You are on page 1of 18

Tourism Management 83 (2021) 104213

Contents lists available at ScienceDirect

Tourism Management
journal homepage: http://www.elsevier.com/locate/tourman

Celebrity endorsements in destination marketing: A three


country investigation
Subhadip Roy a, *, Wioleta Dryl b, Luciana de Araujo Gil c
a
Indian Institute of Management Ahmedabad Sargam Marg, Vastrapur Ahmedabad, 380015, India
b
University of Gdańsk, Faculty of Management, Department of Marketing, Armii Krajowej 101, 81-824 Sopot, Poland
c
Universidad Diego Portales, Business Department, Av. Manuel Rodríguez Sur 415, Santiago, Región Metropolitana, Chile

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

Keywords: The present study extends research on the role of celebrity endorsements in destination marketing by exploring
Celebrity endorsements various facets of the effect of celebrity endorsements in destination marketing on the consumer. More specif­
Destination marketing ically, theories of source credibility, congruence, social identity and consumer cosmopolitanism, are used to
Destination advertising
build research questions that investigate the relative effectiveness of a celebrity endorsed tourism advertisement
Multi-country exploration
Experimental design
vis a vis a generic advertisement and the boundary conditions governing the same such as destination type
(local/global), celebrity country of origin and consumer level factors. The research questions are addressed using
four experimental studies in sequence. The same four experiments are run in three countries with different socio-
cultural backgrounds to enhance generalization, with a combined sample size of 1073 respondents. Major
findings suggest that a celebrity endorser is effective for a destination advertisement. Significant cross-country
differences were observed in consumer affect depending on the choice of celebrity (local or global) and the
destination type (i.e., domestic or international). The effects are also moderated by consumer cosmopolitanism.
The study has multiple theoretical and managerial implications.

1. Introduction endorsers to showcase destinations to their potential customers (Chan


et al., 2018) and even tourism ancillaries such as airlines (Kotsi and
Celebrity endorsements for marketing and product promotion is a Valek, 2018).
century-old phenomenon (Erdogan, 1999) and celebrity endorsers have Globally, actors and sports celebrities are the most relatable pool of
been used by marketers to promote products, services and ideas (Knoll endorsers (Yen and Teng, 2015). Notable examples include actor Chris
and Matthes, 2017; Park and Yim, 2020; Schimmelpfennig and Hunt, Hemsworth for Tourism Australia, singer Taylor Swift for New York,
2020). Recently, destination marketers have joined the bandwagon and singer Rihanna for Barbados, actor Jackie Chan for Indonesia and actor
are using celebrities to promote their destinations to both domestic and Bae Yong-Joon for Korea. While destination marketers have used
international tourists. Celebrity endorsers of a destination brand, also indigenous celebrities to market destinations, there are instances where
known as tourism ambassadors, are selected for their universal appeal destination marketers have used celebrities from the major target mar­
across key visitor markets and are often the highlight of marketing kets (a different country) to attract tourists from those markets. For
campaigns (van der Veen, 2009; van der Veen and Song, 2010). Celeb­ example, James Cameron was selected as the ambassador of New Zea­
rity endorsers can play an important role in selection processes of a land tourism to attract tourists from Canada; Arnold Schwarzenegger to
destination by a prospect and subsequent travel behavior (Glover, 2009; attract US tourists to Madrid; and actor Ranveer Singh to attract Indian
Teng & Chen, 2020). Marketers are considering celebrity endorsements tourist to Switzerland (Bapna, 2017; Jamal, 2016). Celebrity endorse­
of tourist destinations as an important choice to boost destination return ments have benefitted destination marketers financially. For example,
on investment (ROI) and spread word of mouth about a destination the Tourism Australia campaign having Chris Hemsworth generated
brand to a wider set of audiences than the regular tourism campaigns revenues of more than $60 million in publicity (Owen, 2015). However,
(Kotler et al., 1993). This trend, visible both in the western and the the hefty fees charged by these ambassadors (if they are not doing it for
eastern countries, have promoted marketing campaigns using celebrity free), may backfire if the audience shows resentment to such

* Corresponding author.`
E-mail addresses: subhadipr@iima.ac.in (S. Roy), wioleta.dryl@ug.edu.pl (W. Dryl), luciana.dearaujo@udp.cl, toluciana.dearaujo@udp.cl (L. de Araujo Gil).

https://doi.org/10.1016/j.tourman.2020.104213
Received 26 September 2019; Received in revised form 21 August 2020; Accepted 26 August 2020
Available online 7 September 2020
0261-5177/© 2020 Elsevier Ltd. All rights reserved.
S. Roy et al. Tourism Management 83 (2021) 104213

endorsement deals. For example, singer Pitbull had been Florida’s global on the total expenditure (Stienmetz et al., 2015). The DAR model has
tourism ambassador for a year starting in August 2015. However, there been empirically validated in subsequent studies (Choe et al., 2017;
was a big controversy over the tourism marketing funding, once he was Stienmetz et al., 2015; Stienmetz and Fesenmaier, 2014) which implies
coaxed into revealing the details of his high paid contract (Peltier, that all the variables i.e. attitude towards the advertisement, attitude
2017). Linking a destination brand with a single person might be risky towards the destination, visiting intention and decision on the
since the destination is relying on the demographic reach of the celebrity expenditure-value balance are important for destination advertising.
and the brand values that he/she is associated with. However, the reach The next subsection provides an overview of literature on celebrity
of a celebrity may be influenced (both positively and negatively) by the endorsements.
choices in his/her personal life (Vhora, 2018). Thus, a celebrity
endorsed destination marketing campaign may go both ways. Despite 2.2. Overview of celebrity endorsements research
the apparent potential of the celebrity endorsement campaigns for
tourism destinations, researchers have not investigated this phenome­ The literature on celebrity endorsements has ample evidence that
non thoroughly (van der Veen and Song, 2010). In addition, the existing proper utilization of a celebrity endorser is effective in promotion of a
investigations have confounding evidence on the effectiveness of ce­ product and getting the right response from the consumers, in terms of
lebrity endorsers in destination branding (Lee et al., 2008; van der Veen, attitudes and intentions (Park and Yim, 2020; Priester and Petty, 2003;
2008). Roy et al., 2012, 2015, 2019) and in terms of market based outcomes
Based upon this premise, and inferences from celebrity endorsement (Agnihotri and Bhattacharya, 2018; Derdenger et al., 2018). The major
research (with a focus on destination marketing), three research ques­ streams of research in celebrity endorsements are: a) the ‘source credi­
tions (RQ) are posed: 1) whether a celebrity endorsement would work bility’ stream; b) the congruence stream; and c) the meaning transfer
better than a generic tourism ad across different countries; 2) If yes to stream.
q1, then how these effects may change depending on the celebrity type The ‘source credibility’ researchers propose the perceived credibility
(local/global); and 3) the destination type (domestic/international). The of the celebrity endorser to be the reason behind the effectiveness of the
exploration to these RQs are conducted using four controlled experi­ endorser on the consumers. Three dimensions of source credibility
ments (in sequence) conducted simultaneously with similar samples in through which an endorser may influence the consumers are Trust­
three countries that are socio-culturally different (India, Brazil and worthiness, Expertise and Attractiveness (Ohanian, 1990; Pornpitakpan,
Poland). The following sections outline the literature review, the back­ 2003). Researchers in celebrity endorsement have observed a positive
ground to each section and the findings and implications of the study effect of each of these dimensions on consumer attitudes and behavior,
before the study is concluded. such as: Kamins (1990) and Priester and Petty (2003) in trustworthiness;
Maddux and Rogers (1980) and Ohanian (1991) in expertise; Silvera and
2. Literature review Austad (2004) and Gilal et al. (2020) in attractiveness.
The ‘congruence’ studies are built on the premise of celebrity-brand
2.1. Destination advertising effectiveness congruence (fit) and explain the effectiveness of the endorsement
through “match” between the celebrity and the brand (Choi and Rifon,
The effectiveness of advertising campaigns is a key issue for desti­ 2012; Kirmani and Shiv, 1998). While the concept of fit was initially
nation managers and thus, in-depth exploration of tourism advertising restricted to celebrity-product ‘fit’, subsequent studies explored multiple
campaigns may contribute to the effectiveness of destination adver­ aspects of congruence in endorsements such as: celebrity-brand per­
tisements (Amar et al., 2017). There are multiple ways to measure the sonality (Mishra et al., 2015) or celebrity-target audience age (Roy et al.,
effectiveness of destination advertising. One of the widely practiced 2015) and celebrity-target market country of origin (Roy et al., 2019)
techniques is ‘‘Conversion study’’ that usually includes a delayed congruence.
response survey questionnaire sent (by email or telephone) to the The third stream of research in celebrity endorsements is the ‘the
audience with a time lag after the advertisement is aired that has meaning transfer model’ that is based on the conceptualization of ce­
question related to their travel behavior and attitudes (Woodside, 2010). lebrity endorsements as a meaning transfer process by McCracken
The conversion models considers various measures such as: Gross Con­ (1989). McCracken (1989) proposed a model where the celebrity was
version Rate (GCR): i.e. the percentage of travelers who requested travel considered as a storehouse of meaning that get transferred to the
information and then traveled to the specific destination (Choe et al., product brand via advertisements and social processes. The consumers
2017); Net Conversion Rate (NCR): GCR adjusted by excluding the subsequently consumed these meanings by possession and consumption
travelers who planned their trip before exposure to tourism advertising of goods and services. Researchers have put relatively less focus on
(Stergiou and Airey, 2003); and Net Influence Rate (NIR): a combined ‘meaning transfer’ compared to the other two streams and have explored
measure that includes both the timing and perceived influence of the celebrity traits as a proxy for meaning (Saldanha et al., 2020; Walker
destination advertising, and related trip expenditures. However, the et al., 1992) and found the same to get transferred to the endorsed
existing measures were criticized by Park et al. (2013) as they do not product.
capture the complete process of advertising response behaviour.
Subsequently, Park et al. (2013) proposed the Destination Adver­ 2.3. Celebrity endorsements in destination marketing
tising Response (DAR) model where they created a four-stage face­
t-based advertising response model. The major stages of the DAR are The literature on the effects of celebrity endorsers in destination
attitude towards advertisement, attitude towards destination, purchase marketing is relatively less than the volume of literature on mainstream
trip facets and total trip expenditure (Stienmetz et al., 2015). The DAR celebrity endorsements. Celebrity endorsement research has been
model posits destination advertising response to be a hierarchical pro­ focused mainly on products and services, while the effects of endorsers
cess that has the stages mentioned before. In stage 1, the potential tourist in the context of destination marketing remained under-researched (van
is exposed to the destination advertising resulting in attitude towards der Veen, 2008). Early researchers in this area (e.g., Holloway and
the advertisement (positive/negative). In stage 2, this attitude towards Robinson, 1995; Pritchard and Morgan, 2001) suggested that creating a
the advertisement leads to individual attitude towards the destination. linkage between a well-known celebrity and a destination could benefit
In stage 3, the exposed audience considers whether or not to visit the the promotion of a destination immensely through free publicity and
advertised destination, and the purchase intentions of various aspects. tourist attention. In addition, a celebrity could foster the development of
Lastly, in stage 4, each travel-related decision is evaluated based on their the destination image using his/her own associations (Gartner, 1994).
contribution to the overall perceived value of the trip leading to decision Aronson et al. (1999) observed that a destination marketing program

2
S. Roy et al. Tourism Management 83 (2021) 104213

with a celebrity endorser might generate a higher degree of tourist cosmopolitanism. A schematic diagram that integrates all the study
receptiveness, but the strength of the receptiveness would depend on the hypotheses is given in Fig. 1. All the mentioned studies were conducted
destination brand-celebrity image linkage. According to Zhang et al. using parallel experiments in three countries with different socio-
(2020), credibility factors such as trustworthiness and expertise of a cultural backgrounds to aid generalizability. The next few sections are
celebrity endorser may favorably influence destination brand love of built in the following manner: the theoretical development to each study
tourists directly and indirectly through para-social interaction. Re­ is presented first, leading to the hypothesis specific to the study,
searchers agree that celebrity-endorsed destination advertising may following which, the methodology and data collection for that particular
help in differentiating a destination from others (Pike et al., 2011) and study is presented. The findings from each study are then used to sup­
may also generate awareness among the target audience (Campelo et al., port/reject the respective hypotheses but also as a base for the next
2011). study.
Researchers exploring the underlying mechanisms that may lead to
the success of celebrity endorsements in destination marketing found the 3. Study 1
celebrity endorser may increase destination credibility which, in turn,
may lead to favorable attitudes towards the destination (Chan, 2010). 3.1. Rationale
Lee et al. (2008) proposed that the consumer’s involvement with the
celebrity would influence the success of the celebrity endorsed desti­ The rationale behind study 1 is drawn from the concepts of source
nation brand. This implies that the higher the involvement of the ce­ credibility (Hovland and Weiss, 1951; McGuire, 1985). The concept of
lebrity with the destination brand (fostered by the celebrity aura or source credibility has been used frequently in endorsement research
image), the higher would be the future behavioral intentions of the (such as Spry et al., 2011; Winterich et al., 2018; Schouten et al., 2020).
target audience (Kim et al., 2012; Yen and Teng, 2015). In addition, Hovland and Weiss (1951) define source credibility as ‘the quality of a
researchers observed that positive affective consumer reactions to source of communication that allows the recipients of the communica­
celebrity-endorsed destination advertisements are more likely to lead to tion to agree (at the attitudinal or behavioral level) to the communica­
a favorable evaluation of the destination (van der Veen and Song, 2014) tion content/message’. A celebrity could be considered as a credible
that may result in favorable attitudes towards the destination (Chan source by his/her attractiveness, trustworthiness and expertise (Oha­
et al., 2018) and destination loyalty (Teng and Chen, 2020). It is worth nian, 1990; 1991). Researchers have observed that a credible endorser
noting that the research on the effectiveness of celebrity endorsements may be able to influence the consumer’s attitude towards the adver­
in tourism have been around the first three stages of the DAR model. tisement, brand evaluations, purchase intentions and recall (Amos et al.,
Quite similar to mainstream endorsement research, research in ce­ 2008; Ohanian, 1991).
lebrity endorsements for destination marketing are not free from criti­ It is expected that a celebrity endorser would bring more credibility
cisms. For example, Lee et al. (2008) did not find a significant effect of as compared to a generic advertisement for a destination, since the
empathic involvement of the consumer with the celebrity on the desti­ advertisement would have an identified source. According to Hussain
nation image. Likewise, van der Veen (2008) argued that the selection of et al. (2020), trustworthiness of a celebrity positively affects advertise­
a destination by the consumer depends on the consumer emotions and a ment and brand credibility. However, the mentioned effects are
celebrity endorser may not be able to change such decisions. van der moderated by the consumers’ ethnicity, which justifies the need for
Veen (2008) also argued for the ‘right’ fit between the celebrity and the international research in this area. Since source credibility has been
destination, but cautioned on the difficulties associated with measuring found to be effective across the globe (such as USA e.g. Amos et al.,
the fit and its effect on the consumers. Glover (2009) built upon this 2008; U.K. e.g. Thwaites et al., 2012; India e.g. Roy et al., 2012; and
critique and proposed a threefold match between ‘destination image, China e.g. Hung et al., 2011), it is argued that a tourism advertisement
celebrity image and the tourist’s self-image’. Glover (2009) argued that with a celebrity endorser would be more effective irrespective of the
this tripartite fit might result in effective destination advertising. destination origin and this effect would hold good for all destinations.
This effectiveness would be manifested in the consumer attitudes toward
2.4. Research gaps and objectives the advertisement, destination and visiting intention. Thus, the first set
of hypotheses of the study are proposed as:
Summarizing the review, it is observed that a) the quantum of
research is less in the context of celebrity endorsements of destination H1a A celebrity endorser will be more effective in generating positive
brands; b) there are certain questions related to the effectiveness of the consumer cognitive and affective attitudes towards the adver­
endorsement that need to be answered; and c) there has been rarely any tisement compared to a generic advertisement.
study that corroborates findings across different countries to enhance H1b A celebrity endorser will be more effective in generating positive
generalizability. Using these gaps as the starting point, the present consumer attitudes towards the destination compared to a
research aims to address three objectives in the present study using a generic advertisement.
sequence of experimental designs. First, theoretical support from source H1c A celebrity endorser will be more effective in generating positive
credibility theory is drawn to conduct an exploration of the notion of destination visiting intentions compared to a generic
whether a celebrity endorsed advertisement would be more effective advertisement.
than a generic tourism ad for the same destination (study 1, corre­
sponding to H1a-c). The following study (study 2, corresponding to H2a- 3.2. Choice of country
c) draw inferences from the congruence theory and social identity theory
to explore whether a domestic or a foreign (to the domestic target The study was conducted in three countries, namely India, Brazil and
audience) celebrity would work better for a domestic tourist destination; Poland. There were multiple reasons behind the selection of these three
and whether this effect would be moderated by the country-specific countries. First, these three countries belong to three different conti­
cultural factors (study 3, corresponding to H3a-c). The final study nents and offer significant diversity in culture. India is a recently glob­
(study 4, corresponding to H4a-c and H5) builds on arguments of con­ alized nation, and a rapidly growing emerging market with a lot of
sumer cosmopolitism and investigates whether the effectiveness of ce­ diversity in terms of religious practices, languages, social systems and
lebrity endorser would depend upon the interaction between the choice wealth structure (Sahoo, 2006; Saran et al., 2007). Brazil globalized
of the celebrity (i.e. whether the celebrity should be local or global) and before India and in-spite of its diversity, has a common theme of
the destination type i.e. whether the destination is domestic or inter­ multi-culturalism running through the country that has resulted from
national and whether the effects would be influenced by consumer internal migration and the effects of mass media (Schneider, 2018).

3
S. Roy et al. Tourism Management 83 (2021) 104213

Fig. 1. The conceptual model.

Poland is a country where the most important cultural factors affecting effect) and the reasons behind the same. Based on the findings, three
the people are the norms and values that dominate the society, the beach destinations were selected in each of the three countries respec­
characteristics of the nation and the accepted norms of behavior (Sto­ tively (as beach destinations were found to be universally popular across
lecka-Makowska, 2016). To further support the differences between the the three countries). The destinations selected were Goa in India, Ilha
countries, inference is drawn from Gert Hofstede (hofstedeinsights.com, Grande in Brazil and Sopot in Poland.
2019) who illustrate that India, Brazil and Poland are very different on
five of the cultural dimensions suggested by Hofstede except power 3.3.2. Celebrity selection
distance. Pre-test 1 was conducted to select the celebrity endorser in each
While these three countries have cultural differences, all three country respectively using free word elicitation. A new sample of 50
countries have a significant amount of celebrity endorsements in the adult respondents in each country (mean age: 32 years; approximately
marketplace (Boeing and Schurhaus, 2014; Datta, 2019; Kowalewska, 40 percent female) was asked to write down the names of five national
2018). In addition, these countries have tourist destinations that are celebrities that they would like to see as the endorser of the respective
comparable in terms of the destination type (such as beaches). Hence, a beach destination in their country. Based on frequency counts and
simultaneous exploration in these countries would allow enough maintaining the similarity of celebrity profession, cricketer Sachin
generalizability of the findings both in terms of common and contrasting Tendulkar from India, and soccer stars Neymar from Brazil and Robert
elements. Lewandowski from Poland were selected for study 1.
Post the celebrity selection; Pre-test 2 (in each country separately)
was conducted to ensure that all three celebrities were at the same level
3.3. Methodology of credibility. This was required as past research on source credibility in
endorsements found differential levels of credibility of the endorser to
Experimental design was used for the study, as the study objectives/ influence consumer choice. Hence, this was done to avoid the possibility
hypotheses were dependent on categorical conditions imposed on the of this confounding at a later stage. A new sample of 100 adult re­
dependent variables and were more fitting to a controlled experimental spondents in each country was selected using judgmental sampling for
set-up. Hence, there was a need to pre-select the destination and the pre-test 2 that belonged to the same population as Pre-test 1 (and the
celebrity endorser so that fictitious ads could be created as stimuli on final study) (mean age: 33 years; 48 percent female). The sample in each
which the consumer attitudes would be tested. Vargas et al. (2017) country was asked to provide credibility ratings for the celebrity selected
suggest careful crafting of stimuli for experimental studies in advertising for that particular country (i.e., Tendulkar, Neymar or Lewandowski). A
as the accuracy and validity of the findings majorly depend on it. Thus, a six-item abridged version of Ohanian’s (1990; 1991) source credibility
series of focus group discussions and quantitative pre-tests were con­ scale was used to measure credibility (5-point semantic differential
ducted to select the stimuli for the studies that are comparable, relevant scales: attractive–unattractive, classy–not classy, expert–not an expert,
to the study objectives and do not create unnecessary confounding later knowledgeable–unknowledgeable, honest–dishonest, and trust­
on. Details of the same for study 1 is given in the next sub-sections: worthy–untrustworthy). All three celebrities were found to have average
credibility ratings above 3. Following this, an ANOVA was run to test the
3.3.1. Choice of destination presence of any inter-country difference. The ANOVA results (F = 0.53;
To select the destination, two Focus Group Discussions (FGD) in each p > 0.1) and subsequent pairwise comparisons did not indicate any
country were conducted. Within each country, one FGD consisted of statistically significant difference in credibility ratings of any celebrity.
student respondents and the other consisted of working professionals. Thus, the choice of the celebrity endorser was justified. To summarize, a
This was done to ensure within-group homogeneity, and at the same 3-step procedure using three different techniques (FGD, Word Elicitation
time, maintain across group heterogeneity concerning demographics Technique, Survey Rating) were used to select a destination with com­
that would give us comprehensive understanding. The gender ratio was mon characteristics across the three countries, and a celebrity in each
approximately 1:1 across the six FGDs. The total number of participants country with similar levels of credibility. Next, the procedure for the
(all six FGDs) was 50 with average group size around 8 across countries creation of the stimuli for the experiments is discussed.
and near to equal group sizes for student/non-student participants. The
participants were asked to discuss the destination that they would like to
visit (that they have not visited earlier, to take care of the memory

4
S. Roy et al. Tourism Management 83 (2021) 104213

3.3.3. Treatments 3.3.5. Data collection


Study 1 used a 2 (Advertisement type: Generic vs. Endorser) x 3 Study 1 (and the other studies) was a controlled experiment con­
(Country: India vs. Brazil vs. Poland) full factorial between-subject ducted among a sample of adult respondents (Please refer to Table 1 for
design. Fictitious print advertisements were used for treatments that sample demographics of all the studies). The use of student respondents
were designed jointly by the researchers and a graphic designer to make in experiments was criticized by Vargas et al. (2017), as they limit
them look like real advertisements. Six treatments were thus oper­ generalizability. Thus, a more general audience was needed to add
ationalized in this manner. All the advertisements had similar layouts robustness to the findings. In this pursuit, judgmental sampling was used
with the destination in the background and the celebrity in the fore­ and the researchers approached relevant respondents with an incentive
ground, closest possible graphics, and similar wording and length of text to participate in the experiment. The screening criteria were two: 1) The
to minimize extraneous variances (Refer to Figs. 2–7 for sample adver­ respondent had not visited the respective destination (in the treatment)
tisements from all studies). The treatment booklet contained an opening in the last ten years, and 2) The respondent was planning to take a
brief followed by the focal advertisement. This was followed by the vacation in the next one year. The researchers themselves supervised the
measures of the dependent variables and the manipulations checks. The sample selection and the conduction of the experiment. The shortlisted
final page contained questions related to respondents’ demographics. participants (in each country) were asked to assemble at the research lab
(in this case, the university) at a specified date and time. Once assem­
3.3.4. Variables and measures bled, they were assigned randomly to one of the two treatment rooms
The dependent variables for Study 1 were the attitude towards the where the researchers/trained assistants of the researchers ran the
advertisement (AD), attitude towards the destination (AB), and visiting experiment. Initially, the respondents were given an oral brief (and in
intention (VI). Based on prior literature, both cognitive and affective print on the first page of the treatment booklet). This was followed by a
components of attitude towards the advertisement were measured. Five print advertisement in the next page that the respondents were asked to
items were used to measure the cognitive component: unpersuasi­ view for a minute. The subsequent pages had the dependent variable
ve–persuasive, uninformative–informative, unbelievable–believable, (DV), manipulation checks, and respondent demographics. When the
ineffective–effective, and unconvincing–convincing (Burton and Lich­ respondents finished, they were thanked, debriefed, given a gift coupon
tenstein, 1988; van der Veen and Song, 2014). Likewise, five items were from a city store and the session was closed. The total sample size in
used to measure the affective component: bad–good, unappea­ study 1 was 217 (approximately 70 respondents per country and 35 per
ling–appealing, unattractive–attractive, unpleasant–pleasant, and unli­ treatment group).
kable–likable (Janiszewski, 1988; van der Veen and Song, 2014).
Attitude towards the destination was measured using a five-item se­
mantic differential scale that was adopted from MacKenzie and Lutz 3.4. Results
(1989) but was used in the context of destination marketing by van der
Veen and Song (2014). The participants were asked to respond to the The manipulation checks did not reveal any significant difference in
question, “How would you describe your overall attitude towards XX the endorser credibility across the countries (F = 0.69, p > 0.1). Thus,
(destination) after watching this ad?”. The responses were recorded on: the manipulations were successful. A multivariate analysis of variance
good–bad, like–dislike, pleasant–unpleasant, positive–negative, and (MANOVA) was first run to identify the effects of the treatments on the
favorable-unfavorable. Similarly, to measure visit intention, re­ dependent variables at an overall level. Following this, individual
spondents were asked: “How likely is it that you will visit XX (destina­ ANOVAs for each of the four dependent variables were conducted to
tion) in the next 12 months?”, followed by the items: assess the effect of treatments and the interactions.
probable–improbable, likely–unlikely, and possible–impossible (Lam The MANOVA results found a significant main effect of the adver­
and Hsu, 2006; van der Veen and Song, 2014). The same source credi­ tisement type (i.e., generic vs. celebrity) (F4,208 = 1152.79, p < 0.01) at
bility measures used in pre-test 2 were used for manipulation check. An the overall level. However, the ‘advertisement type × country interac­
additional question asked “Do you recognize the celebrity?”. Besides tion’ was also found to be significant (F8,216 = 10.77, p < 0.01). To
these, there were questions on respondent demographics. better understand the data pattern, individual ANOVAs and between-
cell contrasts were examined. The ANOVA results are presented

Fig. 2. Generic Ad, Poland.

5
S. Roy et al. Tourism Management 83 (2021) 104213

Fig. 3. Celebrity Ad, Poland.

Fig. 4. Celebrity Ad, Brazil.

separately for each dependent variable. 2.47; tC-G = 18.38; p < 0.01). However, the average difference between
attitudes generated by the generic ad and the celebrity ad was less in
3.4.1. Cognitive attitude toward the ad (CA) Poland compared to India and Brazil (the possible reason behind the
Consistent with the MANOVA findings, the ANOVA results indicated significant interaction effect).
a significant effect of advertisement type (F1,211 = 1547.51, p < 0.01)
and the interaction term (F2,211 = 9.38, p < 0.01) on the cognitive 3.4.2. Affective attitude toward the ad (AA)
attitude. Planned contrasts revealed that the celebrity ad (subscript C) The results indicated a similar effect of advertisement type (F1,211 =
was more effective than the generic ad (subscript G) for all countries 1211.22, p < 0.01) and the interaction (F2,211 = 19.98, p < 0.01) on the
(MCIndia = 4.01; MGIndia = 1.99; tC-G = 25.86; p < 0.01), (MCBrazil = 3.96; affective attitude. Planned contrasts revealed that the celebrity ad was
MGBrazil = 2.02; tC-G = 22.86; p < 0.01) and (MCPoland = 4.03; MGPoland = more effective than the generic ad for all countries (MCIndia = 4.03;

6
S. Roy et al. Tourism Management 83 (2021) 104213

Fig. 5. Celebrity Ad, India.

Fig. 6. Study 2, global celebrity, India.

7
S. Roy et al. Tourism Management 83 (2021) 104213

Fig. 7. Study 4, global celebrity, international destination.

contrasts revealed similar findings as CA and AA (MCIndia = 3.97; MGIndia


Table 1
= 1.97; tC-G = 21.69; p < 0.01), (MCBrazil = 4.02; MGBrazil = 1.94; tC-G =
Sample Demographics (all countries combined).
21.01; p < 0.01) and (MCPoland = 3.91; MGPoland = 2.39; tC-G = 15.35; p
Demographics Study 1 Study 2 Study 3 Study 4 < 0.01).
Gender
Male 113 118 109 218 3.4.4. Visit intention (VI)
Female 104 103 112 196
Even in this case, there was a significant effect of advertisement type
Age
18-25 41 44 43 81 (F1,211 = 708.79, p < 0.01) and the interaction term (F2,211 = 8.71, p <
25-35 44 42 41 85 0.01) on the visit intentions towards the respective destination. Planned
35-45 53 60 59 102 contrasts revealed findings (MCIndia = 4.05; MGIndia = 1.95; tC-G = 17.44;
45-55 46 45 48 91 p < 0.01), (MCBrazil = 4.01; MGBrazil = 1.96; tC-G = 16.43; p < 0.01) and
55 and above 33 30 29 55
(MCPoland = 3.99; MGPoland = 2.53; tC-G = 11.47; p < 0.01) quite similar
Occupation
Private Service 52 51 53 100 to the previous cases.
Government Service 44 42 44 70
Self Employed 62 68 63 131 4. Discussion
Student 33 32 35 67
Homemaker 17 15 12 26
Others 9 13 13 20 The aim of Study 1 was to determine whether consumers would be
Number of domestic leisure trips in a year more influenced by a celebrity endorsed tourism advertisement vis a vis
0 26 22 24 41 a generic advertisement and whether this effect would be varying across
1-3 168 169 172 352 countries. The findings indicated that irrespective of the country, a ce­
3 and above 23 20 24 21
lebrity endorsed advertisement would be more effective than a generic
Number of foreign leisure trips in a year
0 76 69 78 147 tourism advertisement in creating consumer attitudes and affect.
1-3 124 127 120 243 Thereby, support for H1 (a through c) was found and it could be inferred
3 and above 17 25 22 24 that celebrities bring additional value to the advertisements, leading to
Note: All studies had unique sample without any overlap of respondents. more favorable consumer attitudes. In the case of all four dependent
variables, similar effects were observed. However, this brings us to the
MGIndia = 1.99; tC-G = 22.79; p < 0.01), (MCBrazil = 4.09; MGBrazil = 1.98; next question on the type of celebrity to be used. While countries have
tC-G = 22.24; p < 0.01) and (MCPoland = 3.97; MGPoland = 2.59; tC-G = ; p the option to use a local or a global celebrity, which celebrity would be
< 0.01). preferred more by domestic consumers (in case they are exposed to both
communication)? This leads to study 2.
3.4.3. Attitude toward the destination (AD
The results here also showed a significant effect of advertisement
type (F1,211 = 1084.44, p < 0.01) and the interaction term (F2,211 = 9.42,
p < 0.01) on the attitude towards the respective destination. Planned

8
S. Roy et al. Tourism Management 83 (2021) 104213

5. Study 2 country was asked to name their three favorite global sports celebrities.
Based on the frequency counts in each country and commonality across
5.1. Rationale countries, soccer star Cristiano Ronaldo was selected as the global ce­
lebrity. The generic ad in each country was replaced with the global
A significant volume of research explored the concept of celebrity- celebrity (e.g., Fig. 6). The main experiment was conducted in the same
brand congruence or fit and its influence on consumer attitudes. The fashion in each country as in study 1 but with a new sample (sample size:
congruence hypothesis originated in psychology, where Mandler (1982) 221, approximately 35 respondents per cell).
defined congruence as ‘a structural correspondence between two en­
tities.’ Researchers in marketing and consumer behavior afterward have 5.3. Results
considered ‘congruence’ as a ‘match’ between two entities or simply a
measure of two entities going together. Several domains in marketing The first test was for any possible difference in the endorser credi­
research have explored the effects of congruence between two or more bility within the country. Paired t-tests comparing mean credibility of
entities and its effect on the consumers such as, product design (Wänke the global to the local celebrity did not reveal any significant difference
et al., 2007); brand extensions (Kim and John, 2008); retailing (Span­ (India: tL-G = 0.37, p > 0.1; Brazil: tL-G = 0.32, p > 0.1; Poland: tL-G =
genberg et al., 2006) and sponsorships (Weeks et al., 2008). Celebrity 0.41, p > 0.1)1. The ANOVA results also did not show any significant
endorsement researchers are one of the highest users of this theory. difference in any celebrity across the countries. Thus, the manipulations
Notable examples include Kamins, (1990), Mittelstaedt et al. (2000), Till were successful. Similar to study 1, a multivariate analysis of variance
et al. (2008), and Roy and Bagdare (2015). Most of these studies found (MANOVA) was run first, followed by individual ANOVAs for each of the
perceived congruence between a celebrity and a brand to generate more four dependent variables.
favorable consumer attitudes than an incongruent pair. However, re­ The MANOVA results indicated a significant main effect of the ce­
searchers have subsequently investigated the congruence between ce­ lebrity type (i.e. local vs. global) (F5,211 = 10.18, p < 0.01) at the overall
lebrity and the consumer at multiple levels of attributes. For example, level. However, a significant effect of the country (F10,422 = 9.76, p <
Chang (2008) found consumers to express stronger perceptions towards 0.01) and the ‘celebrity type x country’ interaction (F10,422 = 156.67, p
the advertisement when there was a match between the chronological < 0.01) was also observed. To better understand the patterns in the
age of the endorser and the consumer. However, Carlson et al. (2020) results, individual ANOVAs and between-cell contrasts were examined.
found identification of the consumer with an endorser to be more
important than the enodorser-brand ‘fit’. Consumers who strongly 5.3.1. Cognitive attitude toward the ad
identify with the endorser respond positively to the recommendation, Consistent with the MANOVA findings, the ANOVA results indicated
even if the match between the endorser and the brand is poor. In a significant effect of celebrity type (F1,215 = 6.74, p < 0.01), country
addition, identification with the endorser is consistently linked to pur­ (F2,215 = 18.67, p < 0.01) and the interaction term (F2,215 = 558.25, p <
chase intentions over multiple time points. 0.01) on the cognitive attitude. Planned contrasts (Fig. 8a) revealed that
Similarly, Roy et al. (2015) found women consumers to have stron­ the effect of the celebrity was varying with country (MLIndia = 2.07;
ger attitudes and purchase intentions towards a celebrity endorser who MGIndia = 4.05; tL-G = 25.35; p < 0.01), (MLBrazil = 2.01; MGBrazil = 4.07;
matched their chronological age. Continuing this line of thought, Roy tL-G = 22.34; p < 0.01) and (MLPoland = 2.88; MGPoland = 2.57; tL-G =
et al. (2019) extended congruence to celebrity-consumer fit based on the 3.65; p < 0.05). The global celebrity was more effective in India while
country of origin. Uy and Wu (2020) indicated the importance of the the local celebrity was more effective in Brazil. On the other hand, both
congruence of the celebrity’s and consumer’s country of origin, but also celebrities were close to each other in creating consumer attitudes at the
pointed to the fact that the effect of localized (vs. standardized) celebrity cognitive plane in Poland.
endorsements depended greatly on the patriotism of an individual. It
could be argued that a match or congruence between the country of 5.3.2. Affective Attitude toward the ad
origin of the celebrity and the consumer would heighten the consumers’ Similar to the cognitive attitude, the ANOVA results for affective
attitudes and evaluations of the destination. Thus, it is argued that a attitude towards the ad indicated a significant effect of celebrity type
local celebrity would be more preferred for a destination within the (F1,215 = 9.97, p < 0.01), country (F2,215 = 9.46, p < 0.01) and the
country for the domestic consumers (and thereby oppose the use of interaction term (F2,215 = 434.01, p < 0.01). Comparable to cognitive
foreign celebrities). Thus, the next set of hypotheses of the study are attitude, planned contrasts (Fig. 8b) revealed that the effect of the ce­
proposed as: lebrity was varying with country (MLIndia = 1.99; MGIndia = 3.98; tL-G =
21.09; p < 0.01), (MLBrazil = 4.03; MGBrazil = 1.95; tL-G = 19.57; p < 0.01)
H2a A local celebrity endorser would be more effective in generating and (MLPoland = 2.95; MGPoland = 2.49; tL-G = 0.82; p > 0.10). Thus, even
positive consumer cognitive and affective attitudes towards the here the global celebrity was more effective in India while the local
destination advertisement compared to a global celebrity. celebrity was more effective in Brazil and there was no significant dif­
H2b A local celebrity endorser would be more effective in generating ference in the consumer attitudes in Poland (for the local and global
positive consumer attitudes towards the destination compared to celebrities).
a global celebrity.
H2c A local celebrity endorser would be more effective in generating 5.3.3. Attitude toward the destination
positive destination visiting intentions compared to a global Similar to attitudes towards the ad, a significant effect of celebrity
celebrity. type (F1,215 = 25.92.74, p < 0.01), country (F2,215 = 13.42, p < 0.01)
and the interaction term (F2,215 = 533.36, p < 0.01) was observed on the
5.2. The experiment attitudes towards the destination. Planned contrasts (Fig. 8c) revealed
that the effect of the celebrity was varying with country (MLIndia = 2.02;
Study 2 used the same dependent variables as used in study 1. MGIndia = 3.95; tL-G = 24.71; p < 0.01), (MLBrazil = 4.03; MGBrazil = 2.01;
However, the independent variable in this study was the celebrity type tL-G = 20.41; p < 0.01) and (MLPoland = 3.07; MGPoland = 2.38; tL-G =
(manipulated as local/global). The other independent variables: the 8.13; p < 0.01), and the results were very similar to those of the attitudes
country and the destination were retained from study 1. Thereby, a 2 towards the ad.
(celebrity: local vs. global) X 3 (country: India/Brazil/Poland) full
factorial design was used. Pre-test 3 was conducted with the objective of
selecting the global celebrity. A sample of 100 respondents in each 1
L = Local Celebrity; G = Global Celebrity.

9
S. Roy et al. Tourism Management 83 (2021) 104213

Fig. 8. a): Study 2, DV: CA; IV: Celebrity Type, Country. b): Study 2, DV: AA; IV: Celebrity Type, Country. c): Study 2, DV: AD; IV: Celebrity Type, Country. d): Study
2, DV: VI; IV: Celebrity Type, Country.

5.3.4. Visit intention theory (Tajfel and John, 1986). Scholars in social-identity theory opine
In consonance to the attitudes, a significant effect of celebrity type that human beings respond favorably to those stimuli that are consistent
(F1,215 = 8.88, p < 0.01), country (F2,215 = 7.72, p < 0.01) and the with their accessible identity (i.e., the identity whose mental represen­
interaction term (F2,215 = 265.74, p < 0.01) was observed on the visit tations are salient) (Reicher et al., 2012; Tajfel and Turner, 1986). For
intentions. Planned contrasts (Fig. 8d) revealed findings (MLIndia = 2.09; example, researchers in consumer ethnicity found consumers high on
MGIndia = 3.97; tL-G = 16.49; p < 0.01), (MLBrazil = 3.93; MGBrazil = 2.08; their ethnic identity (or having it accessible) to respond more favorably
tL-G = 15.36; p < 0.01) and (MLPoland = 3.06; MGPoland = 2.41; tL-G = to advertisements where the spokesperson’s ethnic identity matches
5.40; p < 0.01) very similar to the other DVs. with the consumers’ ethnic identity (Deshpandé and Stayman, 1994;
Forehand et al., 2002). In similar thoughts, researchers found brand
associations that are consistent with the cultural identity of the target
5.4. Discussion
consumers to create more favorable consumer attitude towards the same
brands (Aaker et al., 2001). Wheeler et al. (2005) explain this phe­
To summarize, the global celebrity was more effective in India while
nomenon using consumer psychology. They opine that consumers tend
the local celebrity was more effective in Brazil at an overall level. On the
to preserve positive self-views, and thereby, identity-consistent infor­
other hand, both celebrities were close to each other on their effec­
mation is evaluated as more relevant for processing objectives such as
tiveness in Poland. Thereby, H2 (a through c) was not supported. This
brand attitudes and hence is given greater valence than
implies that the choice of a celebrity for destination marketing may not
identity-inconsistent information.
work universally for all countries. Given that there was no significant
In this context, Arnett (2002) proposed the concept of local and
difference in the celebrity credibility across the celebrity type (local vs.
global identities while discussing the psychological consequences of
global) and countries, it could be argued that there was some consumer
globalization on consumers. Arnett (2002) proposed that consumers
level variable that was instrumental behind such findings. Kim and
tend to have both local and global identities with different valence
Richardson (2003) called for research that investigates the underlying
depending on their environment. He conceptualized local identity as a
mechanism that influences the consumers in the context of destination
mental representation in which individuals ‘had faith in and respect for
marketing. Hence, study 3 was conducted to explore the possible causes
local traditions and customs, recognized the uniqueness of local com­
leading to such findings.
munities and were interested in local events’. On the other hand, global
identity consisted of mental representations, where the individuals
6. Study 3 believed in the beneficial effects of globalization, recognized more
commonalities (with the global community) rather than differences, and
6.1. Rationale were interested in global culture and events. In a world that is getting
increasingly globalized, a consumer’s global and local identity may be
To explain the cross-country differences in the effectiveness of the important for his/her product decisions. Consumers oriented with local
local vs. global celebrity, inferences may be drawn from social identity

10
S. Roy et al. Tourism Management 83 (2021) 104213

identity would feel one with their local community and identify with Table 2
local ways of life. On the contrary, consumers high on global identity PROCESS results: Study 3.
would feel themselves to be a part of the global community and identify DV IV β (S.E.) t value R2
with global lifestyles (Tu et al., 2012). Interestingly, this has been a
Cognitive Attitude (CA) Celebrity − 14.45 − 18.59** 0.83
scarcely researched area in celebrity endorsement. (0.78)
Based on the premise, consumers high on global identity should be Country − 7.37 − 14.27**
more affected by a global celebrity, whereas consumers high on local (0.52)
identity should be more affected by a local celebrity. This might explain LGI − 13.67 − 20.59**
(0.66)
the findings from study 2. Thereby, the next set of hypotheses is pro­ Celebrity X 4.68 (0.32) 14.49**
posed as: Country
Celebrity X LGI 8.84 (0.41) 21.47**
H3a A consumer’s local/global identity would moderate the effect of Country X LGI 4.58 (0.29) 16.05**
Celebrity X − 2.96 − 16.57**
the celebrity endorser (local/global) on the cognitive and affec­
Country X LGI (0.18)
tive attitude towards the destination advertisement. Affective Attitude (AA) Celebrity − 12.10 − 16.56** 0.81
H3b A consumer’s local/global identity would moderate the effect of (0.78)
the celebrity endorser (local/global) on attitudes towards the Country − 6.58 − 12.61**
destination. (0.52)
LGI − 12.38 − 18.47**
H3c A consumer’s local/global identity would moderate the effect of (0.67)
the celebrity endorser (local/global) on the visiting intentions. Celebrity X 4.24 (0.33) 13.02**
Country
6.2. The experiment Celebrity X LGI 8.13 (0.42) 19.57**
Country X LGI 4.20 (0.29) 14.55**
Celebrity X − 2.78 − 15.40**
Study 3 used the same treatments with the same independent and Country X LGI (0.18)
dependent variables used in study 2. The only added variable was a Attitude Towards Celebrity − 14.15 − 17.47** 0.80
measure of local-global identity (LGI), adapted from the short (8 items) Destination (AD) (0.81)
local-global identity scale by Tu et al. (2012). Thereby, a 2 (celebrity: Country − 7.13 − 13.24**
(0.54)
local vs. global) X 3 (country: India/Brazil/Poland) full factorial design LGI − 13.09 − 18.92**
with moderator LGI was applied in study 3. The main experiment was (0.69)
conducted in the same way in each country as done in studies 1 and 2 Celebrity X 4.63 (0.34) 13.77**
(with a new set of respondents). The total sample size in study 3 was 220 Country
Celebrity X LGI 8.62 (0.43 20.09**
(with approximately 35 respondents per cell).
Country X LGI 4.39 (0.30) 14.74**
Celebrity X − 2.92 − 15.68**
6.3. Results Country X LGI (0.19)
Visit Intention (VI) Celebrity − 13.51 − 13.24** 0.73
Since the moderating effect of LGI on consumer attitudes was of in­ (1.02)
Country − 6.68 − 9.86**
terest, Process (Hayes, 2013) was used instead of regular MANOVA in (0.68)
study 3. Thus, independent linear regression models for all four DVs LGI − 12.66 − 14.53**
using PROCESS Model 3 (Hayes, 2013) was run. Both country and the (0.87)
celebrity type were included as independent variables, and LGI was Celebrity X 4.40 (0.42) 10.38**
Country
included as the moderator. All two-way interactions and the three-way
Celebrity X LGI 8.43 (0.54) 15.60**
(country X celebrity X LGI) interaction were also included (Table 2). Country X LGI 4.30 (0.37) 11.48**
Similar results were observed for all the DVs. Thus, the results for visit Celebrity X − 2.87 − 12.20**
intentions (VI) are reported here (the others are reported in Table 2). A Country X LGI (0.23)
reasonably high R square was observed (R2 = 0.73; F7,213 = 81.21; p < Note: LGI = Local/Global Identity; ** implies p < 0.01.
0.01) that implied that the regression results were acceptable. The main
effect of the celebrity (β = − 13.51; p < 0.01) and the country (β =
with locally oriented consumers. Thereby, H3 (a through c) was sup­
− 6.68; p < 0.01) were significant on VI, as was the main effect of LGI (β
ported. However, one question remained. What would be the choice of
= − 12.66; p < 0.01). However, the negative effects signified the pos­
the celebrity endorser for the consumers in a country for an international
sibility of an interaction. The two-way interactions were all positive and
tourist destination? There are instances of endorsers who have the same
significant, i.e. Celebrity X Country (β = 4.39; p < 0.01); Celebrity X LGI
country of origin as the tourist destination (e.g. Chris Hemsworth for
(β = 8.43; p < 0.01); Country X LGI (β = 4.3; p < 0.01). Finally, a
Australia) and same country of origin as the target county (e.g. Ranveer
negative three-way interaction effect Celebrity X Country X LGI (β =
Singh for Switzerland). Thus, it warrants an investigation of which type
− 2.86; p < 0.01) was noted, which indicated that some of the two-way
would be more effective and whether there would be any consumer level
interactions were effectively suppressed depending on the celebrity type
factors that would influence this effect. This question leads to the final
or the country (to be specific, individual characteristics of the re­
study.
spondents belonging to that country). The results were replicated for all
the dependent variables.
7. Study 4

6.4. Discussion
7.1. Rationale

The findings from study 3 indicated that there is no ‘one size fits all’
The arguments in study 4 stem from the congruence theory. Ac­
strategy. In other words, the choice of a celebrity endorser for a tourist
cording to the congruence theory, a fit between the endorser and the
destination may not be universally applicable and would depend on the
product would increase consumer evaluations (Choi and Rifon, 2012;
consumer levels factors in a particular country. A global celebrity would
Mishra et al., 2015; Roy and Bagdare, 2015; Roy et al., 2015). Thus, a
be more effective in a country where consumers are more globally ori­
celebrity congruent with the destination (in terms of belonging to the
ented, whereas a local celebrity would be more preferred in a country

11
S. Roy et al. Tourism Management 83 (2021) 104213

same country) would generate more favorable consumer evaluations. H4c A celebrity congruent with the destination (in terms of country of
This is named as C-D congruence. However, if one considers congruence origin), will generate more favorable destination visiting in­
between the endorser and the target audience (as in Roy et al., 2015), it tentions as compared to an incongruent celebrity.
would suggest that a celebrity congruent with the target audience (in H5 The effect mentioned in H4 (a through c) will be moderated by
terms of belonging to the same country) would be more effective. This is consumer cosmopolitanism.
named as C–C congruence. This creates a paradox. C-D congruence
suggests the use of a local celebrity for a local destination and a global 7.2. The experiment
celebrity for an international destination to the same target audience.
On the other hand, C–C congruence suggests the use of a local celebrity Study 4 used the same dependent variables used in the previous
irrespective of the destination (again to the same audience). studies (i.e., cognitive and affective attitudes; attitudes toward desti­
Given this paradox, the construct of consumer cosmopolitanism is nation and visit intention). The new independent variable was the
introduced. The word cosmopolitanism is derived from the Greek destination type (manipulated as Domestic vs. International). The
‘kosmopolitês’ that means ‘citizen of the world’ (Turner, 2002). In the additional measure in study 4 was cosmopolitanism (COS), adapted
context of marketing, Cleveland et al. (2014) defined cosmopolitanism from the short 5-item cosmopolitanism scale of Cleveland et al. (2014)
as “a learned disposition: a general orientation reflecting a set of values, that was found valid in multiple cultures [sample items: I like to observe
opinions, and competencies held by certain individuals; specifically, a people of other cultures, to see what I can learn from them; I enjoy
genuine, humanitarian appreciation for, desire to learn from and ability exchanging ideas with people from other cultures and countries;
to engage with, peoples of different cultures” (p. 269). Consumer measured on a Likert scale (1 strongly agree … …..5 strongly disagree)].
cosmopolitanism is becoming increasingly significant with rapid glob­ Thereby, study 4 had a 2 (celebrity: local vs. global) X 3 (country:
alization and is being used for segmenting consumers within nations and India/Brazil/Poland) X 2 (destination: domestic vs. international) full
at a global level (Cleveland et al., 2009, 2014; Nijssen and Douglas, factorial design with moderator cosmopolitanism.
2008). Pre-test 4 was conducted to select an international destination. A
Highly cosmopolitan consumers would be more open towards other sample of 80 respondents from each country was asked to write their
people and cultures and would be interested to learn from them (Levy three favorite international destinations (country) that they would like
and Sznaider, 2007). They would also appreciate the cross-cultural to visit. Based on frequency counts and commonality across the three
differences and treat them as a variety (Featherstone, 2002). Highly countries, the UK was selected as the international destination. Next,
cosmopolitan individuals are more willing to engage personally with pre-test 5 was conducted to select the global celebrity for the interna­
other cultures with the objective understanding nuances of other cul­ tional destination. Thus, a fresh sample of 75 respondents from each
tures (Nijssen and Douglas, 2008). Hence, they “are positively disposed country was asked to write the name of three UK born celebrities that
towards consuming products from different countries” (Riefler et al., they would like to see as the ambassador for UK tourism. Based on
2012, p. 287). Cosmopolitan consumers display a favorable stance to­ frequency counts and commonality of the profession (with the other
ward products and services from different countries (Riefler et al., celebrities), selected soccer star David Beckham was selected as the in­
2012). Since they are more oriented towards the global consumer cul­ ternational celebrity. Bournemouth was selected as the UK tourist
ture, they continuously indulge in products, places, and experiences destination for the treatment. Bournemouth is a beach destination and
originating from foreign cultures (Riefler et al., 2012). Thereby, they has been rated as one of the best beach destinations in the UK and also at
respond positively to positioning strategies that highlight global con­ par with the other destinations. David Beckham was shown endorsing
sumer culture (Alden et al., 1999; 2006) and more likely to consume Bournemouth for the treatment advertisement for the international
products originating from foreign places and culture (Alden et al., 1999; destination (e.g. Fig. 7) while he was replaced by the local celebrity
Cleveland et al., 2009). The same may be applicable for tourists as (concerning the country) for the other treatment group.
tourism is highly related to foreign places, people, culture and traditions The main experiment was conducted using the same procedure in
(Hannerz, 1990). each country as done in the previous studies (with a different sample).
Now, congruence theory and cosmopolitanism are brought together The total sample size in study 4 was 414 (with approximately 35 re­
to build the case for study 4. C-D congruence suggests the use of a ce­ spondents per cell). In this case, first MANOVA and ANOVA tests were
lebrity matching with the destination (country of origin). C–C congru­ run to identify the attitudinal patterns and reveal country-specific
ence suggests the use of a local celebrity irrespective of the destination. findings (if any). Subsequently, PROCESS model 3 (Hayes, 2013) was
It could be argued that for the more cosmopolitan consumer, using a applied to identify the strength and direction (if any) of the moderation
foreign celebrity endorser may show a greater connection with global effects.
consumer culture, and consequently generate favorable consumer atti­
tudes (e.g. Roy et al., 2019; Steenkamp et al., 2003). Thus, high 7.3. Results
cosmopolitanism may override C–C congruence and even C-D congru­
ence for the case where the celebrity is local. For the less cosmopoli­ First, tests for any possible difference in the endorser credibility
tanism consumer, C–C congruence would surely hold, while C-D within country was conducted. Paired t tests comparing mean credibility
congruence may suffer for the global celebrity. Thereby, the congruence of the global to the local celebrity did not reveal any significant differ­
theory in tourism advertising would hold but would be influenced by ence (India: tL-G = 0.55, p > 0.1; Brazil: tL-G = 0.63, p > 0.1; Poland: tL-G
consumer cosmopolitanism. Summarizing the argument, the next set of = 0.92, p > 0.1) similar to study 2 and 3.
hypotheses are proposed: The MANOVA results indicated a significant effect of the country
(F8,798 = 23.94, p < 0.01), the celebrity type (F4,399 = 579.85, p < 0.01)
H4a A celebrity congruent with the destination (in terms of country of and the destination type (F4,399 = 89.07, p < 0.01). A significant effect of
origin), will generate more favorable cognitive and affective all two-way interactions, i.e. Country X Celebrity (F8,798 = 152.4, p <
attitude towards the advertisement as compared to an incon­ 0.01); Country X Destination (F8,798 = 70.33, p < 0.01); and Celebrity X
gruent celebrity. Destination (F4,399 = 817.84, p < 0.01) was observed. A significant
H4b A celebrity congruent with the destination (in terms of country of three-way interaction of Country X Celebrity X Destination (F8,798 =
origin), will generate more favorable attitudes towards the 153.23, p < 0.01) was also observed. Next, the individual dependent
destination as compared to an incongruent celebrity. variables were investigated (to avoid repetitiveness and save space, the
interaction plots only for cognitive attitude towards the ad and visit
intentions are provided).

12
S. Roy et al. Tourism Management 83 (2021) 104213

7.3.1. Cognitive attitude Towards the ad < 0.01). To better understand the three-way interaction, mean com­
Consistent with the MANOVA findings, the ANOVA results indicated parisons taking one country at a time was conducted.
a significant effect of the country (F2,402 = 29.67, p < 0.01), the celebrity Results were very similar to that of cognitive attitudes. In India, the
type (F1,402 = 699.14, p < 0.01) and the destination type (F1,402 = global celebrity was preferred over the local celebrity for both the des­
122.08, p < 0.01). All the two way interactions were also found sig­ tinations (domestic: MLDomestic = 1.96; MGDomestic = 3.98; tL-G = 23.81; p
nificant, i.e. Country X Celebrity (F2,402 = 338.16, p < 0.01); Country X < 0.01; international: MLInternational = 2.09; MGInternational = 3.94; tL-G =
Destination (F2,402 = 98.16, p < 0.01); and Celebrity X Destination 21.80; p < 0.01). For Brazil, the local celebrity was more effective for
(F1,402 = 861.28, p < 0.01) as was the three-way interaction of Country the local destination (MLDomestic = 4.08; MGDomestic = 2.05; tL-G = 22.02;
X Celebrity X Destination (F2,402 = 259.09, p < 0.01). To better un­ p < 0.01) and the global celebrity for the international destination
derstand the three-way interaction, mean comparisons taking one (MLInternational = 1.96; MGInternational = 4.11; tL-G = 25.34; p < 0.01). For
country at a time was conducted (Fig. 9a). Poland, the global celebrity was more preferred for the international
At first, the case of India is considered. Both for the domestic and destination (MLInternational = 2.51; MGInternational = 4.02; tL-G = 17.80; p <
international destination, a global celebrity is preferred over the local 0.01) while the local celebrity was found to be marginally more effective
celebrity (domestic: MLDomestic = 2.00; MGDomestic = 4.01; tL-G = 23.69; p for a domestic destination (MLDomestic = 2.56; MGDomestic = 2.05; tL-G =
< 0.01; international: MLInternational = 2.01; MGInternational = 4.10; tL-G = 6.01; p < 0.01).
24.63; p < 0.01). Next, the case of Brazil is considered. Here, the local
celebrity was found to be more effective for the local destination 7.3.3. Attitude towards the destination
(MLDomestic = 3.89; MGDomestic = 2.06; tL-G = 21.57; p < 0.01) and the The ANOVA results for attitude towards the destination were similar
global celebrity to be more effective for the international destination to that of the consumer attitudes. A significant effect of the country
(MLInternational = 2.03; MGInternational = 3.98; tL-G = 22.98; p < 0.01). The (F2,402 = 24.38, p < 0.01), the celebrity type (F1,402 = 576.65, p < 0.01)
findings from Poland was different from both India and Brazil. The and the destination type (F1,402 = 94.88, p < 0.01) was observed. All the
global celebrity was found to be more effective for the international two way interactions were found significant, i.e. Country X Celebrity
destination (MLInternational = 2.52; MGInternational = 3.98; tL-G = 17.21; p < (F2,402 = 286.83, p < 0.01); Country X Destination (F2,402 = 106.83, p <
0.01) while the local celebrity was found to be marginally more effective 0.01); and Celebrity X Destination (F1,402 = 896.88, p < 0.01) as was the
than the global celebrity for a domestic destination (MLDomestic = 2.48; three-way interaction of Country X Celebrity X Destination (F2,402 =
MGDomestic = 2.01; tL-G = 6.02; p < 0.01). 291.26, p < 0.01). To better understand the three-way interaction, mean
comparisons taking one country at a time was performed.
7.3.2. Affective attitude Towards the ad Results were very similar to that of attitudes. In India, the global
The ANOVA results for affective attitude were similar to that of the celebrity was preferred over the local celebrity for both the destinations
cognitive attitudes. A significant effect of the country (F2,402 = 21.66, p (domestic: MLDomestic = 1.99; MGDomestic = 3.95; tL-G = 23.10; p < 0.01;
< 0.01), the celebrity type (F1,402 = 563.21, p < 0.01) and the desti­ international: MLInternational = 2.01; MGInternational = 3.98; tL-G = 23.22; p
nation type (F1,402 = 85.63, p < 0.01) was observed. All the two way < 0.01). For Brazil, the local celebrity was more effective for the do­
interactions were also found significant, i.e. Country X Celebrity (F2,402 mestic destination (MLDomestic = 4.11; MGDomestic = 2.06; tL-G = 24.16; p
= 259.07, p < 0.01); Country X Destination (F2,402 = 84.26, p < 0.01); < 0.01) and the global celebrity for the international destination
and Celebrity X Destination (F1,402 = 817.90, p < 0.01) as was the three- (MLInternational = 1.99; MGInternational = 4.07; tL-G = 24.51; p < 0.01). For
way interaction of Country X Celebrity X Destination (F2,402 = 310.83, p Poland, the global celebrity was more preferred for the international

Fig. 9. a. Study 4. DV: CA; IV: Destination and Celebrity, for India, Brazil, Poland. b). Study 4. DV: VI; IV: Destination and Celebrity, for India, Brazil, Poland.

13
S. Roy et al. Tourism Management 83 (2021) 104213

destination (MLInternational = 2.53; MGInternational = 4.06; tL-G = 18.03; p < Table 3


0.01) while the local celebrity was found to be marginally more effective PROCESS results: Study 4.
for a domestic destination (MLDomestic = 2.52; MGDomestic = 1.99; tL-G = DV IV β (S.E.) t value R2
6.25; p < 0.01).
Cognitive Attitude (CA) Celebrity − 5.73 − 9.90** 0.63
(0.58)
7.3.4. Visit intention Destination − 3.74 − 5.84**
The ANOVA results for visit intentions were similar to that of the (0.59)
consumer attitudes. There was a significant effect of the country (F2,402 COS − 1.08 − 4.11**
(0.26)
= 24.53, p < 0.01), the celebrity type (F1,402 = 375.12, p < 0.01) and the Celebrity X 3.38 9.02**
destination type (F1,402 = 40.60, p < 0.01). All the two way interactions Destination (0.37)
were significant, i.e. Country X Celebrity (F2,402 = 146.67, p < 0.01); Celebrity X COS 1.16 6.99**
Country X Destination (F2,402 = 71.25, p < 0.01); and Celebrity X (0.17)
Destination X COS 0.32 1.87˟
Destination (F1,402 = 548.37, p < 0.01) as was the three-way interaction
(0.17)
of Country X Celebrity X Destination (F2,402 = 172.18, p < 0.01). Next, Celebrity X − 0.46 − 4.28**
mean comparisons taking one country at a time was performed to un­ Destination X COS (0.11)
derstand three-way interactions (Fig. 9b). Affective Attitude (AA) Celebrity − 6.10 − 9.85** 0.59
Results were very similar to that of attitudes. In India, the global (0.78)
Destination − 3.97 − 6.25**
celebrity was preferred over the local celebrity irrespective of the (0.52)
destination (domestic: MLDomestic = 2.02; MGDomestic = 3.91; tL-G = 17.78; COS − 1.19 − 4.23**
p < 0.01; international: MLInternational = 2.08; MGInternational = 3.97; tL-G (0.67)
= 19.09; p < 0.01). In Brazil, the local celebrity was more effective for Celebrity X 3.71 9.26**
Destination (0.33)
the local destination (MLDomestic = 4.08; MGDomestic = 2.20; tL-G = 16.62;
Celebrity X COS 1.23 6.90**
p < 0.01) and the global celebrity for the international destination (0.42)
(MLInternational = 1.85; MGInternational = 3.99; tL-G = 18.92; p < 0.01). For Destination X COS 0.43 2.32*
Poland, the global celebrity was more preferred for the international (0.29)
destination (MLInternational = 2.39; MGInternational = 3.99; tL-G = 16.16; p < Celebrity X − 0.54 − 4.67**
Destination X COS (0.18)
0.01) while the local celebrity was found to be marginally more effective
Attitude Towards Celebrity − 5.97 − 9.65** 0.60
for a domestic destination (MLDomestic = 2.47; MGDomestic = 1.92; tL-G = Destination (AD) (0.62)
5.17; p < 0.01). Destination − 3.72 − 5.86**
Thereby H4 (a through c) was supported for Brazil and to some (0.64)
COS − 1.08 − 3.84**
extent for Poland but not for India. The findings indicated the presence
(0.28)
of a moderation effect of cosmopolitism. Hence, the presence (if any) the Celebrity X 3.59 8.96**
effect of cosmopolitism using PROCESS model 3 (Hayes, 2013) was Destination (0.40)
explored. Celebrity X COS 1.16 6.54**
(0.18)
Destination X COS 0.33 1.80˟
7.3.5. Moderation effects
(0.18)
Since the country-wise were already analyzed, country was not Celebrity X − 0.48 − 4.20**
modeled as an independent variable in the next set of analyses. As in Destination X COS (0.12)
study 3, independent linear regression models for all four DVs using Visit Intention (VI) Celebrity − 5.91 − 8.97** 0.56
(0.66)
PROCESS Model 3 (Hayes, 2013) was run. In this case, the celebrity type
Destination − 3.97 − 5.87**
(local vs. global) and the destination type (domestic vs. international) (0.68)
were modeled as independent variables, and COS as the moderator. All COS − 1.03 − 3.45**
two-way interactions and the three-way (celebrity X destination X COS) (0.30)
interaction were also included (Table 3). Results were similar for all the Celebrity X 3.71 8.72**
Destination (0.43)
DVs except affective attitude towards the advertisement (AA). Thus, it is
Celebrity X COS 1.16 6.11**
reported. The R square for affective attitude was 0.59 (F7,406 = 83.91; p (0.19)
< 0.01). The main effect of the celebrity (β = − 6.10; p < 0.01) and the Destination X COS 0.39 2.01˟
destination (β = − 3.97; p < 0.01) were significant on AA, as was that of (0.20)
Celebrity X − 0.53 − 4.29**
COS (β = − 1.19; p < 0.01). All two-way interactions, i.e. Celebrity X
Destination X COS (0.12)
Destination (β = 3.71; p < 0.01); Celebrity X COS (β = 1.23; p < 0.01)
and Destination X COS (β = 0.43; p < 0.05) were significant. However, Note: Coefficients in bold are non-significant; COS = Cosmopolitanism; ** im­
the Destination × COS interaction was only marginally significant. In all plies p < 0.01; * implies p < 0.05; ˟ implies non-significant.
other cases, the same interaction was non-significant (Table 3). A sig­
nificant negative three-way interaction effect of Celebrity X Destination replicated for cognitive attitudes towards the ad and the visit intentions.
X COS (β = − 0.54; p < 0.01) was observed, that supported the Therefore, H5 was supported.
cross-country differences.
Next, the results for attitude towards the destination (AD) is reported 7.4. Discussion
(rest given in Table 3). The overall regression yielded a R square value of
0.6 (F7,406 = 86.35; p < 0.01). The main effect of the celebrity (β = The findings from study 4 reiterated that a single strategy of celebrity
− 5.97; p < 0.01), the destination (β = − 3.72; p < 0.01) and COS (β = endorsements for tourist destinations might not work universally. While
− 1.08; p < 0.01) were significant on AD. The two-way interactions consumers in some nations may prefer a celebrity endorser-destination
Celebrity X Destination (β = 3.59; p < 0.01) and Celebrity X COS (β = match in terms of country of origin (such as Brazil), others may be
1.16; p < 0.01) were significant while that of Destination X COS (β = only affected by the “foreignness” of the celebrity endorser and let it
0.33; p > 0.10) was not found significant. The three-way interaction outweigh C-D congruence (such as India). For some countries, these
Celebrity X Destination X COS (β = − 0.48; p < 0.01) was negative and effects may hold good in certain cases but not for all (such as Poland,
significant, supporting cross-country differences. These results were where the difference between the effectiveness of the local and the

14
S. Roy et al. Tourism Management 83 (2021) 104213

global celebrity for the domestic destination was marginal). Thus, H4 favorable consumer attitudes and behavioral intentions compared to a
was partially supported. The rationale behind such behavior may be generic advertisement. Hence, the present study supports the body of
deep-rooted in the consumer’s psychology and be motivated by factors knowledge that advocates the use of celebrity endorsers (Ohanian, 1990;
such as consumer cosmopolitanism (H5). Pornpitakpan, 2003; Priester and Petty, 2003; Roy et al. 2012, 2019;
Roy et al., 2015; Silvera and Austad, 2004) and augments the literature
8. Summary of results in celebrity endorsements in a novel context (as suggested by Schim­
melpfennig and Hunt, 2020).
In this section, the results from all the four studies and the related Having discussed that, the study finds contradictory evidence to the
hypotheses are summarized. The objective of study 1 was to establish well-used congruence theory (Kamins, 1990; Mittelstaedt et al., 2000;
whether a celebrity endorser would be more effective in destination Roy and Bagdare, 2015; Till et al., 2008). Findings from study 2 indicate
marketing compared to a generic tourism advertisement irrespective of that celebrity endorser-destination congruence may not hold good
country culture (H1a-c). Support for all the hypotheses were observed as across countries, contrary to Glover (2009). While in some cases, a ce­
the celebrity endorsed advertisement was found to generate more lebrity congruent with the destination origin may be more effective,
favorable attitudes towards the advertisement (both cognitive and af­ such as the case of Brazil, it may not work for a newly globalized country
fective); attitudes towards the destination and visit intentions compared such as India. This also brings us to the country-specific consumer fac­
to a generic advertisement. tors that may moderate the effect of celebrity destination congruence on
Following this, study 2 was conducted to further explore the choice consumer attitudes (Study 3). The average level of ‘globalness’ of a
of the endorser. In study 2, the objective was to explore whether a local newly globalized country (in the present case, India) may be much
celebrity would be more effective than a global celebrity endorser for a higher than a country that was globalized earlier (such as Poland). This
domestic destination for the domestic travelers (H2a-c). The support for high level of global orientation may influence a consumer to accept a
the hypothesis was not universal across all countries. A full support for global celebrity compared to a local celebrity, even when the destination
H2 was found in Brazil where the local celebrity generated more is within the country. On the other hand, a country with a higher level of
favorable attitudes compared to the global (foreign) celebrity. However, local orientation may be more receptive to a local celebrity as compared
in case of India, the results were just the opposite, i.e., the global ce­ to a global one. Thereby, the findings support the notion that
lebrity was more effective than the local celebrity. In case of Poland, the country-specific consumer level factors may override the expected out­
both the celebrities were generating traveler attitudes at similar levels. comes of congruence (Tu et al., 2012; Roy et al., 2019).
Study 3 was conducted with the aim of understanding the internal This line of thought is extended in Study 4, where a third dimension
mechanism (or country specific nuances) of the traveler that was leading of the destination is added. Here, a bipartite congruence is introduced,
to the findings as in study 2. Thus, the objective of study 3 was to explore where on the one hand, the celebrity-destination congruence is dis­
the moderating effect of the consumer’s local/global identity on the cussed and on the other, the celebrity-consumer congruence is dis­
effectiveness of the endorser type (local/global). The findings suggested cussed. Given the notion of celebrity destination (C-D) congruence, one
a significant moderating effect of the consumers’ local/global identity could assume that a local celebrity would be more preferred for a do­
on the effectiveness of the endorser. Thus, H3a-c were supported. While mestic destination, whereas a global celebrity would be more preferred
individuals with more global identity would be more influenced by a for an international destination. On the contrary, congruence between
global celebrity, the individuals with more local identity would prefer a the celebrity endorser and the target audience (C-C) would imply that a
local celebrity. local celebrity would be more preferred irrespective of the destination.
Study 4 started with the question on whether the results from study 3 Findings from Study 4 partially support one of these notions. While for
should hold good for an international destination. Thus, the objective in Brazil, support for C-D congruence is observed, for India, neither C-D nor
study 4 was to explore whether a celebrity congruent with the destina­ C-C congruence was supported. This is where inference from consumer
tion (i.e. local celebrity for local destination and global celebrity for cosmopolitanism is drawn (e.g., Cleveland et al., 2009; Cleveland et al.,
international destination) would be more effective and whether these 2014; Nijssen and Douglas, 2008). The moderating effect of consumer
effects would be moderated by consumer cosmopolitanism. Even in this cosmopolitanism in Study 4 suggests that countries with a high level of
case, the findings were not universal. In India, the global celebrity was cosmopolitanism may prefer a global celebrity overriding the congru­
preferred irrespective of the destination type. In Brazil, the local ce­ ence between celebrity destination or celebrity consumer. While coun­
lebrity was preferred for the domestic destination while the global ce­ tries that are moderately high on cosmopolitanism may prefer a global
lebrity was more effective for the international destination. In case of celebrity for an international destination, the same audience may prefer
Poland, travelers were more or less indifferent between a local and a local celebrity for a domestic destination, thereby supporting C-D
global celebrity for a domestic destination, but were more influenced by congruence.
the global celebrity for the international destination. Thus, H4a-c was To summarize, the study finds partial support for the congruence
only supported in Brazil. The moderating effect pf consumer cosmo­ hypothesis and emphasizes on the consumer level characteristics such as
politanism was supported as it was found that individuals with higher local-global identity or consumer cosmopolitanism in influencing the
degree of cosmopolitanism would prefer a global celebrity while the effectiveness of the celebrity-endorsed destination advertising.
ones with lower degree of cosmopolitanism would prefer a local celeb­
rity. In the next section, the theoretical implications of the findings are 8.2. Managerial Implications
discussed.
There are multiple implications for the destination marketers that
9. Implications emanate from the present study. First, and the obvious one, the study
suggests justifiability of use of a celebrity endorser for the marketing of a
9.1. Theoretical implications tourist destination. Use of a credible celebrity in destination marketing
would generate favorable attitudes towards the advertisement and the
The present study contributes to the major thoughts in celebrity destination, leading to visit intentions. In addition, the findings were
endorsements and destination marketing and augments the existing consistent across countries with very different cultures. This implies that
literature (Aronson et al., 1999; Chan et al., 2018; van der Veen and a celebrity as an endorser may be effective irrespective of the cultural
Song, 2014). The findings from Study 1 (H1) support the notion of a context.
celebrity bringing credibility to an advertisement. Thereby, the celebrity Second, having said that an endorser would be more effective than a
endorsed advertisement was found to be more effective in generating generic marketing campaign, the subsequent studies pose challenges in

15
S. Roy et al. Tourism Management 83 (2021) 104213

front of the marketer on the choice of the celebrity. First, the case of the Luciana Gil: Writing - review & editing.
destination marketer who wants to market to the domestic audience is
discussed. The study findings demonstrate that consumer level factors Impact statement
could vary across and within countries, such as local-global identity, and
this may influence the effectiveness of the endorser. More specifically, in Research in the applications of celebrity endorsements in destination
a country where the majority of the consumers are globally oriented, a marketing is much less compared to the volume of research in main­
destination marketing campaign may be successful if the endorser is a stream celebrity endorsements. This is where we make a contribution by
global celebrity (not from the same country). On the other hand, a local providing a comprehensive picture of the various aspects of celebrity
celebrity may be more effective in a country where consumers are endorsed advertisements and supplementary guidelines for practice. The
locally oriented. Thereby, an a-priori understanding of the consumers fact that the same studies are repeated in three countries with different
perceived “localness”/“globalness” would help the marketer make a socio-cultural set-up, further enhances the generalizability of the find­
better choice of the endorser. ings. Hence, the present study makes a theoretical contribution to a
Third, the study findings also have significant implications for the hitherto under-researched area in tourism management. The findings
destination marketer, who wants to market a destination to a different emphasize on the need for a clear understanding of the socio-cultural
audience than the domestic market. Even in such a case, it is found that factors determining tourism in a country by the practitioner before
country-specific consumer factors would influence the choice of the crafting a destination marketing communication campaign. The findings
celebrity endorser. While more cosmopolitan consumers would be more from the present study will guide practitioners/policy makers in desti­
open to a foreign celebrity endorsing an international destination, less nation advertising in the selection and application of an endorsement
cosmopolitan consumers may prefer a local celebrity. Thus, while a strategy and crafting the advertisement plan to make it more effective.
match between the destination and the celebrity (in terms of the country
of origin) may be effective for some countries, it may be more sensible to
use a local celebrity matching the target audience in some other coun­ Declaration of competing interest
tries. Even in this case, the countries planning a destination marketing
campaign for multiple target countries may not find a single solution. None.
Based on the degree of cosmopolitism, they may have to use a different
celebrity for specific target countries rather than a common celebrity for Acknowledgements
all targets. To summarize, findings from the study note that even though
celebrity endorsements are relatively more beneficial, they need to be The author would like to thank the anonymous reviewers for their
tailored depending on the characteristic of the target audience. constructive feedback on the earlier draft of the manuscript.

9. Limitations and further research


References
The present study brings together the research streams of celebrity Aaker, J. L., Benet-Martinez, V., & Garolera, J. (2001). Consumption symbols as carriers
endorsements and destination marketing and contributes to the knowl­ of culture: A study of Japanese and Spanish brand personality constructs. Journal of
edge on celebrity endorsements in destination marketing with the help personality and social psychology, 81(3), 492–508.
Agnihotri, A., & Bhattacharya, S. (2018). The market value of celebrity endorsement:
of four empirical studies, each done consecutively in three countries. evidence from India reveals factors that can influence stock-market returns. Journal
The major findings challenge some existing notions in celebrity of Advertising Research, 58(1), 65–74.
endorsement and add to the body of knowledge in destination market­ Alden, D. L., Steenkamp, J. B. E., & Batra, R. (1999). Brand positioning through
advertising in Asia, North America, and Europe: The role of global consumer culture.
ing. The study has a few limitations that could be considered as scope for Journal of marketing, 63(1), 75–87.
future research. First, the study only measured attitudes and behavioral Alden, D. L., Steenkamp, J. B. E., & Batra, R. (2006). Consumer attitudes toward
intentions. Whether the visit intentions lead to actual visit was beyond marketplace globalization: Structure, antecedents and consequences. International
Journal of Research in Marketing, 23(3), 227–239.
the scope of the present study. Future studies could explore this in a
Amar, J., Droulers, O., & Legohérel, P. (2017). Typography in destination advertising: An
longitudinal study where the actual behavior could be mapped. Second, exploratory study and research perspectives. Tourism Management, 63, 77–86.
even though the study had theoretical generalizability, future studies December.
could explore multiple destinations and multiple celebrity professions. Amos, C., Holmes, G., & Strutton, D. (2008). Exploring the relationship between celebrity
endorser effects and advertising effectiveness: A quantitative synthesis of effect size.
For example, one could argue that soccer and beaches go well together. International Journal of Advertising, 27(2), 209–234.
Thus, would the same effects hold good if a soccer star is paired with a Arnett, J. J. (2002). The psychology of globalization. American Psychologist, 57(10),
mountain destination, or say a music star is paired with a beach. Third, 774–783.
Aronson, E., Wilson, T. D., & Akert, R. M. (1999). Social psychology (3rd ed.). New York,
the opinion of a single individual was only considered in the study. The NY: Addison Wesley Longman.
choice of travel as a tourist may not always be a single individual’s Bapna, A. (2017). Ranveer Singh is the first Indian ambassador for Switzerland Tourism.
decision and maybe a joint decision. Consequently, future research Economic Times. Retrieved from https://economictimes.indiatimes.com/ranveer-s
ingh-is-the-first-indian-ambassador-for-switzerland-tourism/articleshow/57750871.
should explore whether a peer group influence moderates the effect of a cms.
celebrity endorser on the consumer choices. Despite the limitations, the Boeing, R., & Schurhaus, C. (2014). The effect of celebrity endorsement on Brazilian
present study has made a novel and timely contribution to destination consumer behavior: Does it really matter? International Business Research, 7(5),
49–58.
marketing both from the academic and practitioner points of view. Burton, S., & Lichtenstein, D. R. (1988). The effect of ad claims and ad context on attitude
toward the advertisement. Journal of Advertising, 17(1), 3–11.
Credit roles Campelo, A., Aitken, R., & Gnoth, J. (2011). Visual rhetoric and ethics in marketing of
destinations. Journal of Travel Research, 50(1), 3–14.
Carlson, B. D., Donavan, D. T., Deitz, G. D., Bauer, B. C., & Lala, V. (2020). A customer-
Subhadip Roy, Conceptualization, Subhadip Roy, Wioleta Dryl, focused approach to improve celebrity endorser effectiveness. Journal of Business
Luciana Gil: Data curation, Subhadip Roy: Formal analysis, Funding Research, 109(March), 221–235.
Chan, J. K. (2010). Building sustainable tourism destination and developing responsible
acquisition, (Not applicable) Subhadip Roy, Wioleta Dryl, Luciana Gil,
tourism: Conceptual framework, key issues and challenges. Tourism Development
Investigation, Subhadip Roy, Methodology, Subhadip Roy, Wioleta Dryl, Journal-An International Research Journal, 8(1), 24–32.
Luciana Gil: Project administration Resources, (Not applicable) Sub­ Chang, C. (2008). Chronological age versus cognitive age for younger consumers:
hadip Roy: Software Supervision (Not applicable) Subhadip Roy: Vali­ Implications for advertising persuasion. Journal of Advertising, 37(3), 19–32.
Chan, G. S. H., Lee, A. L. Y., & Wong, C. H. M. (2018). Celebrity endorsement in
dation, Subhadip Roy: Visualization, Subhadip Roy, Wioleta Dryl, advertisement on destination choice among generation Y in Hong Kong. International
Luciana Gil: Roles/Writing - original draft, Subhadip Roy, Wioleta Dryl, Journal of Marketing Studies, 10(2), 16–27.

16
S. Roy et al. Tourism Management 83 (2021) 104213

Choe, Y., Stienmetz, J. L., & Fesenmaier, D. R. (2017). Measuring destination marketing: Maddux, J., & Rogers, R. (1980). Effects of source expertness, physical attractiveness,
Comparing four models of advertising conversion. Journal of Travel Research, 56(2), and supporting arguments on persuasion: A case of brains over beauty. Journal of
143–157. Personality and Social Psychology, 39(2), 235–244.
Choi, S. M., & Rifon, N. J. (2012). It is a match: The impact of congruence between Mandler, G. (1982). The structure of value: Accounting for taste. In M. S. Clark, &
celebrity image and consumer ideal self on endorsement effectiveness. Psychology S. T. Fiske (Eds.), Affect and cognition: The 17th annual carnegie symposium (pp. 3–36).
and Marketing, 29(9), 639–650. Hillsdale, NJ: Lawrence Erlbaum.
Cleveland, M., Laroche, M., & Papadopoulos, N. (2009). Cosmopolitanism, consumer McCracken, G. (1989). Who is the celebrity endorser? Cultural foundations of the
ethnocentrism, and materialism: An eight-country study of antecedents and endorsement process. Journal of Consumer Research, 16(3), 310–321.
outcomes. Journal of International Marketing, 17(1), 116–146. McGuire, W. J. (1985). Attitudes and attitude change. In L. Gardner, & E. Aronson (Eds.),
Cleveland, M., Laroche, M., Takahashi, I., & Erdoğan, S. (2014). Cross-linguistic Handbook of social psychology (pp. 233–346). New York: Random House.
validation of a unidimensional scale for cosmopolitanism. Journal of Business Mishra, A. S., Roy, S., & Bailey, A. A. (2015). Exploring brand personality–celebrity
Research, 67(3), 268–277. endorser personality congruence in celebrity endorsements in the Indian context.
Datta, A. (2019). Virat Kohli is India’s most valued celebrity brand. Fortune India. Psychology and Marketing, 32(12), 1158–1174.
January Retrieved from https://www.fortuneindia.com/people/virat-kohli-is-india Mittelstaedt, J. D., Riesz, P. C., & Burns, W. J. (2000). Why are endorsements effective?
s-most-valued-celebrity-brand/102849 . (Accessed 6 March 2019). Sorting among theories of product and endorser effects. Journal of Current Issues and
Derdenger, T. P., Li, H., & Srinivasan, K. (2018). Firms’ strategic leverage of unplanned Research in Advertising, 22(1), 55–65.
exposure and planned advertising: An analysis in the context of celebrity Nijssen, E. J., & Douglas, S. P. (2008). Consumer world-mindedness, social-mindedness,
endorsements. Journal of Marketing Research, 55(1), 14–34. and store image. Journal of International Marketing, 16(3), 84–107.
Deshpandé, R., & Stayman, D. M. (1994). A tale of two cities: Distinctiveness theory and Ohanian, R. (1990). Construction and validation of a scale to measure celebrity
advertising effectiveness. Journal of Marketing Research, 31(1), 57–64. endorsers’ perceived expertise, trustworthiness, and attractiveness. Journal of
Erdogan, B. Z. (1999). Celebrity endorsement: A literature review. Journal of Marketing Advertising, 19(3), 39–52.
Management, 15(4), 291–314. Ohanian, R. (1991). The impact of celebrity spokespersons’ perceived image on
Featherstone, M. (2002). Cosmopolis: An introduction. Theory, Culture & Society, 19 consumers’ intention to purchase. Journal of Advertising Research, 31(1), 46–54.
(1–2), 1–16. Owen, E. (2015, April 28). Celebrity endorsements: Eight stars who are also tourism
Forehand, M. R., Deshpandé, R., & Reed, I. I. (2002). Identity salience and the influence ambassadors. Retrieved from https://www.travelandleisure.com/articles/celebrit
of differential activation of the social self-schema on advertising response. Journal of y-endorsements-eight-stars-who-are-also-tourism-ambassadors. (Accessed 6 March
Applied Psychology, 87(6), 1086–1099. 2019).
Gartner, W. C. (1994). Image formation process. Journal of Travel & Tourism Marketing, 2 Park, S., Nicolau, J. L., & Fesenmaier, D. R. (2013). Assessing advertising in a
(2–3), 191–216. hierarchical decision model. Annals of Tourism Research, 40(1), 260–282.
Gilal, F. G., Paul, J., Gilal, N. G., & Gilal, R. G. (2020). Celebrity endorsement and brand Park, S. Y., & Yim, M. Y. C. (2020). Do celebrity endorsements benefit familiar luxury
passion among air travelers: Theory and evidence. International Journal of Hospitality brands? A perspective from social adaptation theory. Journal of Current Issues and
Management, 85 ([online]). Research in Advertising, 41(1), 20–35.
Glover, P. (2009). Celebrity endorsement in tourism advertising: Effects on destination Peltier, D. (2017, March 16). Many Destinations Use Celebrities for Marketing but Pitbull
image. Journal of Hospitality and Tourism Management, 16(1), 16–23. Fiasco is a Warning Sign. Retrieved from https://skift.com/2017/03/16/many-destin
Hannerz, U. (1990). Cosmopolitans and locals in world culture. Theory, Culture & Society, ations-use-celebrities-for-marketing-but-pitbull-fiasco-is-a-warning-sign/. (Accessed
7(2–3), 237–251. 6 March 2019).
Hayes, A. F. (2013). Model templates for PROCESS for SPSS and SAS. [White paper]. Pike, S., Murdy, S., & Lings, I. (2011). Visitor relationship orientation of destination
Retrieved from http://www.afhayes.com/public/templates.pdf January 2019. marketing organizations. Journal of Travel Research, 50(4), 443–453.
Holloway, J.С., & Robinson, С. (1995). Marketing for tourism (3rd ed.). Reading: Reino Pornpitakpan, C. (2003). Validation of the celebrity endorsers’ credibility scale: Evidence
Unido, Adisson Wesley Longman. from Asians. Journal of Marketing Management, 19(1–2), 179–195.
Hovland, C. I., & Weiss, W. (1951). The influence of source credibility on communication Priester, J. R., & Petty, R. E. (2003). The influence of spokesperson trustworthiness on
effectiveness. Public Opinion Quarterly, 15(4), 635–650. message elaboration, attitude strength, and advertising effectiveness. Journal of
Hung, K., Chan, K. W., & Caleb, H. T. (2011). Assessing celebrity endorsement effects in Consumer Psychology, 13(4), 408–421.
China: A consumer-celebrity relational approach. Journal of Advertising Research, 51 Pritchard, A., & Morgan, N. J. (2001). Culture, identity and tourism representation:
(4), 608–623. Marketing cymru or wales? Tourism Management, 22(2), 167–179.
Hussain, S., Melewar, T. C., Priporas, C. V., Foroudi, P., & Dennis, C. (2020). Examining Riefler, P., Diamantopoulos, A., & Siguaw, J. A. (2012). Cosmopolitan consumers as a
the effects of celebrity trust on advertising credibility, brand credibility and target group for segmentation. Journal of International Business Studies, 43(3),
corporate credibility. Journal of Business Research, 109(March), 472–488. 285–305.
Jamal, A. (2016, November 9). shah rukh khan becomes brand ambassador of dubai. Roy, S., & Bagdare, S. (2015). The role of country of origin in celebrity endorsements:
Retrieved from https://www.connectedtoindia.com/shah-rukh-khan-becomes-brand Integrating effects of brand familiarity. Journal of Global Marketing, 28(3–5),
-ambassador-of-dubai-101.html. (Accessed 6 March 2019). 133–151.
Janiszewski, C. (1988). Preconscious processing effects: The independence of attitude Roy, S., Gammoh, B. S., & Koh, A. C. (2012). Predicting the effectiveness of celebrity
formation and conscious thought. Journal of Consumer Research, 15(2), 199–209. endorsements using the balance theory. Journal of Customer Behaviour, 11(1), 33–52.
Kamins, M. A. (1990). An investigation into the “match-up” hypothesis in celebrity Roy, S., Guha, A., & Biswas, A. (2015). Celebrity endorsements and women consumers in
advertising: When beauty may be only skin deep. Journal of Advertising, 19(1), 4–13. India: How generation-cohort affiliation and celebrity-product congruency moderate
Kim, K., Hallab, Z., & Kim, J. N. (2012). The moderating effect of travel experience in a the benefits of chronological age congruency. Marketing Letters, 26(3), 363–376.
destination on the relationship between the destination image and the intention to Roy, S., Guha, A., Biswas, A., & Grewal, D. (2019). Celebrity endorsements in emerging
revisit. Journal of Hospitality Marketing & Management, 21(5), 486–505. markets: Align endorsers with brands or with consumers? Journal of International
Kim, H., & John, D. R. (2008). Consumer response to brand extensions: Construal level as Business Studies, 50(3), 295–317.
a moderator of the importance of perceived fit. Journal of Consumer Psychology, 18 Sahoo, A. K. (2006). Issues of identity in the Indian diaspora: A transnational perspective.
(2), 116–126. Perspectives on Global Development and Technology, 5(1–2), 81–98.
Kim, H., & Richardson, S. L. (2003). Motion picture impacts on destination images. Saldanha, N., Mulye, R., & Rahman, K. (2020). A strategic view of celebrity
Annals of Tourism Research, 30(1), 216–237. endorsements through the attachment lens. Journal of Strategic Marketing, 28(5),
Kirmani, A., & Shiv, B. (1998). Effects of source congruity on brand attitudes and beliefs: 434–454.
The moderating role of issue-relevant elaboration. Journal of Consumer Psychology, 7 Saran, A., Guo, C., & Kulviwat, S. (2007). Marketing strategy in transition economies:
(1), 25–47. The case for India. International Journal of Business Research, 7(2), 85–93.
Knoll, J., & Matthes, J. (2017). The effectiveness of celebrity endorsements: A meta- Schimmelpfennig, C., & Hunt, J. B. (2020). Fifty years of celebrity endorser research:
analysis. Journal of the Academy of Marketing Science, 45(1), 55–75. Support for a comprehensive celebrity endorsement strategy framework. Psychology
Kotler, P., Haider, D., & Rein, I. (1993). There’s no place like our place! the marketing of and Marketing, 37(3), 488–505.
cities, regions, and nations. Futurist, 27(6), 14. Schneider, R. M. (2018). Brazil: Culture and politics in a new industrial powerhouse.
Kotsi, F., & Valek, N. S. (2018). Flying with nicole kidman or jennifer aniston? Brand Boulder, CA: Routledge.
funnel stages’ influence on brand personality. Journal of Travel & Tourism Marketing, Schouten, A. P., Janssen, L., & Verspaget, M. (2020). Celebrity vs. Influencer
35(3), 365–376. endorsements in advertising: The role of identification, credibility, and product-
Kowalewska, K. (2018). The effectiveness of celebrity endorsements in Polish magazines. endorser fit. International Journal of Advertising, 39(2), 258–281.
Annales. Etyka w Życiu Gospodarczym, 21(4), 31–44. Silvera, D. H., & Austad, B. (2004). Factors predicting the effectiveness of celebrity
Lam, T., & Hsu, C. H. (2006). Predicting behavioral intention of choosing a travel endorsement advertisements. European Journal of Marketing, 38(11/12), 1509–1526.
destination. Tourism Management, 27(4), 589–599. Spangenberg, E. R., Sprott, D. E., Grohmann, B., & Tracy, D. L. (2006). Gender-congruent
Lee, S., Scott, D., & Kim, H. (2008). Celebrity fan involvement and destination ambient scent influences on approach and avoidance behaviors in a retail store.
perceptions. Annals of Tourism Research, 35(3), 809–832. Journal of Business Research, 59(12), 1281–1287.
Levy, D., & Sznaider, N. (2007). Memories of europe: Cosmopolitanism and its others. In Spry, A., Pappu, R., & Cornwell, B. T. (2011). Celebrity endorsement, brand credibility
C. Rumford (Ed.), Cosmopolitanism and europe (pp. 158–177). Liverpool: Liverpool and brand equity. European Journal of Marketing, 45(6), 882–909.
University Press. Steenkamp, J. B. E., Batra, R., & Alden, D. L. (2003). How perceived brand globalness
MacKenzie, S. B., & Lutz, R. J. (1989). An empirical examination of the structural creates brand value. Journal of International Business Studies, 34(1), 53–65.
antecedents of attitude toward the ad in an advertising pretesting context. Journal of Stergiou, D., & Airey, D. (2003). Inquiry conversion and tourism website effectiveness:
Marketing, 53(2), 48–65. Assumptions, problems and potential. Tourism and Hospitality Research, 4(4),
355–366.

17
S. Roy et al. Tourism Management 83 (2021) 104213

Stienmetz, J. L., & Fesenmaier, D. R. (2014). Effects of channel, timing, and bundling on Zhang, H., Xu, H., & Gursoy, D. (2020). The effect of celebrity endorsement on
destination advertising response. Tourism Analysis, 19(1), 97–104. destination brand love: A comparison of previous visitors and potential tourists.
Stienmetz, J. L., Maxcy, J. G., & Fesenmaier, D. R. (2015). Evaluating destination Journal of Destination Marketing & Management, 17 [online]. WEB REF https://www.
advertising. Journal of Travel Research, 54(1), 22–35. hofstede-insights.com/product/compare-countries/information. (Accessed 22 March
Stolecka-Makowska, A. (2016). Kultura jako determinanta zmian zachowań nabywczych 2019).
konsumentów podlegających akulturacji. Studia Ekonomiczne, 270, 265–274.
Tajfel, H., & John, C. T. (1986). The social identity theory of intergroup behavior. In
S. Worchel, & W. G. Austin (Eds.), Psychology of intergroup relations (pp. 7–24).
Chicago, IL: Nelson-Hall. Subhadip Roy is an Associate Professor of marketing at the
Teng, H. Y., & Chen, C. Y. (2020). Enhancing celebrity fan-destination relationship in Indian Institute of Management Ahmedabad, India. His areas of
film-induced tourism: The effect of authenticity. Tourism Management Perspectives, 33 research are experience marketing, celebrity endorsements and
brand management. His papers have appeared in Journal of
([online]).
Thwaites, D., Lowe, B., Monkhouse, L. L., & Barnes, B. R. (2012). The impact of negative International Business Studies, Industrial Marketing Manage­
ment, Journal of Advertising Research and Psychology &
publicity on celebrity ad endorsements. Psychology and Marketing, 29(9), 663–673.
Till, B. D., Stanley, S. M., & Priluck, R. (2008). Classical conditioning and celebrity Marketing, among others.
endorsers: An examination of belongingness and resistance to extinction. Psychology
and Marketing, 25(2), 179–196.
Turner, B. S. (2002). Cosmopolitan virtue, globalization and patriotism. Theory, Culture &
Society, 19(1–2), 45–63.
Uy, S., & Wu, Y. (2020). When luxury brands meet China: The effect of localized celebrity
endorsements in social media marketing. Journal of Retailing and Consumer Services,
54 ([online]).
Vargas, P. T., Duff, B. R., & Faber, R. J. (2017). A practical guide to experimental
advertising research. Journal of Advertising, 46(1), 101–114.
van der Veen, R. (2008). Analysis of the implementation of celebrity endorsement as a Wioleta Dryl is Assistant Professor of marketing at the Uni­
destination marketing instrument. Journal of Travel & Tourism Marketing, 24(2–3), versity of Gdańsk, Poland. Here research interests are in the
213–222. areas of luxury branding and tourism marketing. Her papers
van der Veen, R. (2009). Celebrity endorsement effectiveness for print destination have appeared in Economics and Business Review and Journal
advertising. European Journal of Tourism Research, 2(2), 186. of Intercultural Management among others.
van der Veen, R., & Song, H. (2010). Exploratory study of the measurement scales for the
perceived image and advertising effectiveness of celebrity endorsers in a tourism
context. Journal of Travel & Tourism Marketing, 27(5), 460–473.
van der Veen, R., & Song, H. (2014). Impact of the perceived image of celebrity endorsers
on tourists’ intentions to visit. Journal of Travel Research, 53(2), 211–224.
Vhora, F. B. (2018, July 19). Do Bollywood celebrities help foreign tourism boards that sign
them up?. Retrieved from https://www.cnbctv18.com/views/do-bollywood-celebriti
es-help-foreign-tourism-boards-that-sign-them-up-340541.htm . (Accessed 12 March
2019).
Walker, M., Langmeyer, L., & Langmeyer, D. (1992). Commentary: Celebrity endorsers:
Do you get what you pay for? Journal of Services Marketing, 6(4), 35–42.
Wänke, M., Herrmann, A., & Schaffner, D. (2007). Brand name influence on brand Luciana Gil is an Associate Professor of marketing at the Uni­
perception. Psychology and Marketing, 24(1), 1–24. versidad Diego Portales, Chile. Her research areas are in the
Weeks, C. S., Cornwell, T. B., & Drennan, J. C. (2008). Leveraging sponsorships on the effects of culture in consumption and luxury branding. Her
Internet: Activation, congruence, and articulation. Psychology and Marketing, 25(7), papers have appeared in European Journal of Marketing and
637–654. Consumption Markets and Culture among others.
Wheeler, S. C., Petty, R. E., & Bizer, G. Y. (2005). Self-schema matching and attitude
change: Situational and dispositional determinants of message elaboration. Journal
of Consumer Research, 31(4), 787–797.
Winterich, K. P., Gangwar, M., & Grewal, R. (2018). When celebrities count: Power
distance beliefs and celebrity endorsements. Journal of Marketing, 82(3), 70–86.
Woodside, A. G. (2010). Tourism advertising and marketing performance metrics. In
A. G. Woodside (Ed.), Tourism-marketing performance metrics and usefulness auditing of
destination websites (pp. 1–14). Bingley, UK: Emerald.
Yen, C. H., & Teng, H. Y. (2015). Celebrity involvement, perceived value, and behavioral
intentions in popular media-induced tourism. Journal of Hospitality & Tourism
Research, 39(2), 225–244.

18

You might also like