to Predict Trial Purchase 364–371.
of Net Products: A Comparison of Models and
Friedman, H. H., and L. Friedman. “Endorser Methods.” Journal of Marketing Research 26, 3 Laurent, G., and J.-N. Kapferer. “Measuring
Effectiveness by Product Type.” Journal of Adver- (1989): 336–345. Consumer Involvement Profiles.” Journal of Mar-
tising 19, 5 (1979): 63–71. keting Research 22, 1 (1985): 41–53.
Jöreskog, K. G., and D. Sörbom.LISREL 7: User’s
Friedman, H. H., S. Termini, and R. Washington. Reference Guide. Mooresville, IN: Scientific Soft- Lindell, M. K., and D. J. Whitney. “Account-
“The Effectiveness of Advertisements Utilizing ware, 1989. ing for Common Method Variance in Cross-
Four Types of Endorsers.” Journal of Advertising Sectional Research Designs.” Journal of Applied
5, 3 (1976): 22–24. Joseph, W. B. “The Credibility of Physically Psychology 86, 1 (2001): 114–121.
Attractive Communicators: A Review.” Journal
Goldsmith, R. E., B. A. Lafferty, and S. J. New- of Advertising 11, 3 (1982): 15–24. MacKenzie, S. B., R. J. Lutz, and G. E. Belch.
ell. “The Impact of Corporate Credibility and “The Role of Attitude toward the Ad as a
Celebrity Credibility on Consumer Reaction to Kahle, L. R., and P. M. Homer.“Physical Attrac-
Mediator of Advertising Effectiveness: A Test of
Advertisements and Brands.” Journal of Advertis- tiveness of Celebrity Endorser: A Social Adapta-
Competing Explanations.” Journal of Marketing
tion Perspective.” Journal of Consumer Research
ing 29, 3 (2000): 43–54. Research 23, 2 (1986): 130–143.
11, 4 (1985): 954–961.

Hair, J. F., R. E. Anderson, R. L. Tatham, and MacKenzie, S. B., and R. A. Spreng. “How Does
Kamins, M. A. “Celebrity and Non-Celebrity
W. C. Black.Multivariate Data Analysis (6th ed.). Motivation Moderate the Impact of Central and
Advertising in a Two-Sided Context.” Journal of
Upper Saddle River, NJ: Prentice Hall, 2006. Peripheral Processing on Brand Attitudes and
Advertising Research 29, 3 (1989): 34–42.
Intentions?” Journal of Consumer Research 18, 4
Horai, J., N. Naccari, and E. Fatoullah. “The (1992): 519–529.
Kapferer, J.-N., and G. Laurent. “Consumer
Effects of Expertise and Physical Attractive-
Involvement Profiles: A New Practical Approach
ness upon Opinion Agreement and Liking.” Mardia, K. V.“Mardia’s Test of Multinormality.”
to Consumer Involvement.” Journal of Advertis-
Sociometry 37, 4 (1974): 601–606. In Encyclopedia of Statistical Science, vol. 5, S. Kotz
ing Research 25, 6 (1985a): 48–56.
and N. L. Johnson, eds. New York: Wiley, 1985.
Hovland, C. I., I. L. Janis, and H. H. Kelley.Com-
Kapferer, J.-N., and G. Laurent. “Consumers’
munication and Persuasion: Psychological Studies of Mathur, L. K., I. Mather, and N. Rangan. “The
Involvement Profile: New Empirical Results.”
Opinion Change. New Haven, CT: Yale Univer- Wealth Effects Associated with a Celebrity
In Advances in Consumer Research, vol. 12, E. C.
sity Press, 1953. Endorser: The Michael Jordan Phenomenon.”
Hirschman and M. B. Holbrook, eds. Provo, UT:
Journal of Advertising Research 37, 3 (1997): 67–73.
Association for Consumer Research, 1985b.
Hovland, C. I., and W. Weiss. “The Influence of
Source Credibility on Communication Effec- McCracken, G.“Who Is the Celebrity Endorser?
Keller, K. L.“Conceptualizing, Measuring, and
tiveness.” Public Opinion Quarterly 15, 4 (1951): Cultural Foundations of the Endorsement Pro-
Managing Customer-Based Brand Equity.” Jour-
635–650. cess.” Journal of Consumer Research 16, 3 (1989):
nal of Marketing 57, 1 (1993): 1–22.
310–321.
Hoyle, R. H., and A. T. Panter. “Writing About Kelman, H. C. “Processes of Opinion Change.”
Structural Equation Models.” In Structural Equa- Public Opinion Quarterly 25, 1 (1961): 57–78. McGuire, W. J.“Attitudes and Attitude Change.”
tion Modeling: Concepts, Issues, and Applications, In Handbook in Social Psychology, L. Gardner and
R. H. Hoyle, ed. Thousand Oaks, CA: Sage, Knoll, J., and J. Matthes. “The Effectiveness E. Aronson, eds. New York: Random House,
1995. of Celebrity Endorsements: A Meta-Analysis.” 1969.

30 JOURNAL OF ADVERTISING RESEARCH March 2018

 Enhancing Brand Credibility Via Celebrity Endorsement thearf.org

Mishra, A. S., R. Subhadip, and A. A. Bai- Podsakoff, P. M., S. B. MacKenzie, J.-Y. Lee, and Sichtmann, C.“An Analysis of Antecedents and
ley . “Exploring Brand Personality–Celebrity N. P. Podsakoff. “Common Method Biases in Consequences of Trust in a Corporate Brand.”
Endorser Personality Congruence in Celebrity Behavioral Research: A Critical Review of the European Journal of Marketing 41, 9 (2007):
Endorsements in the Indian Context.” Psychol- Literature and Recommended Remedies.” Jour- 999–1015.
ogy & Marketing 32, 12 (2015): 1158–1174. nal of Applied Psychology 88, 5 (2003): 879–903.
Speck, P. S., D. W. Schumann, and C. Thompson.
Mitchell, A. A., and J. C. Olson. “Are Product Podsakoff, P. M., and D. W. Organ.“Self-Reports “Celebrity Endorsements: Scripts, Schema and
Attribute Beliefs the Only Mediator of Advertis- in Organizational Research: Problems and Roles: Theoretical Framework and Prelimi-
ing Effects on Brand Attitude?” Journal of Mar- Prospects.” Journal of Management 12, 4 (1986): nary Tests.” Advances in Consumer Research 15,
keting Research 18, 3 (1981): 318–332. 531–544. 1 (1988): 69–76.

Mittal, B. “A Comparative Analysis of Four Poggi, G. “Why Super Bowl 50 is Poised to be Taylor, C. R. “Some Interesting Findings about
Scales of Consumer Involvement.” Psychology ‘Celeb Bowl’” Advertising Age, 1 February 2016. Super Bowl Advertising.” International Journal of
and Marketing 12, 7 (1995): 663–682. Advertising 35, 2 (2016): 167–170.
Putrevu, S., and L. R. Kenneth. “Comparative
Mowen, J. C., and S. W. Brown. “On Explaining and Noncomparative Advertising: Attitudinal Till, B. D., and M. Busler. “The Match-Up
and Predicting the Effectiveness of Celebrity Effects under Cognitive and Affective Involve- Hypothesis: Physical Attractiveness, Expertise,
Endorsers.” Advances in Consumer Research 8, 1 ment Conditions.” Journal of Advertising 23, 2 and the Role of Fit on Brand Attitude, Purchase
(1981): 437–441. (1994): 77–91. Intent and Brand Beliefs.” Journal of Advertising
29, 3 (2000): 1–13.
Ohanian, R. “Construction and Validation of a Rao, A. R., L. Qu., and R. W. Ruekert. “Signal-
Scale to Measure Celebrity Endorsers’ Perceived ing Unobservable Quality through a Brand Till, B. D., and T. A. Shimp.“Endorsers in Adver-
Expertise, Trustworthiness, and Attractiveness.” Ally.” Journal of Marketing Research 36, 2 (1999): tising: The Case of Negative Celebrity Informa-
Journal of Advertising 19, 3 (1990): 39–52. 258–268. tion.” Journal of Advertising 27, 1 (1998): 67–82.

Ohanian, R. “The Impact of Celebrity Spokes- Rao, A. R., and R. W. Ruekert.“Brand Alliances Tomkovik, C., Yelkur, R., and Christians, L.“The
persons’ Perceived Image on Consumers’ Inten- as Signals of Product Quality.” Sloan Manage- USA’s Biggest Marketing Event Keeps Getting
tion to Purchase.” Journal of Advertising Research ment Review 36, 3 (1994): 87–97. Bigger: An In-Depth Look at Super Bowl Adver-
31, 1 (1991): 46–54. tising in the 1990s.” Journal of Marketing Commu-
Richardson, H. A., M. J. Simmering, and M. C. nications 7, 3 (2001): 89–108.
Ostrom, T. M. “The Relationship between the Sturman.“A Tale of Three Perspectives: Examin-
Affective, Behavioral, and Cognitive Compo- ing Post Hoc Statistical Techniques for Detection Wernerfelt, B. “Umbrella Branding as a Signal
nents of Attitude.” Journal of Experimental Social and Correction of Common Method Variance.” of New Product Quality: An Example of Signal-
Psychology 5, 1 (1969): 12–30. Organizational Research Methods 12, 4 (2009): ing by Posting a Bond.” Rand Journal of Econom-
762–800. ics 19, 3 (1988): 458–466.
Petroshius, S. M., and K. E. Crocker.“An Empiri-
cal Analysis of Spokesperson Characteristics Rodgers, W. C., and K. C. Schneider.“An Experi- Williams, L. J., M. R. Buckley, and J. A. Cote.
on Advertisement and Product Evaluations.” mental Evaluation of the Kapferer–Laurent Con- “Lack of Method Variance in Self-Reported
Journal of the Academy of Marketing Science 17, 3 sumer Involvement Profile Scale.” Psychology Affect and Perceptions at Work: Reality or Arti-
(1989): 217–225. and Marketing 10, 4 (1993): 333–345. fact?” Journal of Applied Psychology 74, 3 (1989):
462–468.
Petty, R. E., J. T. Cacioppo, and D. Schumann. Roy, S.“Meaning Transfer in Celebrity Endorse-
“Central and Peripheral Routes to Advertising ments: An Explanation Using Metaphors.” Jour- Wood, L. A., and D. F. Poltrack. “Measuring
Effectiveness: The Moderating Role of Involve- nal of Marketing Communications (2016): ePub The Long-Term Effects of Television Advertis-
ment.” Journal of Consumer Research 10, 2 (1983): ahead of print. http://dx.doi.org/10.1080/13527 ing.” Journal of Advertising Research 55, 2 (2015):
135–146. 266.2016.1197294 123–131.

March 2018 JOURNAL OF ADVERTISING RESEARCH 31

What We Know About Celebrity Endorsement in Advertising

APPENDIX
Examples of Celebrity Endorsement in International Airline Advertising

32 JOURNAL OF ADVERTISING RESEARCH March 2018