Source credibility is traditionally defined as the ability or willingness of a message source to provide reliable and truthful information (Kelman and Hovland 1953). Traditionally, source credibility has been conceived as consisting of two dimensions: source expertise and source trustworthiness (Mills and Jellison 1967; Rhine and Severance 1970). Source expertise refers to the extent of which a source is perceived of being knowledgeable on the message topic whereas source trustworthiness refers to the text of which the source is believed to make unbiased conclusions about a message topic. Thus, although expertise may affect a message source of making valid claims, trustworthiness is likely to influence a message source willingness to communicate valid claims (Hovland et al., 1953; McCracken, 1989). While most research does not manipulate source expertise and trustworthiness, source credibility research continues to underline the general effects of source credibility (O’Keefe 2002). It is likely, however, that assumed source expertise and trustworthiness could have different individual and situational factors (Moore, Hausknecht, and Thamodaran 1986; DeBono and Klein 1993; Pornpitakepan 2004). With regard to expertise, consumers have been found to not believe advertising messages when they do not perceive that they know what they are talking about i.e. that they are experts in the related area of the message that they are passing (Karmarkar and Tormala, 2010). To explain this, in the sales context, an expert salesperson generated far greater sales than the non-expert salesperson did (Woodside & Davenport, 1974). These propositions were confirmed by Speck, Schumann and Thompson  whose study found that celebrities with expertise to the endorsed product would develop significantly higher brand recall than a non-celebrity; however, the difference was not statistically important. More precisely, a positive attitude towards the endorsed product may be driven by more reliable information about the product provided by the expertise of the celebrity .
Trust has been recognised as an essential part in human relations, communication and marketing (Gambetta 1988; Luhmann 1980). Trust has been the subject of little systematic study in advertising, given the large amount of research in other disciplines. Credibility, a related trust construct, has been studied in advertising for some time, but literature argues that credibility and trust are separate constructs that have distinct conceptual characteristics (Doney and Cannon, 1997; Hovland, Janis, and Kelly, 1953; Ohanian, 1990; Swan et al., 1988). Soh, Reid, and King (2009) defined advertising trust as “confidence that advertising is a reliable source of product/service information and willingness to act on the basis of information conveyed by advertising”. The endorser has been shown to affect buyer purchasing decisions as it offers useful knowledge (Power et al., 2008). In an online environment where products cannot be touched or checked, the endorser communications important information on the product (Degeratu et al., 2000). Similarly, endorser trust beliefs or expectations regarding the actions of the endorser have been shown to affect consumer behaviours such as loyalty (Chaudhuri and Holbrook, 2001; Lau and Lee, 1999) and purchasing intentions (Delgado-Ballester et al . , 2003; Elliott and Yannopoulou, 2007) as they mitigate perceived uncertainties about the product (Power et al., 2008). Delgado-Ballester et al (2003) found that confidence in the credibility of the endorser had an effect on purchasing intentions. Lau and Lee (1999) consider that expectations regarding the trustworthiness and expertise of the endorser have an effect on repurchase intentions. Chaudhuri and Holbrook (2001) consider that the trust in the benevolence of the endorser has an influence on loyalty. Power et al., (2008) suggests that endorser trust has a positive effect on endorser attractiveness, which in effect, influences consumer purchasing intentions. Previous literature therefore provides support for the proposal that trustworthiness has a positive impact on purchase intentions.
Attractiveness is the extent to which one is physically pleasing and appealing. Studies such as Amos et al. (2008) and Wang and Scheinbaum (2017) have centred the term ‘attractiveness’ around ‘one who is viewed as beautiful, classy or elegant.’ According to McGuire (1969), the attractiveness of the source as perceived by consumers has a direct impact on the effectiveness of the advertising message. Stafford et al . (2002) notes that the likeability and approach of consumers towards advertising has an impact on the promotional results of celebrity endorsement. Praxmarer Decision (2009) observed the congeniality factor in the transfer of a favourable attitude towards the brand and found a direct relationship to purchasing intention. Thus, information sources that are highly likeable increase the effectiveness of the advertising message, as these sources increase the attention and recall of the message (Jain & Posavac, 2001, p. 179). Also, the effectiveness of physically attractive celebrities used to communicate advertising messages was supported concerning attitude change (Debevec and Kernan 1984), consumer behaviour and increasing purchase intentions (Petroshius and Crocker 1989; Petty and Cacioppo 1981; Tingchi Liu et al. 2007). This shows that the attractiveness of the endorser purchase intentions of the endorsed product (Kahle and Homer, 1985; Kamins, 1990). People like to be around attractive people because they are appealing to look at and because looking like them feels rewarding. Moreover, the positive characteristics of attractive people seem to ‘rub off’ those around them as a result of associational learning (Sigall and Landy, 1973) Physically more attractive celebrities are perceived as stronger, mentally healthy, sexually warm, intelligent and socially skilled compared to physically less attractive celebrities (Eagly, Ashmore, Makhijani, and Longo, 1991). Ohanian (1991) found that knowledgeable, honest and physically attractive celebrities are perceived to be trustworthy and thereby promote positive perceptual and behavioural reactions from consumers.
Marketers and academics (Patzer 1985) share the belief that physical attractiveness of endorsers has a significant influence on the effectiveness of advertising and marketing practices (Erdogan, 1999). However, studies have shown, that while attractive endorsers an effect on advertising evaluation (Baker and Churchill 1977), they usually have little or no impact on the purchasing intention (e.g. Baker and Churchill 1977).The unexpected findings from Caballero and Solomon (1984) study in which respondents were more receptive to low attractiveness when buying tissue, led Caballero, Lumpkin and Madden (1989) to use to use the reinforcement theory of attraction of Byrne (1971) which suggests that the effect of an attractive source would depend on the properties of the endorsed product. They observed the willingness of consumers to purchase a videotaped advertised product with three different levels of attractiveness. The authors were unable to establish a relationship between the properties of the product and the attractiveness, i.e. the association between the willingness to purchase and the level of attractiveness. The study considered low involvement and frequent purchases of coffee and cologne. In retrospect, this was considered to be a limitation, as the level of consumer transactions in these types of product categories may have had an impact on the findings. They suggested that the findings may differ in different attractiveness-related product categories such as cosmetics and fashion clothing which enhance physical attractiveness. According to a study by Kamins (1990) there is a correlation between attractiveness on the perceptions of advertising of attractiveness-related products. The study shows that when an attractive celebrity endorsed an attractiveness-related product, e.g. beauty enhancers or camouflage products, they were perceived as more credible compared an unattractive celebrity that endorsed the same product. Thus, physical attractiveness is a strong source of influence on people through their appearance and has a direct effect on their purchase intention.
Match-up Between Product and Endorser
The match-up hypothesis theory shows that the congruence between the characteristics of the celebrity endorser and the attributes of the product is critical in order to allow the correct transfer of meaning and increase the credibility of the source used to endorse the product (Erdogan, 2010). In line with this, celebrity endorsement has a significant effect on consumer attitude towards advertising when there is a congruence that is consistent between the endorser and the product being endorsed. Congruence between celebrity and endorsed product is considered to be a key factor of celebrity endorsement effectiveness (Erdogan, Baker, & Tagg, 2001; Kahle and Driver, 1985; Kamins, 1990; Kamins and Gupta, 1994). Friedman and Friedman (1979) suggested that the influence of celebrities on advertising differs depending on the type of product being endorsed. They concluded that a stronger match-up between celebrity and product fit, as perceived by consumers, leads to a higher degree of brand evaluation and a higher level of endorsement effectiveness. The degree of congruence between the celebrity endorser and the product has an influence on consumer perception towards the brand and celebrity endorser (Till and Busler, 2000). For example, a celebrity athlete was most effective as an endorser for the energy bar that enhanced athletic performance compared to a non-athletic product unrelated to athletic performance. This suggests that the target audience will only react to the advertisement if the image of the celebrity endorser matches strongly with the image of the brand or product. This is because the celebrities selected convey a message that suggests a higher perceived similarity between the elements selected for their self-image and the product. In support of this statement, the research by Choi and Rifon (2007) found that a strong match-up between the endorser and the product appeared to give positive attitudinal and behavioural responses from the target audience.
Kanungo and Pang (1973) conducted a preliminary study to determine the perceived physical attractiveness of a group of male, female and male-female endorsers with different product types. The study found that the male model for the car produced a more favourable attitude towards the product than the female model, while the male model for the sofa produced an unfavourable attitude towards the product than the female model (Kanungo and Pang, 197). The physically attractive celebrity endorsers were liked more and had more positive impact when the product was related to a and attractiveness such as cosmetics (Choi and Rifon, 2012; Kahle and Homer, 1985). This is in contrast to technology-related products which were less likely to have a similar effect due to the insignificance of physical attractiveness to the product type (Choi and Rifon, 2012; Till & Busler, 2000). In addition, to the findings of Kahle and Homer (1985) on product-endorser fit, Kamins (1990) noticed that a physically attractive endorser was less effective at endorsing an unrelated product to their area of expertise even when the product was attractive. In addition to attractiveness, Till and Busler (2000) considered expertise to be another potential match-up element influencing the source-object congruence effect. Two very similar design and method studies were conducted to test the hypotheses. The study looked at the congruence effect of physical attractiveness on advertising effectiveness, while the second study used expertise as a match-up factor. This study did not identify the interaction effect of the physical attractiveness by product type as opposed to previous research (e.g. Kahle and Homer, 1985; Kamins, 1990). In addition, the second study demonstrated the importance of expertise as a match-up factor. The study uncovered that the perceived expertise of the endorser was more relevant than the perceived attractiveness for matching a product to achieve greater endorsement effectiveness.
Match-up Between Endorser and Consumer Self-image
Previous research on the match-up hypothesis has primarily focused on celebrity endorsers and the product being endorsed (Kamins & Gupta, 1994). More recent research, however, has shown that the effect of congruence between celebrity endorsers and consumer self-image is based on consumer personality and self-concept. According to Escalas (2004) consumers are able to create a relationship with the brand and develop a self-brand connection. Self-brand connection suggests that, as a person purchases a brand, they integrate the brand their self-concept, thus incorporating the brand as part of their extended self (Escalas & Bettman, 2005). A strong self-brand connection is more likely to exist where the personal experience of the consumer with the brand is directly linked to the identity of the brand and the brand itself meets their psychological needs (Moore & Homer, 2008). The self-concept of consumers is a key factor in shaping their consumption behaviours. Self-image congruence allows marketers to anticipate various aspects of consumer behaviour (Hosany & Martin, 2012; Quester, Karunaratna, & Kee, 2000). Consumers are more likely to be persuaded to buy the endorsed product if their self-image matches that of the celebrity endorser (Hosany & Martin, 2012). As Choi (2012) points out, a strong match between the celebrity image and either the self-image of the endorsed product or the consumer is more effective than a weak match between them. Consumer self-concept has been conceptualised from a multidimensional perspective (Burns 1979; Rosenberg 1979). Previous research identified four dimensions of self-concept to study and forecast consumer behaviours: (1) real self-concept, how a individual perceives himself or herself, (2) potential self-concept, how an individual would like to view himself or herself, (3) external self-concept, how customers believe others interpret them, and (4) potential external self-concept, how a person would like to be viewed by others.
According to Sirgy (1982), self-concept can be categorised into four (4) dimensions: self-concept (how an individual perceives themselves), ideal self-concept (how an individual would like to perceive themselves), social self-concept (how an individual thinks others perceive them), and ideal social self-concept (how an individual would like to be perceived by others; Noble & Walker, 1997; Schiffman, Bednall, O’Cass, Paladino, & Kanuk, 2005; Sirgy & Su, 2000). In recent research, the two dimensions of self-concept that have received a vast body of empirical support and theoretical consideration are actual self-concept and ideal self-concept (Graeff, 1996). Accordingly, Sirgy (1982; 1985) defined self-image value under two dimensions: (1) actual self-image and (2) ideal self-image. Self-image congruence is the fit between consumers’ self-concept (e.g., actual self, ideal self, social self and ideal social self) and brand personality (Aaker, 1999, Sirgy, 1982). Self-congruence, self-image congruence, image congruence and self-congruity are used much in the same way to explain this congruency effect (Hosany & Martin, 2012; Sirgy, 1985). Self-image congruence is a strong factor influencing consumer behaviour (Kressmann, Sirgy, Herrmann, Huber, & Lee, 2006; Sirgy, Johar, Samli, & Claiborne, 1991; Hosany & Martin, 2012; Sirgy & Samli, 1985, 1991) For example, prior research has shown that self-image congruence influences consumers’ choice (Hosany & Martin, 2012; Quester et al., 2000), perceived quality (Hosany & Martin, 2012; Kwak & Kang, 2009), brand loyalty through functional congruity (Hosany & Martin, 2012; Kressmann et al., 2006), brand preferences (Hosany & Martin, 2012; Jamal & Goode, 2001), attitudes and consumer buying behaviours (Hosany & Martin, 2012; Ibrahim & Najjar, 2008), contributes to more favourable attitudes toward product and brands (Ekinci & Riley, 2003; Hosany & Martin, 2012; Sirgy et al.,1997) and affects advertising effectiveness (Bjerke & Polegato, 2006; Hong & Zinkhan, 1995; Hosany & Martin, 2012).
Langner et al. (2013) suggested that consumers show more positive attitudes and stronger purchase intentions towards brands that they feel are compatible with their self-image and show negative attitudes towards brands that they feel are not incompatible with their self-image. According to Park et al. (2010) a brand is also judged on the grounds of whether it represents itself or how important the brand is to the customer. Graeff (1996) suggested that there are two types of self-brand congruence; self-congruity and ideal congruity. Self-congruity is a match between the image of the brand and the actual self-image of the consumer, while ideal congruity is a match between the image of the brand and the ideal self-image of the consumer (Graeff, 1996). More recently, brands serve not only their primary roles, but also their psychological and social objectives as indicators for consumers to communicate their identity and to facilitate social interaction with others (Aaker, 1996; Belk, 1988; Choi & Rifon, 2012; Sirgy, 1982). Sirgy et al. (1997) suggested that the images consumers hold of various products interact with their self-concept and generate a self-image congruence during the consumer buying process. A celebrity endorser can influence the self-construction of the consumer through this process. In addition, McCracken (1989) suggested that consumers develop their own perceptions of the product when buying a celebrity-sponsored product and use it to form a positive self-concept of themselves. In summary, self-image congruence can influence a number of consumer behaviours, such as brand preferences (Hosany & Martin, 2012; Jamal & Goode, 2001), attitudes and buying behaviours (Hosany & Martin, 2012; Ibrahim & Najjar, 2008). Consumers pursue self-esteem by enhancing their ideal self-image and finding celebrities with inspirational personalities with positive associations to enhance their ideal self (Boon & Lomore, 2001; Choi & Rifon, 2012; Escalas & Bettman, 2003).