National or Global? Moderated Mediation Impact of Sports Celebrity Credibility on Consumer’s Purchase Intention

The paper aims to examine the role of localness of a sports celebrity and a level of product involvement in the mediated impact of the perceived product quality on the relationship between a sports celebrity’s credibility and consumers’ purchase intention in a developing economy. A total of 253 respondents participated in an experiment including the localness of 2 sports celebrities (global vs national) and2 product involvement (high vs low) levels. The study determined that sports celebrity credibility had both direct and indirect effects on respondents’ intention to buy, but product involvement moderated the direct impact of sports celebrity credibility on the consumer’s intention to buy a product. The direct impact was noticed in the case of low involvement products, and no impact was observed in the situation of high involvement. Besides, the study showed that global sports celebrities enjoyed a higher level of attractiveness, but the trustworthiness was higher for national celebrities. Finally, moderation analysis showed that the mediation effect of the localness of a sports celebrity on the relationship between credibility and intention to purchase depended on the type of measured effect – direct or indirect. This study expands the research on the effects of celebrity credibility on the consumer’s intention to purchase in developing economies.


Introduction
The use of celebrity endorsement has been a popular trend for more than a century (Erdogan, 1999). Companies invest large sums of money in having celebrity endorsers in their advertisements hoping to achieve returns on their investments (Roozen & Claeys, 2010). Celebrities were used in 25% of commercials in the United States and even 70% of advertisements in Korea and Japan (Euromonitor International, 2014). One kind of celebrity, sports celebrities, are used quite often for advertising products (Kunkel, Walker & Hodge, 2018;Lee & Koo, 2015;Liang & Lin, 2018), since it is believed that consumers were intended to have psychological relationships with them, copy their behaviour and lifestyle or values (Foong & Yazdanifard, 2014). A growing number of publications since 2010 reflects an academic concern with the subject as well (Rocha, de Oliveira & Giraldi, 2019). Knoll and Matthes (2017) pointed out a growing interest in celebrity endorsements in emerging countries such as India and China due to cultural differences. Other authors emphasize higher importance of celebrity endorsements on consumers' behaviour in emerging markets such as Asia and Latin America than developed countries such as the United States and the United Kingdom (Paek, 2005;Kulkarni, 2012). Whereas the phenomenon of celebrity endorsement has been widely investigated in developed countries (Bergkvist & Zhou 2016), several aspects like sports celebrity's localness and a level of product involvement are relatively understudied.
Some previous studies reported the celebrity's direct impact on intentions to purchase (Ahmed, Farooq & Iqbal, 2014;Nguyen & Huynh, 2018;Ong & Ong, 2015). However, other studies showed that celebrities could have just an indirect impact on it since such impact was made through an increase of product quality perception (Chang, 2014;Dwivedi, Johnson & McDonald, 2016;Srivastava, 2011). Such results raise the question of whether celebrities drive consumer's intention to purchase products directly or through a mediator -the perceived quality of products. Jorge (2015) argued that celebrity is a global phenomenon and is related to a symbol of globalisation. However, studies of the localness of a celebrity are very rare and present contradicting results. On one side, some results argued that global celebrities (including sports celebrities) had a higher impact on intention to buy products than national ones ( Jorge, 2015;Keat, 2015), which could happen due to "the value of voice"the number of followers or higher awareness or even attractiveness (Kucuk, 2020). On another side, national celebrities represent a certain culture, share the same values, or can even be heroes, which increases their reliability or trustworthiness. Since consumers need honest and trustworthy information, the national (or local) celebrities could be more influential (Kucuk, 2020). Some authors noticed the popularity of national celebrities in developing countries (Dwivedi et al., 2016;Saxena, 2008). Even though the national celebrity is perceived as depending largely on local media ( Jorge, 2015), the importance of it might grow due to the popularity and widespread adoption of social media as well as unique features of Web 2.0 for self-promotion ( Jin, 2018).
Finally, the characteristics of an advertised product could moderate a celebrity's ability to influence consumers' purchase decisions (Schouten, Janssen & Verspaget, 2020). Eslami and Ghasemaghaei (2018) stated that consumers were more engaged in information search and evaluation when they search for high involvement products, compared to those who are looking for low involvement products due to the financial and social risk. Thus, a celebrity's credibility could increase the perceived quality of high-involvement products and stimulate their purchasing (Al Zoubi & Bataineh, 2011;Punyatoya, 2011;Renton, 2009). On the other hand, Roozen and Claeys (2010), Saleem (2007) voted for using celebrities in advertisements of low-involvement products. The latest research (Schouten et al., 2020) noticed that celebrity effectiveness depended on a product type and recommended paying more attention to it in future studies.
The paper aims to examine the role of localness of a sports celebrity and a level of product involvement in the mediated impact of the perceived product quality on the relationship between a sports celebrity's credibility and consumers' purchase intention. Since the aim of the article is twofold, we first investigate the indirect and direct paths between the sports celebrity credibility, perceived quality and purchase intention. Second, the article evaluates the impact of two moderating variables (the localness of a sports celebrity and a level of product involvement) on both the direct and indirect paths.

Impact of a sports celebrity's credibility on consumer's purchase intention
Celebrities are used in advertisements very often since it is believed that consumers are intended to have psychological relationships with them, copy their behaviour and lifestyle or values (Fraser, Hill & Sotiriadou, 2016;Foong & Yazdanifard, 2014). Professional athletes and coaches are quite often used as celebrities for companies' marketing activities (Bush, Martin & Bush, 2004;Lear, Runyan & Whitaker, 2009;Liu, Huang & Minghua, 2007). Companies use sports celebrities for the advertisement of products with various level of involvement for an increase of brand awareness and sales growth (Al Zoubi & Bataineh, 2011;Liu et al., 2007;Liu & Brock, 2011). Sports celebrities are significant sources of information since they could influence or even form the social values of consumers (Al Zoubi & Bataineh, 2011). Renton (2009) noticed that famous athletes became an unseparated part of advertising campaigns for many brands. Some authors (Liu et al., 2007;Liu & Brock, 2011) stated that sports celebrities advertised products that could be quite distant from their professional activity like soup, soft drink, cars, internet, or financial services.
Previous studies intensively analysed celebrity endorser effectiveness through endorser quality -credibility, and identified a relationship between celebrity credibility and consumer's purchase decision. Chung, Derdenger and Srinivasan, (2013) and El-berse and Verleun (2012) reported a significant increase in sales over the endorsement period. Other studies identified a direct impact of celebrity credibility on an intention to purchase a product (Ahmed et al., 2014;Koo, Ruihley & Dittmore, 2012;Lee & Koo, 2015;Nguyen & Huynh, 2018;Ong & Ong, 2015). Thus, we expect that: H 1 : Credibility of a sports celebrity has a positive direct impact on the consumers' intention to purchase a product. Knoll and Matthes (2017) performed a meta-analysis of previous studies and concluded that celebrity credibility had no direct effect on consumers' purchase intention, but it influenced consumers' attitudes toward advertised objects. Some authors added that people pay more attention to ads with celebrities (Wei & Lu, 2013), which increases awareness of an advertised product (Miller & Allen, 2012). Moreover, some articles stated that celebrity endorsement improved brand image (Lee & Koo, 2015), especially in a case of low awareness of it (Chang, 2014;Chan, Leung & Luk, 2013;Dwivedi, Johnson & McDonald, 2016;Liu et al., 2007;Liu & Brock, 2011). Other authors (Chang, 2014;Dwivedi et al., 2016;Srivastava, 2011) noticed a positive impact of celebrity credibility on consumers' attitudes towards the perceived product quality, which is usually defined as a way in which a customer views a product's overall superiority rather than individual elements of quality (Beneke, Flynn, Greig & Mukaiwa, 2013). Some authors noticed that the credibility of endorsers can also influence consumers' general attitudes towards the product (Amos, Holmes & Strutton, 2008;Goldsmith, Lafferty & Newell, 2000). The positive impact of celebrity endorsement on product quality was found in the studies of sports celebrities as well (Braunstein-Minkove, Zhang & Trai, 2011;Tong, Su & Hu, 2017;Zhou, Mou, Su & Wu, 2020). Thus, we hypothesize that: H 2 : Perceived product's quality mediates the impact of credibility of a sports celebrity on the consumers' intention to purchase a product.

Importance of localness of a sports celebrity
For several decades media and entertainment industries have been dominated by transnational and cross-media concentrating and converging companies, pushing (cultural and wider) globalisation. Besides globalisation of business, with concentrated media ownership structures, vertically and horizontally integrated companies, new media technologies equip celebrities with some globally accessible platforms which are used to promote companies' products and affiliations (Potter & Hill, 2017). Celebrities, according to Turner (2013), play a crucial role in these processes and earn their social capital. According to Bourdieu's (1986, p. 248) field theory, social capital is "the aggregate of the actual or potential resources which are linked to possession of a durable network of more or less institutionalized relationships of mutual acquaintance and recognition". This theoretical background was used for the development of celebrity capi-tal notion, which refers to the accumulation of media visibility through recurrent media representations (Driessens, 2013). Carrillat and Ilicic (2019) added that celebrity capital includes much more than just media visibility, it is related to their expertise, trustworthiness, likeability or credibility. Since the volume of social capital thus depends on the number of people that can be mobilized in an individual's network (Driessens, 2013), we can expect that a global sports celebrity will have higher credibility than a national one. Therefore we hypothesize: H 3 : A global sports celebrity will have higher credibility than a national one.
The celebrity capital (as a kind of social capital) can be converted to economic capital through endorsing consumer goods and brands (Zulli, 2018, Kim, Kang & Lee, 2018 and could have an impact on intention to buy products. Dwivedi et al. (2016) and Keat (2015) found that global sports celebrities had a higher impact on the perception of a product compared with national sports celebrities. Roy and Bagdare (2015) noticed that the consumers might attribute elements of product quality if a known international celebrity is endorsing it, especially when both the celebrity and the brand (or product) are global (or at least non-native). As it was mentioned, celebrities could have a direct impact on the intention to buy products rather than through perception of product quality. Some previous studies noticed that global celebrities (including sports celebrities) had a stronger impact on the intention to buy products than national ones (Dwivedi et al., 2016;Keat, 2015). Thus, we could expect that a global sports celebrity would have a stronger impact (either direct or indirect) on the intention to buy products.
In contrast to this, there is a widespread use of national celebrities in developing countries (Saxena, 2008). National celebrities are famous in developed countries as well as globally due to national versions of global formats such as reality television or sports events (Dhoest, 2005). Chenu (2008) and Ferris (2010) stated that national media pay ample attention to these home celebrities. Therefore, national celebrities quite often have higher awareness in a specific market than in the global one. Moreover, national sports celebrities get the status of national heroes (Dumitriu, 2018) or simply can be perceived as more credible (Toncar, Reid & Anderson, 2007). Due to this, national celebrities could have a stronger impact on consumers' purchase intention than global celebrities (Kay, Mulcahy & Parkinson, 2020). Due to the mentioned we expect that: H 4 : Localness of a sports celebrity will moderate the impact of the sports celebrity's credibility on the consumers' intention to purchase.
H 4A : Localness of a sports celebrity will moderate the direct impact of the sports celebrity's credibility on the consumers' intention to purchase.
H 4B : Localness of a sports celebrity will moderate indirect impact (through perceived quality) of the sports celebrity's credibility on the consumers' intention to purchase. Moderated Mediation Impact of Sports Celebrity Credibility on Consumer's Purchase Intention

Moderating role of product involvement
The concept of involvement has been widely used in marketing for the classification of products according to the level of involvement they arouse (Michaelidou & Dibb, 2008). In the light of this article, we accept the perception of product involvement which means that high involvement is associated with greater levels of importance, personal relevance, and consequences being attached to the object or consumption context (Alexandris, Tsiotsou & James, 2007;Smith & Stewart, 2007;Thorne & Bruner, 2006). By contrast, low involvement products do not involve complex cognitive thinking (Punyatoya, 2011;Roozen & Claeys, 2010), and consumers more frequently purchase low involvement products compared to high involvement ones (Park & Yoon, 2017). There is no consensus about the impact of the type of involvement on endorser credibilitybrand credibility link. On one hand, academic literature supports the view that celebrities are appropriate for high-involvement products (Friedman & Friedman, 1979;Kahle & Homer, 1985) since low involvement products are purchased repeatedly or even routinely and have no impact on consumers' lifestyle (Chung & Zhao, 2003). On the contrary, the literature supports the view that celebrity endorsers should be used for low-involvement products to have an impact on the brand (Roozen & Claeys, 2010;Saleem, 2007). With low-involvement products, consumers spend a limited amount of time reviewing the advertisement. Thus, the visual impression of celebrity endorser is key to obtaining information (Kahle & Homer, 1985). At the same time, celebrity credibility could increase the perceived quality of high involvement products and stimulate purchasing (Al Zoubi & Bataineh, 2011;Punyatoya, 2011;Renton, 2009 Park, Lee and Han (2007) noticed the moderating role of involvement in the product category. Based on the above we state: H 5 : Product involvement will moderate the impact of sports celebrity credibility on the consumers' intention to purchase.
H 5A : Product involvement will moderate the direct impact of sports celebrity credibility on the consumers' intention to purchase.
H 5B : Product involvement will moderate the indirect impact (through perceived quality) of sports celebrity credibility on the consumers' intention to purchase.

Stimuli selection
The effect of sports celebrity credibility on the consumers' intention to purchase products was tested using the 2 x 2 factorial design of an experiment. Two types of sports celebrities -national and global ones were included. Previous studies (Klaus & Bailey, 2008;Spry, Pappu & Cornwell, 2011) noticed that consumers reacted differently to male and female celebrities. Therefore, two male celebrities were chosen. Both of them were related to the most-watched sports events -football (David Beckham) and basketball (Robertas Javtokas). Both of them were well known during their career and had finished it. Another variable of the experiment was the level of product involvement. According to previous studies (Lee & Park, 2014;Punyatoya, 2011;Saleem, 2007), a level of product involvement could have an impact on the results. Therefore two products were chosen -a bottle of water as a low involvement product (Friedmann & Lowengart, 2019) and a personal computer (Cochrane & Quester, 2005) as a high involvement product. Products were without brands to avoid the problem of familiarity and different image of the brand in various groups of respondents. Respondents were informed that they were to imagine that the celebrity advertised the product and were asked to present their opinions. Since the developed situations included the same products from the low and high involvement category, the respondents were likely to encounter problems in objectively evaluating each of the products. Therefore, a fractional factorial design was applied, based on product involvement and the sports celebrity. Questionnaire A included two pictures: firstly, a picture of a personal computer and the global sports celebrity, later a picture of water and the national sports celebrity. Questionnaire B presented a picture of water and the global sports celebrity at the beginning and then a picture of a personal computer and the national sports celebrity.

Participants
The data collection was performed using an internet survey by convenience sampling in Lithuania. According to previous research (Liu et al., 2007;Seimiene & Jankovic, 2014), the sample size was estimated at 250. An internet survey generated 305 responses from Lithuania, but some questionnaires had careless responses (Meade & Craig, 2012) -they were incomplete, some respondents filled questionnaires too fast, and some questionnaires had no variance in responses. Therefore 52 questionnaires were excluded and data analysis was based on 253 respondents: 143 respondents filled Questionnaire A and 110 -Questionnaire B. The respondents included in both samples did not differ by their gender (χ 2 (1) = .53, p=.47), age (χ 2 (2) = 2.356, p=.308) and income (χ 2 (3) = 1.509, p=.680) (see Table 2).

Measures
To be consistent with the previous studies, we used scales employed by scholars in the field. Measurement of sports celebrity credibility was based on the adapted Ohanian (1990) scale, which was used later by Liu and Brock (2011) and Dwivedi et al. (2016). The scale consisted of 15 statements related to three dimensions of credibility -attractiveness, expertise and trustworthiness. Cronbach alpha were 0.893 for the credibility of the national sports celebrity and 0.925 for the global sports celebrity, indicating a suitable degree of reliability. Measurement of perceived product quality was based on the Yoo, Donthu and Lee (2000) scale which consisted of four statements (Cronbach α=0.957 for the low involvement product and α = 0.965 for the high involvement product). Finally, a scale proposed by Capella, Hill, Rapp and Kees (2010) was used for the measurement of intention to purchase the products. The scale consisted of five statements and high psychometric property (Cronbach α=0.937 for the low involvement product and α = 0.924 for the high involvement product). On all scales, the respondents were asked to show the degree of their agreement with the statement on a 5-point Likert scale (1 -strongly disagree, 5 -strongly agree).

The impact of sports celebrity credibility on consumers' purchase intention
To examine relations between the credibility sports celebrities , perceived product quality and consumers' intention to purchase a product, we used the PROCESS macro (Hayes, 2013). Model 4 was employed to assess evidence for moderation, and Model 10 was applied to evaluate the moderated mediation effect. We tested the significance of the effects using bootstrapping procedures. Unstandardized indirect effects for each of 10,000 bootstrapped samples and the 95% confidence interval were computed. Analysis of the mediation effect of the perceived product quality showed that the credibility of a sports celebrity had a positive impact on the perceived quality (

The impact of moderators on the relationship between sports celebrity credibility and consumers' purchase intention
As one could expect, the evaluations of sports celebrity credibility were different. The national sports celebrity had lower credibility (M=3.92, SD=.59) compared to the global one (M=4.04, SD=.65) p=.031, which supports hypothesis H 3 , but differences were noticed within dimensions of credibility. Both sports celebrities had the same evaluation of their expertise, but the global celebrity (M=4.26, SD=.87) was more attractive than the national one (M=3.54, SD=.96) p<.001, while the national celebrity was perceived as more trustful (M=3.59, SD=.87) than the global one (M=3.23, SD=.92) p<.001 (see Table 3). As stated before, the impact of sports celebrity credibility on the consumers' intention to purchase could depend on such variables as the level of product involvement and localness of the celebrity. Since the mediation effect of perceived quality was proved, we had to evaluate the effect of moderation on both relationships: the direct impact of sports celebrity credibility on the intention to purchase and the impact of sports celebrity credibility on the perceived quality. Moderated mediation occurs when either Path a (from the impact of sports celebrity credibility on the perceived quality) or Path b (from the direct impact of sports celebrity credibility on the consumers' intention to purchase), or both are moderated (Edwards & Lambert, 2007).
Analysis of the moderated mediation effect indicated that the overall model was statistically significant (R 2 =0.085, F(5, 500)= 9.308, p<0.001) when the relationship between the impact of sports celebrity credibility and the perceived quality was evaluated. However, analysis of the moderation effect on the relationship between the impact of sports celebrity credibility and the perceived quality was not moderated either by localness of the sports celebrity (b=0.094, 95% CI [-0.213, 0.401], t=0.60, p=0.549) or by the level of product involvement (b=-0.063, 95% CI [-0.358, 0.233], t=-0.42, p=0. 678) (see Table 4 Part A). Thus H 4B and H 5B were rejected.  Analysis of the moderation effect of the direct impact of celebrity credibility on consumers' intention to purchase confirmed that the sports celebrity credibility and the perceived quality have a direct impact on purchase intention R 2 =0.322, F(6, 499)= 39.505, p<0.001 (see Table 4 Part B). Additionally, the level of product involvement also had a direct impact on purchase intention (b=-0.394, 95% CI [-0.554, -0.235], t=-4.86, p<0.001), which proves theoretical assumptions. Moreover, both moderators were important when the direct impact of the sports celebrity credibility on the consumers' intention to purchase was evaluated. The direct impact was moderated by both localness of the sport celebrity (b=-0.388, 95% CI [-0.634, -0.142], t=-3.10, p=0.002) and involvement of product moderated (b=-0.254, 95% CI [-0.498, -0.010], t=-2.05, p=0.041). Thus we support H 4A and H 5A .

Moderation of localness of a sports celebrity and product involvement on the mediated relationship between the sports celebrity credibility and the consumers' intention to purchase
Moderation analysis revealed a different story depending on the used variables. The sports celebrity credibility had a stronger conditional direct impact on the consumers' intention to purchase in the case of the national celebrity than the global celebrity (see Figure 2). Such a situation was noticed in the case of both the products of low and high involvement. The opposite impact of the sports celebrity credibility was observed on the consumers' intention to purchase depending on the level of product involvement. Growing credibility of the sports celebrity had a stronger positive impact on the intention to buy low involvement products than those of high involvement. Finally, the biggest difference in the impact was found due to the moderation of the two variables. Growing credibility of the national sports celebrity had a positive impact on consumers' intention to buy the low involvement product, while in the case of the global sports celebrity and high involvement product the relationship had the opposite direction.

Discussion
Sports celebrities are quite often used in advertisements, since it is believed in the direct impact of celebrity credibility on the consumers' intention to purchase a product. Such a relationship was found in several previous studies (Ahmed et al., 2014, Lee & Koo, 2015Nguyen & Huynh, 2018;Ong & Ong, 2015). On the other hand, many studies reported the impact of sports celebrities only on the perception of products or brands (Chang, 2014;Dwivedi et al., 2016;Renton, 2009;Srivastava, 2011). Therefore, the purpose of this research was to investigate the role of sports celebrity credibility on the consumers' intention to purchase a product. Initial analysis showed that the credibility had both a direct and indirect effect on respondents' intention to buy, which means that both groups of authors were right in their findings. In line with the authors who argued about the direct impact of celebrity credibility, we also found a strong direct impact. However, the analysis of perceived quality as a mediating variable also proved the existence of indirect effect. Such results were quite predictive and consistent with the existing theory. An additional aspect of the theory of celebrity endorsement was the localness of celebrities. The theory of a celebrity's capital supposed global sports celebrities should enjoy a higher level of credibility (Driessens, 2013;Carrillat & Ilicic, 2019). Current study proved that the global sports celebrity had a higher level of credibility. However, deeper analysis showed that a higher level of sports celebrity credibility does not mean a higher level in all dimensions. Global and national sports celebrities had the same and a bit higher than the average level of expertise. It means that being a global celebrity does not guarantee the perception of possessing a higher level of knowledge, skills, and abilities related to a product. Another dimension of credibility -attractiveness -was higher for a global sports celebrity. Since attractiveness is an athlete's physical appearance (Kunkel et al., 2018), it can hardly depend on the localness of the sports celebrity, but rather on a personality. However, global celebrities are known for most of the people from mass media, where the fashion industry makes them much more stunning, while national sports celebrities have fewer possibilities to use all benefits from the fashion industry and could be perceived as less attractive.
Contrary to attractiveness, trustworthiness was higher for the national sports celebrity than the global one as was noticed in a few previous studies (Dumitriu, 2018, Toncar, Reid & Anderson, 2007. Sports celebrities are often involved in scandals that could have some impact on his or her trustworthiness (Chang, 2018). Especially that could happen with global sports celebrities since we know them mainly from mass media or social media. Such results show the need for a more profound analysis of dimensions of the credibility of global sports celebrities in the native country, developed markets and emerging countries.
Just a few studies (Dwivedi et al., 2016;Keat, 2015) tested the differed impact of global/national celebrity on the perception of a product. Moderation analysis showed that the localness of a sports celebrity does not influence the relationship between credibility and perceived quality. In both cases (direct and moderating), we had the same correlation between sports celebrity credibility and the perceived quality of a product. Such results show that the sports celebrity localness has no moderating effect on the indirect impact of credibility on consumers' purchase intention. Besides, the results raise the question if a celebrity's capital could be easily transformed into economic capital according to Bourdieu's (1986) field theory.
The moderation analysis of direct effect showed that both moderating variablessports celebrity localness and product involvement, affected the celebrity' ability to stimulate the consumers' intention to purchase. The results showed that the relationship between the celebrity credibility and consumers' purchase intention exists when the national sports celebrity presents a low involvement product, but there is no relation in the case of the global celebrity. The opposite results were in the case of a high involvement product. Such findings could be explained using two routes of the purchase decision process -extended and routine. Consumers look for information about a high involvement product in various sources of information: technical characteristics on the website of a producer, experts' reviews and users' comments. Thus celebrities could add the symbolic value of being the same or because "they select the best". By contrast, consumers usually have to decide here and now about purchasing a low involvement product. Therefore celebrity trustworthiness could have a significant influence. Finally, ethnocentricity and anti-globalisation could work against global celebrities (Roy & Bagdare, 2015).
A new contribution to the existing knowledge added the product involvement -it mediated the direct impact of sports celebrity credibility on the consumers' intention to buy a product. The direct impact was noticed in the case of low involvement products and no impact was found in the situation of high involvement. Since low involvement products are purchased quite frequently and do not require complex cognitive thinking (Punyatoya, 2011;Roozen & Claeys, 2010), sports celebrities could have a direct impact on purchasing. Attractiveness of a sports celebrity could influence consumers' unplanned purchases, which supports Kahle and Homer (1985) conclusion that the visual impression of celebrity endorser has high importance. In the case of high involvement products, sports celebrities could increase the perceived quality of products and will have just an indirect effect on the consumers' intention to buy. Such results confirm the findings of previous studies (Al Zoubi & Bataineh, 2011;Punyatoya, 2011;Renton, 2009).

Theoretical and practical implications
This study contributes to a better understanding of consumers' decision-making in sports management literature in several ways. First of all, the results confirm the general understanding of the importance of sports celebrities in emerging as well as developed markets. At the same time, we expanded the general understanding by separating two possible ways for the impact of sports celebrities on consumers' purchase intentions: direct impact and through quality perception. Even though direct and indirect impact were almost equal, they depended on product involvement. The direct influence of sports celebrity credibility was found only for the consumers' intention to purchase a low involvement product. Therefore, product involvement should be taken into account for future research.
The main contribution of the study was related to the localness of sports celebrities. A growing amount of marketing campaigns that include national sports celebrities raises a question about their effectiveness compared with global athletes. The study found that the impact of a celebrity's credibility on perceived quality did not differ depending on global and national sports celebrities. However, the localness of the celebrity moderated a direct relationship between source credibility and consumers' purchase intention. Finally, the results of the study raise the question if a celebrity's capital could be converted to economic capital.

Limitations and future research
As with any study, the current study has several limitations demanding further clarification with future research. Previous authors (Belch & Belch, 2013) noticed that research in celebrity endorsement is product-dependent. The current analysis included two products with a different level of involvement, but both of them were goods. Therefore, future research could be based on services or services and goods. Also, the experimental design of the study included just categories of products instead of using products with brands. People have a stronger relationship to brands than a category. Thus, the usage of certain brands in research would make the situation closer to real life. However, the same brand could have a different image depending on a country which would add a question about the generalisation of findings.
Second, this study used only male athletes. On one hand, this allows us to have more comparable results since the main aim was to evaluate the impact of the sports celebrity's localness, but on the other side, that narrows the results. It is worth noticing that both sports celebrities have left active sports activity. Thus, longitudinal research on rising stars could add some light on the celebrity's capital theory and ability to accumulate that capital. Besides, analysis of active sports celebrities would let us evaluate their impact on different audiences like fans.
Third, this research compares the impact of global and national sports celebrities on consumers' purchase intention in one of the emerging and small markets in Europe. Future research should also consider conducting studies in large countries. In the case of the latter, it would be possible to evaluate the impact of global, national, or even local sports celebrities on consumers' purchase intention.