Advertisers in pursuit of endorser’s sports results : the impact of sports celebrity underperformance on consumer attitude toward a brand

(1)

Graduate school of Communication

Persuasive Communication

Master Thesis

Advertisers in Pursuit of Endorser’s Sports Results:

The Impact of Sports Celebrity Underperformance on Consumer Attitude

toward a Brand

Author: Katarína Železná

Student ID: 10864091

Supervisor: dr. B.C. Barbara Schouten

(2)

Abstract

Many studies (e.g. Carrillat, D' Astous, & Christianis, 2014; Fong & Wyer, 2012; Thwaites, Lowe, Monkhouse, & Barnes, 2012) have been dedicated to ascertain what influence negative news, such as scandals in celebrity life, may have on the brand which a celebrity endorses. However, having considered the previous studies, there is one exception; researchers did not take into account another type of negative news – the underperformance of a sports celebrity. Due to the fact, that underperformance of famous athletes is almost relentlessly followed by the media, and often negatively framed; it can potentially have a deleterious effect on the consumer’s perception of sports celebrity characteristics (credibility and attractiveness) and on the consumer’s attitude toward a brand which the celebrity

endorses. Both of the previously mentioned concepts have received great attention from researchers and have been established as the determinants of advertising effectiveness. Therefore, this experimental study focuses on the impact of the sports celebrity

underperformance negatively framed by media on consumer’s perception of sports celebrity

characteristics (attractiveness and credibility) and the brand attitude of the product which the celebrity endorses. This study complements previously conducted study by Koo, Ruihley and Dittmore (2012) and investigates the issue further by actually using a real-life sports celebrity who has competitive experience. It expands on this by taking the consumer’s level of

identification with the sports celebrity, and the customer’s level of involvement with the

endorsed product into account.

The experimental study comprised of 117 Czechs who were either members or non- members of athletic clubs. It revealed that news about an underperformance of a sports celebrity that is negatively framed by the media have a deleterious influence on a consumer’s

perception of a sports celebrity credibility. However, the brand attitude of the consumers was not influenced by their perception of sports celebrity characteristics. Therefore, it is

(3)

ADVERTISERS IN PURSUIT OF ENDORSER’S SPORTS RESULTS: THE IMPACT OF SPORTS CELEBRITY

UNDERPERFORMANCE ON CONSUMER ATTITUDE TOWARD A BRAND 3

concluded that an underperformance of the sports celebrity negatively framed by media has a negative impact on consumer’s perception of sports celebrity credibility.

Keywords: sports celebrity credibility, sports celebrity attractiveness, brand attitude,

(4)

Sprint superstar Asafa Powell was in an advertisement for Nike sports shoes and the famous tennis player Roger Federer has endorsed a Wilson tennis racquet. Sports celebrities are not only visible at stadiums, football fields or swimming pools where they fight for the medals, but they often surround us on posters in bus stations or in television commercials. Practitioners became aware of famous athlete’s persuasive power and have made celebrity endorsement the most popular form of retail advertising (Choi & Rifon, 2007 in Thwaites et al., 2012). In 2010, American companies spent almost $17.2 billion on sports celebrity endorsement (IEG, 2011 in Koo et al., 2012) and the usage of celebrity persuasive powers in advertising is increasing (Thwaites et al., 2012). The big corporations are willing to pay millions of dollars to their endorsers because of the belief that the image of the celebrity will spill over to their brand and pay back in terms of dividends in a corporation’s revenue. This assumption has been supported by several studies which revealed that celebrity endorsement is more effective than other forms of advertising (e.g. Atkin & Block, 1983; Friedman, Termini & Washington, 1976) as it positively influences a consumer’s perception and purchase intention (Goldsmith, Lafferty, & Newell, 2000).

One of the reasons why the celebrity endorser enhances the effectiveness of advertising is rooted in the fact that celebrities are in general perceived as having more expertise and being more trustworthy than other type of endorsers (Thwaites et al., 2012). The endorser’s characteristics such as credibility (Goldsmith et al., 2000; Koo et al., 2012; Lafferty & Goldsmith, 1999), attractiveness (Atkin & Block, 1983), likeability (Kamins, 1989), and popularity have a positive influence on attitude towards the brand. However, does celebrity endorsement bring advantages under any circumstances and what happens when the media start to report negative news about the career of the celebrities?

In an ideal world of advertising, the celebrity endorser would be a star without any possibility to gain a negative reputation. However, with the growing interest of the media

(5)

(Skolnick, 2008 in Lewis & Weaver, 2015) and fascination of the public about the life and career of celebrities, the opposite is true. Any immoral or illegal action which occurs in the life of a celebrity can and has been negatively portrayed by the media. Previous studies (Carrillat, d’Astous, & Lazure, 2013; Thwaites et al., 2012) proved that the credibility and attractiveness of the celebrity endorser can be lowered by the occurrence of negative news such as a scandal in a celebrity’s life. Furthermore, a decrease of the celebrity endorser’s credibility and attractiveness can lower viewers’ attitude toward a brand that is endorsed by the celebrity (Kahle & Homer, 1985). For instance, the cheating scandal of one of the most famous golf players, Tiger Woods, lowered Nike’s profit in golf ball sales by $1.5 million (Chung, Derdenger, & Srinivasan, 2013). This study will focus on different type of negative news in the sports career of the celebrity endorser which is filtered by media –

underperformance of the athlete at world-known competitions. The rationale for doing so, is outlined below.

Sports celebrity underperformance can have a deleterious influence on an athlete’s professional career and reputation due to the media. Because of the sports celebrity’s fame, the media immediately start to report news about underperformances. This media output can be negatively framed such as in the case of the competition collapse of Australian rower Sally Robbinson at the 2004 Olympic Games in Athens. She was immediately narrativised as a ‘choker’, ‘quitter’ and ‘UnAustralian’(McKay & Roderick, 2010). The way news about an

underperformance of a sport celebrity is framed and the values which are implemented in the message can directly negatively affect its recipients’ attitudes toward a portrayed person (Price, Tewksbury, & Powers, 1997 in Lewis & Weaver, 2015). In other words, media coverage of an athlete’s underperformance and the framing of the messages can influence how and what its readers think about the sports celebrity as an athlete and a person, and possibly lower his or her attractiveness and credibility.

(6)

Only one piece of research done by Koo et al. (2012) has investigated how the underperformance of a sports celebrity at world-known competitions can influence a consumer’s perception of the brand which the celebrity endorses. Such an investigation is

especially important due to the fact that the popularity of the sports celebrity is at first gained by an athlete’s on-field performance. The results of previously mentioned study (Koo et al., 2012) indicated that athlete’s underperformance negatively affects consumer’s perception of sports celebrity credibility and brand attitude which the athlete endorses.

However, research conducted by Koo et al. (2012) differs from the proposed

experiment in several aspects. First, Koo et al. (2012) used "fake" sports celebrities in order to fabricate large differences in source credibility. Such artificially extreme manipulation could result in a lack of realism and thereby decrease the external validity of its results (Goldsmith et al., 2000 in Koo et al., 2012). Therefore, this research will use a real sports celebrity with real on-field performance results which will add to the validity and

generalizability of the previous findings. Second, instead of using university students this study will focus on a specific population of members of athletic club, who are potential customers of sports brands and due to their level of identification with the athletes they may have different perception of sport celebrity underperformance than students. As Kelman (1961) proposes, identification occurs when an individual adopts the behaviour or attitudes of another individual or group based on a self-defining relationship. Cohen (2001) mainly focuses on identification with characters that appear in the media, such as sports celebrities, and defines identification as "a mechanism through which audience members experience reception and interpretation of the text from the inside, as if the events were happening to them" (p. 245). Previous research revealed that consumers’ who strongly identify with the

(7)

sports celebrity are, in the event of a scandal in the life of celebrity, less influenced1 by such negative news than individuals with a weak identification (Johnson, 2005). Third, the study incorporates the level of consumer identification with the sports celebrity and product involvement as a moderator.

To sum up, this research investigates whether an underperformance of the sports celebrity at world-known competitions that is negatively portrayed in the media has a

negative impact on the sports celebrity’s credibility and attractiveness. Such an investigation is especially important as both of the celebrity characteristics have been established by many researchers (Debevec & Kernan, 1984; Maddux & Rogers, 1980; Ohanian, 1990 in Lafferty, 1999; Tellis, 2003 in Lee & Koo, 2015) as determinants of advertising effectiveness.

Furthermore, the underperformance of sports celebrity is a common part in the sports career of the celebrity and is relentlesly followed by the media. Next, it explores whether a negative perception of sports celebrity credibility and attractiveness has an impact on consumers’ attitude toward the brand that the sports celebrity endorses. Brand attitude has been defined as "consumer's overall evaluation of a brand" (Keller, 2003; Wilkie, 1990 all in Martensen, Gronholdt, Bendtsen, & Jensen 2007, p. 285). Due to the fact that a consumer’s attitude towards a brand partially predicts the consumer’s behaviour towards a product, therefore, this is an essential concept for measuring advertising effectiveness (Olson & Mitchell, 2000).

In addition, this study explores whether the relation between underperformance that is negatively framed by the media and sports celebrity credibility and attractiveness will differ based on a consumer’s level of identification with the sport celebrity.

1_{in the sense of purchase intention of the product which the celebrity endorses, recommendation of the}

product, belief in the innocence of the celebrity and emotional responses such as feeling proud for being the celebrity’s fan.

(8)

Lastly, the processing of advertising information is influenced by a consumer’s level of product involvement (Petty, Cacioppo, & Schumann, 1983). Chen and Leu (2011) define product involvement as a concept which "represents consumers’ cognitions or subjective feelings about the importance of a product" (p. 154). Previous studies have suggested that the presence of a celebrity endorser acts as a peripheral cue and is more effective with less

involved consumers (Petty et al., 1983; Pughazhendi, Baskaran, Praskash, & Balamurugan, 2012). Therefore, this study will incorporate the level of a consumer’s involvement with the product as a moderator of impact of sports celebrity characteristics (credibility and

attractivity) on the consumer’s evaluation of brand attitude.

To conclude, the following research question is proposed: To what extent does a current athletic underperformance by a sports celebrity at a world-known competition, which is negatively reflected in the news, influence a viewer’s attitude toward the brand that this sports celebrity endorses, and is this effect mediated by a sports celebrity’s levels of

attractiveness and credibility? Is the influence of athletic underperformance that is negatively reported in the media on a celebrity’s credibility and attractiveness moderated by a

consumer’s level of identification with the celebrity? And does the influence of sports celebrity credibilty and attractiveness on brand attitude differ based on the consumer’s level

of involvement with the product which the celebrity endorses?

Due to the fact that underperformance of sports celebrities at world-known

competitions is more common than the occurrence of scandals, and can potentially have a deleterious influence on the brand which the celebrity endorses (Koo et al., 2012), it is in the interest of brand managers and advertising agencies alike to address this research gap. This study aims to bring empirical findings that can help brand managers and advertising agencies to understand how negatively-framed underperformances of sports celebrities at world-know competitions can affect the credibility and attractiveness of sports celebrities, and shape

(9)

consumers’ attitudes toward the brand which the sports celebrity endorses. Such information

should help brand managers and advertising agencies to decide about the usage of sports celebrities in their daily advertising practice, and give them a hint of how to act when sports underperformance of an athlete (as endorser of their brands) occurs.

Considering that only one study (Koo et al., 2012) investigated the effects of underperformance of the sports celebrity on consumer’s perception of sports celebrity

characteristics and attitude toward a brand which the celebrity endorses, this research aims to fulfil the research gap. Furthermore, by incorporating real sports celebrities it complements previously mentioned study of Koo et al. (2012) and increases external validity of the results. Lastly, by including consumer’s level of identification with the sports celebrity this

experimental study expands on previously conducted research (Johnson, 2005; Um, 2013) that investigated the influence of consumer’s identification with sport celebrities on the

perception of scandals (negative news).

Theoretical Background

Sports Performance and Its Impact on Celebrity Credibility and Attractiveness

The career of the sportsman is a spiral going up and down, with defeats and victories. Even the most famous sportsmen who are considered as the greatest in their field know how it feels to lose or fail. When such a disappointment in the career of a top athlete occurs, the media immediately trumpet the sports celebrity’s underperformance. The way the journalist informs the public about it, and what attributes he puts into the story depends on his selection of the frame (Entman, 1993; Nelson, Oxley, & Clawson, 1997).

The news about the underperformance of the sports celebrity can be very negatively framed, for instance by expressing disappointment, lack of sports confidence, and expertise. An example of negative framing could be seen in the news of an American sports network,

(10)

which reported about Australian golfer Greg Norman. Even though Norman had won two British Open golf titles he was still named by this broadcaster as the "worst choke artist of all

time" when he lost three other major championships in spectacular fashion (ESPN, 2007 in

McKey & Roderick, 2010).

On the other hand, when the sport celebrity performs extraordinary well or wins at a world-known competition, such as the Olympic Games in athletics, the World Cup in football, or the Grand Slam in tennis, the achievement of the celebrity is "glorified" in the media and the positive characteristics of the sports celebrity are empowered and amplified. After her second winning performance at the Olympic Games, the javelin thrower Barbora Špotaková was narrativized as a "champion" and "sovereign queen" and her mental strength and serenity were emphasized (Železná, 2014).

Previously mentioned examples point out that frames work on the basis of selection and salienece as they make some aspects more noticable, meaningful, appealing or

memorable to the receiver (Entman, 1993). In other words, frames call attention to certain aspects, while driving away attention from other aspects/attributes (Entman, 1993). Fiske and Taylor (1991 in Entman, 1993) explain that it is this salience of certain information which increases the probability that the recipients will perceive the information, discern meaning and as a result of this action they will more readily process such information and store it in memory.

Previous studies have proved that frames are successful in changing public opinions (e.g., Berinsky & Kinder, 2006; Lecheler,Vreese, & Slothuus, 2009; Nelson et al., 1997). In a series of experiments, Kahneman and Tversky (1979, 1984 all in Lewis & Weaver, 2015) demonstrated how individuals make decisions when similar information is presented as positive or negative frames. They revealed that by exposing the recipients to differently framed scenarios containing otherwise identical information, the subsequent evaluations and

(11)

UNDERPERFORMANCE ON CONSUMER ATTITUDE TOWARD A BRAND 11 judgments of that information were affected and demonstrated how frames direct attention toward specific attributes of content and how responses to that content can be influenced by the framing choices of its producer (Lewis & Weaver, 2015).

The effect of frames on the change of recipients’ beliefs and attitude can be decsribed in terms of an expectancy value model of attitudes (Ajzen & Fishbein, 1980; Anderson, 1981 all in Nelson et al., 1997). Framing changes attitude based on their influence on the

recipients’ weight of the information (Nelson et al., 1997). In other words, frames do not add something new to the receivers’ beliefs about certain issues, but they exert their influence and subsequent attitude change by the activation of the individuals’ existing beliefs and

cognitions (Nelson et al., 1997). Therefore, the framing effects do not depend on whether the receiver accepts the assertions in the message or not. An attitude change can occur even for assertions with which the recipients does not agree with as the frame may still make salient points to one's beliefs (Nelson et al., 1997). The fact that frames do not acquire new

information for attitude change make them different from other popular ways of changing opinion and attitude such as classical conditioning, mood manipulation and subliminal suggestion, etc. (Nelson et al., 1997). Nelson et al. (1997) describe the reason of this difference as follows: "frames operate by activating information already at the recipients‘disposal, stored in long term memory" (p. 1).

However, it is important to note that the effect of frames always depends on the recipient of the message as "the frames that guide the receiver‘s thinking and conclusion may or may not reflect the frames in the text and the framing intention of the communicator" (Entman, 1993, p. 52 – 53). In other words, certain aspects of the text can be highly salient for the recipient if they coincide with any existing schemata2 in a recipient’s belief system

2

As Entman (1993) explains, schematas "connote mentally stored clusters of ideas that guide individuals‘ processing of information " (p. 53).

(12)

(Entman, 1993). On the other hand, due to the receiver’s existing schemata it can be difficult

to notice, interpret or remember the information emphasized in the text (Entman, 1993). Therefore, it is proposed that the way news about the underperformance of sports celebrity is framed and the values which are implemented in the message can directly affect the recipients’ opinion about a portrayed person and attitude toward him/her (Price et al., 1997 in Lewis & Weaver, 2015). In other words, media coverage of sports celebrity underperformance and the framing of the messages can influence how and what its readers think about the performance of the sports celebrity and him/her as a person, and framing can possibly alter a viewer’s perception of the sports celebrity’s characteristics such as the athlete’s attractiveness and credibility. The following paragraph will explain why these characteristics can be influenced by the (under)performance of the sport celebrity at world-known competitions.

Sports Celebrities’ Credibility

Sports celebrities develop their (brand) status based on their continued excellence in their sport (Andrews & Jackson, 2001 in Arai, Ko, & Ross, 2014). They acquire their fame and characteristics such as credibility and attractiveness at first by their on-field performance. As Ohanian (1991) explained the credibility of the endorser has three dimensions – expertise, trustworthiness and attractiveness. Expertise is defined as "the extent to which a

communicator is perceived to be a source of valid assertions" (Hovland, Janis, & Kelley, 1953 in Koo et al., 2012, p. 21) and refers to the knowledge, experience or skills possessed by an endorser (Erdogan, 1999). In the field of sport, expertise is often determined by an

athlete’s on-field performances (Arai et al., 2014; Koo et al., 2012). Arai et al. (2014) have

stressed that athletic expertise includes an athlete’s individual sports achievements and capabilities (e.g., win record, skills, and proﬁciency in their sport). For example, when Usain Bolt won three Olympic medals while running in a certain type/brand of spikes, consumers

(13)

UNDERPERFORMANCE ON CONSUMER ATTITUDE TOWARD A BRAND 13 may have considered him to be an expert because he knows about these spikes, but also because he has won major competitions using these endorsed spikes (Koo et al., 2012). Therefore, we assert that greater athletic success correlates with an individual having a more positive perception of sports celebrity expertise.

The second dimension, of credibility is trustworthiness, and it can also be potentially influenced by sport celebrity performance. Trustworthiness refers to honesty, integrity, believability and dependability of the endorser (Erdogan, 1999). In other words it reflects "consumer’s confidence in the source for providing information in an objective and honest manner" (Ohanian, 1991, p. 47). During the run-up to world-known competitions, such as the Olympic Games, the media filter out and subsequently place certain expectations on the sports celebrity’s’ performance (Summers & Morgan, 2008). The viewer’s trust and perception of an athlete’s performance can spill over toward the sports celebrity as an endorser of the sports product. For example, when the media reports about sports celebrity performance or about a disappointment and a failure, and a viewer accepts this perception or holds this perception on his own, such an evaluation can spill over towards the sports

celebrity as an endorser of a certain sport product. Similarly, if the sportsman gives different explanations about his underperformance or tries to give a justification which the viewer perceives as an excuse, a viewer’s perception of sports celebrity trustworthiness could be lowered. This assumption was confirmed in a study conducted by Koo et al. (2012) which revealed that an athlete with underperformance in his career is perceived as less trustworthy.

The last dimension of celebrity credibility – attractiveness – is connected to the athlete’s physical appearance (Ohanian, 1991) and therefore we do not suppose that this

component would be influenced by the underperformance of a sports celebrity at world-known competitions that are negatively portrayed by media. However, due to the fact that the previous two dimensions of sport celebrity credibility – expertise and trustworthiness – are

(14)

perceived to be influenced by such news, the overall credibility of a sports celebrity can be changed. Therefore, the first hypothesis is offered:

H1a: Consumers who read news about underperformance of the sports celebrity negatively framed by media will have a lower perception of sports celebrity credibility than consumers who read news about the victory of the sports celebrity positively framed by media.

Sports Celebrities’ Attractiveness

The source attractiveness model does not primarily focus on the physical

attractiveness of the endorser but attractiveness is rather viewed more generally as a positive attitude towards the endorser/sports celebrity (Silvera & Austad, 2004) which is caused by admiration or perceived similarity of the sports celebrity (McGuire, 1985 in Silvera & Austad, 2004). In such a case, the overall attractiveness of the endorser is dependent on the extent to which the viewer perceives an endorser as similar, familiar and likeable (McGuire, 1985 in Erdogan, 1999).

Likeability is defined by Erdogan (1999) as "affection for the source as a result of the source’s physical appearance and behaviour" (p. 299) and refers to the extent to which one is

considered as friendly, sincere, likeable, warm, charismatic and appealing (Yilmaz, Telci, Bodur and Iscioglu, 2011). Braunstein and Zhang (2005 in Arai et al., 2014) identiﬁed

likeable personality as one of the "athletic star power factors". This claim was also supported

in the study conducted by Giuliano, Turner, Lundquist and Knight (2003 in Guest & Cox, 2009) which emphasizes that admiration (liking) of the celebrity can be connected to an athlete’s success in their sports field as well as an athlete‘s personal qualities. In other words, whether or not a person will feel affection for a sports celebrity and will admire him depends on various professional celebrity traits such as great technical skills and physical qualities and personal traits such as temperament, decision making and identity (Fleming, Hardman, Jones,

(15)

UNDERPERFORMANCE ON CONSUMER ATTITUDE TOWARD A BRAND 15 & Sheridan, 2005). Specifically, sports celebrities who are perceived as hard-workers, being a good person, being a team-leader, and having self-confidence are admired (Fleming et al., 2005; Giuliano et al., 2003 all in Guest & Cox, 2009). Professional traits can be directly influenced by the successfulness of a sports celebrity, while personal traits can be influenced by athletic performance3, statements from the celebrity about their performance, as well as the analysis of such a performance by the media4.

The second dimension of the attractiveness – similarity – can be also influenced by the performance of sports celebrity. Similarity is defined by Erdogan (1999) as "a supposed resemblance between the source and the receiver of the message" (p. 299) and represents attitudes or judgements that people make about the character of the celebrity based on their perceptions of the athlete’s character and of them as people (Cohen, 2001). Thus people will

evaluate the behavioural expressions of the athlete at the stadium (e.g. reaction to the

victory/underperformance, not shaking the competitors hand after the football match etc.) as well as their statements after the match that are published in the media, and they will compare them to their own characteristics. Furthermore, it has been found that similarity is directly connected to the parasocial relationship (Rubin & McHugh, 1987 in Bui, 2015) with the celebrity. Previous research stressed that people had developed stronger relationships with celebrities who have a high ability to perform their specialised task such as acting, dancing, singing, writing etc. (Rubin & McHugh, 1987 in Bui, 2015). In other words, the quality of a (sports) celebrity performance can influence the perception of the similarity between the celebrity and the viewer.

The last dimension of attractiveness – familiarity – refers to consumer’s "knowledge of the source through exposure" (Erdogan 1999, p. 299). Due to the fact that this dimension is

3_{For example, consumer’s perception of self-confidence of a high jumper who always under jumps the} bar.

4

For instance statements from the journalist, such as that the athlete had lack of self-confidence, was rash or was not capable to lead the team etc.

(16)

connected to consumer’s knowledge of the sport celerity, we do not suppose that this

component would be influenced by one sport celebrity underperformance and its media framing. However, due to the fact that the previous two dimensions of sports celebrity attractiveness – likeability and similarity – are influenced by news about underperformance of sport celebrity negatively framed by media, the overall attractiveness of the sports celebrity can be changed. Therefore, the following hypothesis is offered:

H1b: Consumers who read news about an underperformance of a sport celebrity that was negatively framed by media will have a lower perception of the athlete’s attractiveness than consumers who read news about the victory of a sports celebrity that was positively framed by media.

The Influence of Receiver’s Identification with the Sports Celebrity

People know sports celebrities because of the media. They create their opinion about the celebrities and build an attitude toward them based on the picture that the media provides them with. In most of the cases they are not capable of gaining an impression about the celebrity based on their own real-life experience with the sports celebrity and therefore they are dependent on the information which the media provides them with, such as interviews with the celebrities, broadcasting of the competition, and commentating on the athlete’s performances and feelings.

When the individual is exposed to media personalities over time, it is believed that he will develop a sense of intimacy and identification with that celebrity (Horton & Wohl, 1956 in Basil, 1996). Livingstone (1998 in Cohen, 2001) describes identification as "imagining being in someone else’s shoes and seeing the world through his or her eyes" (p. 250). A

similar explanation was offered by Cohen (2001) who points out that identification is about sharing the same perspective of the character, and feelings of that character.

(17)

UNDERPERFORMANCE ON CONSUMER ATTITUDE TOWARD A BRAND 17 How important the celebrity is to the person and to the person’s self-identity is

determined by the extent to which he identifies with the celebrity (Um, 2013). In other words, the relationship between the celebrity and its observer determines the level of identification (Um, 2013). "Consumers who identify with a celebrity deﬁne their identities based in part on being a fan of the celebrity. The strength of identiﬁcation depends on how important being a fan of the celebrity is to the consumer’s identity" (Johnson, 2005, p. 100).

An individual’s level of identification with the celebrity influences how the person will perceive the celebrity and her/his characteristics. For example, Tzoumaka, Tsiotsou and Siomkos, (2014) demonstrated that identification with the sports team has a significant positive effect on the expertise and trustworthiness of the sports celebrity who is part of that team. Another study conducted by Friedman and Friedman (1979) proved that the

identification of the viewer is related to the attractiveness and likeability of the celebrity endorser. Individuals who highly identify with a celebrity are more likely to adopt her/his thoughts, feelings and behaviour compared to those who identify only a little with the celebrity (Um, 2013).

Identification with the celebrity is one of the factors which influence how the individual will react when he obtains negative news about the celebrity. A study conducted by Johnson (2005) revealed that consumers who strongly identify with a celebrity are less likely to react negatively to bad press, which for example occurs in the case of celebrity association with scandal, compared to consumers with weak identification. The findings showed that individuals who were highly identified with the celebrity took a point of view in favour of the celebrity, as they did not believe the person they admire could be connected with the presented scandal and felt proud to be a fan of the celebrity (Johnson, 2005). Contrary to this, people whose identification with the celebrity was weak felt guilty and ashamed about being connected to the celebrity in any way and did not believe in the

(18)

celebrity’s innocence (Johnson, 2005). Such a unique study has pointed out the importance and influence of an individual’s level of identification with the celebrity when the person is presented with bad news about the object of his admiration.

Similarly, studies (e.g., Nelson et al., 1997; Lecheler, et al., 2009) which investigated the influence of the frames on the receiver’s change of opinion/attitude toward the object of the message highlighted the importance of the receiver’s characteristics. One of the reasons why individual differences should be taken into consideration is the fact that the receiver of the message could have pre-existing meaning structures, or schemas about the object of the article which affect how people process and interpret media information (Raney, 2004; Scheufele, 1999 all in Lewis & Weaver, 2015). "Schemas serve as a network of associations between different aspects of a certain stimulus, influencing our perceptions, understanding, and interpretations" (Lewis & Weaver, 2015, p. 222-223). Furthermore, individuals who are highly identified with the celebrity invest more time into creating a relationship with the celebrity (Boon & Lomore, 2001) and therefore it can be claimed that they have greater knowledge about the celebrity and perceive information about the celebrity as more important. Such importance is a crucial variable in the formation and change of attitude toward the object of the message because the individual is motivated to engage in more active processing of information to accumulate more relevant knowledge about the topic and to act on their own conviction (Lecheler et al., 2009).

Funk and Pritchard (2006) demonstrated how difficult is to change previously hold attitude or opinion about certain sports celebrities. They revealed that committed individuals, such as fans of the sports celebrity who highly identify with the celebrity, had more stable beliefs and feelings about their favourite team despite being faced with negative or positive publicity (Funk & Pritchard, 2006). Therefore, it is perceived that negative publicity about an

(19)

UNDERPERFORMANCE ON CONSUMER ATTITUDE TOWARD A BRAND 19 athlete may not be influential when the receiver of the message is a committed fan of the athlete, team or sport (Funk & Pritchard, 2006).

Based on the previously mentioned theories and findings, we propose that highly identified individuals who receive negatively framed news about the underperformance of a sports celebrity they admirer will, due to their pre-existing associations with the sport celebrity, invoke more elaboration and different schemas when processing such content and will elaborate deeper about the claims in the article compared to the less identified

individuals. This pre-existing knowledge and attitude toward the sports celebrity will be more difficult to change than in case of less identify individuals. Therefore the individuals who highly identify with the sports celebrity will be less influenced by news about the

underperformance of the sports celebrity negatively framed by the media and will perceive the sports celebrity as more credible and attractive than less identified individuals. To sum this up, the following hypotheses are offered:

The perception of an athlete’s credibility (H2a) and attractiveness (H2b) will be lower only for viewers who read news about an underperformance of a sports celebrity that was negatively framed by the media and are less identified with the sports celebrity and not for those who read news about an underperformance of a sports celebrity that was negatively framed by the media and are highly identified with the sports celebrity.

The Impact of Sports Celebrity Characteristics on Brand Attitude

Companies realize that the efficiency of the endorser in the advertising is influenced by his/her characteristics and spend considerable time and money to select the celebrity that could have the most positive and powerful impact on the effectiveness of the advertising (Ohanian, 1990 in Lafferty & Goldsmith, 1999). Besides other characteristics5, brand

(20)

managers and advertising strategists look into two important characteristics of the endorser – his/her credibility and attractiveness.

The reason why they search for a credible source is that many previous studies have proved that a more credible source is more persuasive (Lirtzman & Shuv-Ami, 1986; Maddux & Rogers, 1980; Sternthal, Dholakia, & Leavitt, 1978). Furthermore, the majority of

empirical evidence (Ajzen & Fishbein, 1975; Atkin & Block, 1983; Goldberg & Hartwick, 1990 all in Lafferty & Goldsmith, 1999; Lee & Koo, 2015) has shown the positive effect of a highly credible endorser on consumers’ attitudes toward the brand compared with a less credible endorser.

Similarly, the overall attractiveness of the endorser influences the effectiveness of the message. The previously mentioned Source Attractiveness Model proposed by McGuire (1985 in Lee & Koo, 2015) indicates that the more a consumer knows, likes, and is similar to the endorser, the more attractive he or she will be to the consumer and the easier it will be for the consumer to accept the message (Tellis, 2003 in Lee & Koo, 2015). Several studies (Caballero & Pride, 1984; Debevec & Kernan, 1984) have demonstrated that an attractive source generates positive consumer responses and is effective with respect to changing consumers’ attitudes (Lee & Koo, 2015). In other words, the attractiveness of the endorser can influence viewer’s attitude toward the brand in the advertising.

Based on previously stated findings (Caballero & Pride, 1984; Debevec & Kernan, 1984; Lee & Koo, 2015) which indicate that the endorser’s credibility and attractiveness positively influence viewer’s attitude toward the brand, the following hypotheses are offered:

H3a: The higher the perceived credibility of the sports celebrity endorser, the more positive the attitude toward the brand.

H3b: The higher the perceived attractiveness of the sports celebrity endorser, the more positive the attitude toward the brand.

(21)

UNDERPERFORMANCE ON CONSUMER ATTITUDE TOWARD A BRAND 21 The Moderation Effect of Product Involvement on Brand Attitude

When individuals view an advertisement with the sports celebrity endorser they choose which information they will process and how they will process it. Such a selection is made based on individual’s level of involvement. In 1984, Petty and Cacioppo proposed the

Elaboration Likelihood Model (ELM) and put the individual’s level of involvement into the context of information processing and attitude change. The ELM explains different ways of processing stimuli (advertising) based on the individual’s level of involvement and their outcomes on consumer’s attitude change. The main assumption of the ELM is that different

methods of inducing persuasion may work best depending on the individual’s level of

elaboration (Petty et al., 1983). When consumers are higly involved with the product or issue, they are more motivated to devote the cognitive effort required to evaluate the true merits of an issue or product (Petty et al., 1983). In such a case, attitude change is affected through information-processing in the central route and the ideal manipulation of the advertising should require extensive issue related or product-relevant thought (Petty et al., 1983). The ELM brings an important implication for advertising professionals as it demonstrates that different features of the advertisement may be most effective for an audience with a certain level of involvement (Petty et al., 1983). For example, advertising with strong arguments which support the purchase of the product should be successful for highly involved consumers.

Contrary to this, when the customer is less involved with the product/issue, the attitude change is affected through information-processing in the peripheral route and the most effective manipulaton of the advertising allows a person to evaluate an issue or product without engaging in extensive thinking about the issue or product-relevant information (Petty et al., 1983). Such a consumer will rather focus on peripheral cues, such as the attractiveness

(22)

and the credibility of the product’s endorser than on product-related information (Petty et al.,

1983).

The ELM was supported by several studies which have demonstrated that the

presence of the celebrity endorser, which acts as a peripheral cue, is more effective with less involved consumers (Petty et al., 1983; Pughazhendi et al., 2012) as they may simply accept or reject the message on the basis of source characteristics without carefully examining the arguments (Yoon, Kim, & Kim, 1998 in Sallam, 2011). Contrary to this, highly involved consumers are relatively immune to the effect of source credibility (Yoon et al., 1998 in Sallam, 2011) and attractiveness as they will tend to focus on product/issue-related information.

In summary, based on the findings of previously mentioned studies which have demonstrated that the effectiveness of the source, namely the sports celebrity endorser, is influenced by a consumer’s level of product involvement (Cacioppo, Petty, Kao, & Rodriguez, 1986), the following hypothesis is proposed:

A viewer’s level of product involvement moderates the direct effect of a consumer’s perception of sports celebrity characteristics on brand attitude. Specifically, the viewer’s brand attitude will be lower only for those who are less involved with the product and have low perception of the sports celebrity’s credibility (H4a) and attractiveness (H4b) compared to those who are highly involved with the product and have low perception of the celebrity’s credibility and attractiveness.

(23)

- News about underperformance of sport celebrity negatively framed by media (negative news)

- News about a victory of sport ceebrity positively framed by media (positive news) - Neutral news about

the upcoming competition of sport celebrity (neutral news) Brand attitude Product involvement (high/low) Level of identification

with sport celebrity (high/low) Brand attitude Brand attitude Conceptual Model Type of news

Sport celebrtity characteristics

H1a H1b H3a H3b H2a H4a H2b H4b Methods Experimental Design

For the purpose of this study a between-group experimental design (see Table 1) was chosen to assess the differences in sports celebrity credibility, attractiveness, and viewer’s attitude toward the brand endorsed by the celebrity based on the celebrity’s sports

performance portrayed by the media. This design involved two experimental conditions and a control condition with each participant being assigned to only one condition: news about an underperformance of the sports celebrity negatively portrayed by the media vs. news about a victory of the sports celebrity positively portrayed by the media vs. neutral news about the sports celebrity (control condition). In order to create and choose the approporiate stimuli

Sport celebrity credibility and attractiveness

(24)

(news about an underperformance/victory of the sports celebrity; sports celebrity endorser, and a fictitious product for the advertisement) two pre-tests were conducted.

Design Schema Table 1

Experimental design

News about a victory of the sports celebrity positively framed by

the media

News about an

underperformance of the sports celebrity negatively framed by

the media

Condition 1 Condition 2

Sample

To test our hypotheses, an online survey was conducted. In order to recruit the respondents, a convenience sampling technique was used. In comparison to probability sampling techniques, this method is inexpensive and does not require access to the whole population (Bryman, 2003). Furthermore, the possibility to generate large amount of data in a relatively short time span was an important aspect of method selection. Due to the usage of the convenience sample the selection of the participants was not random (Bryman, 2003). The respondents were recruited through the researcher‘s personal ties via social network - Facebook. The Facebook messages containing the study’s URL have been sent to 166 people.

The target population of the study consisted of 21 – 39-year-old males and females of Czech nationality. The age range was selected for two reasons. First, people of such an age range have higher levels of exposure to celebrities. Previous studies revealed that young adults often emulate their favourite media figures and adopt the media figures’ values and

personality characteristics (Boon & Lomore, 2001; Gomillion & Giuliano, 2011; Matthews, 2003). Second, young adults who had significant existing and potential purchasing power, and disposable income were targeted.

(25)

UNDERPERFORMANCE ON CONSUMER ATTITUDE TOWARD A BRAND 25 In order to obtain a varying and diverse range of identification with the sports celebrity (athlete), our target population consisted of both current and former members of athletic clubs and non-members of the athletic clubs. Based on Social Identity Theory (Tajfel & Turner, 1979), it is perceived that members of athletic clubs will identify with the athlete to a greater extent than non-members of athletic clubs. Furthermore, athletes as consumers of sports facilities are of significant interest to sport brands advertisers.

The survey was filled out by a total of 117 participants. Our final sample consists of 99 participants (55.9% male and 44.1% female), the mean age of the participants was 26.43 (SD = 3.50) years. 18 participants were dropped from the study because they had not heard about the athlete Zuzana Hejnová before their participation in the survey or did not specify whether they had heard of her before. The level of education was fairly high with 67.8% of participants holding a university degree. No incentives were offered in this study as a motivation to fill out the questionnaire.

Procedure

The online experiment was conducted by using the Qualtrics research survey site (for the whole survey see Appendix A). The participants took part in the experiment using a computer in their natural environment and were randomly assigned to one of three conditions. Before the start of the survey they were presented with an invitation for the study and

informed about the consent procedure. Furthermore, they were asked whether they

understood the text and if they wished to participate in the study. Respondents who gave their informed consent and wished to take part in the study were displayed with the survey.

First, each group of participants filled in the first couple of questions measuring their identification with Zuzana Hejnová (see below for stimulus material) and personal

knowledge of this athlete. Second, depending on the group in which the participants were, they were exposed to one of three pieces of news about performance (the negatively framed

(26)

news about the underperformance of Zuzana Hejnová vs. positively framed article about the victory of Zuzana Hejnová vs. neutral news about Zuzana Hejnová) of Zuzana Hejnová.

Afterwards, the respondents were asked how they perceived the performance of Zuzana Hejnová at the competition reported in the article (very negative/very positive and very

successful/very unsuccessful6) as manipulation check. Third, all of the participants answered questions about Zuzana Hejnová’s credibility and attractiveness. Afterwards, all the

respondents viewed the advertisement of the same product – sports shoes – which was endorsed by Zuzana Hejnová and indicated their attitude toward the brand and attitude

toward the ad (the latter was included as control variable). The participants were given instructions to read the online news and watched the advertisement in their usual manner and could look at both stimuli (news and advertisement) for as long as they wanted to. Last, the respondents filled in the last part of the questionnaire answering questions indicating their product involvement, demographics (age, gender, education) and control variables (athletic identification, athletic membership). The whole procedure lasted approximately 10 minutes. Participants were recruited over a two-week timeframe in, December 2015.

Pre-tests

Two pre-tests with 9 and 8 participants of our target population were conducted. In order to obtain the appropriate sports celebrity with different levels of consumer

identification, and to select the product with a diverse level of consumer’s involvement, the first pre-test with 9 participants7 was conducted.

6_{Due to the fact that the neutral article (control condition) did not report about the performance of} Zuzana Hejnová but only informed about the future participation of Zuzana Hejnová at upcoming competition a

question about the perception of the negativity of the article toward Zuzna Hejnová has been incorporated as the manipulation checked question.

7_{Age (M = 26.67, SD = 3.24, min = 24, max = 35), gender (55.6% females, 44.4% males), members of} the athletic club 44.4%.

(27)

UNDERPERFORMANCE ON CONSUMER ATTITUDE TOWARD A BRAND 27 Previous studies (Kamins, 1989; Money, Shimp, & Sakano, 2006 all in Thwaites et al., 2012) confirmed the importance of using real sports celebrities in the experiment because of its advantage of lending realism in examining the effect under consideration. Therefore, four currently competing Czech athletes (Zuzana Hejnová, Vítězslav Veselý, Jaroslav Bába, Lukáš Melich) who during their career underperformed8

at world-known competitions, as well as having won a medal in such a competition, were selected for the first pre-test. Participants were supposed to indicate whether they had ever heard about each athlete and afterwards assess their level of familiarity and identification with all four athletes. Czech hurdler and runner Zuzana Hejnová was selected as almost 90% of the respondents had heard about her before, obtained the most variation in the identification score and was fairly

familiar (for descriptive statistics see Appendix B).

The second part of the first pre-test was devoted to the consumer’s involvement with the particular sports products. In order to select the product with a diverse level of

consumer’s involvement, three different types of sports products were pretested. The reason

for using the sports product is described in the theory section. The sports celebrity uses these sports product during their competition or training, and therefore, the product is linked to their performance to a greater extent than a non-sports product. Previous studies (Till & Busler, 1998, 2000) revealed that expertise is the most important component of the celebrity endorser and therefore, the sports product would be influenced to a greater extent by sports celebrity performance than a non-sports product. Three different types of the sports products

8_{Underperformance at world-known competitions was based on the season rankings of the athletes}

(which was based on their previous performances during the season). The underperformance was considered to be a result that was 3 ranks or worst below the ranking of the athlete before the world-known competition. For

example, if the athlete was prior to a world-known competition such as European or World Championship in athletics based on the rankings second and they ranked at such a competition worst then 8th then it was

(28)

which the athlete used during their training/competitions were selected: running shoes, a sport drink and sports gel (as a sportsman’s dietary supplement). A sports drink as the

product with highest variation of consumer product involvement was selected (for descriptive statistics see Appendix B).

In order to select the most negative/positive frame and to verify the advertisement stimuli, a second pre-test9 with 8 participants was conducted. Two positive and two negative news messages that reported about the performance of Zuzana Hejnová were created based on inspiration from the most popular Czech online news websites (iDnes.cz, Sport.cz, iSport.cz). Participants were asked to what extent they perceived the performance of Zuzana Hejnová (about which they were informed of in the article) as successful /unsuccessful and

positive/negative. Furthermore, in order to check the professionality, the respondents had to indicate whether they perceived the articles as professional compared to the online sports news they read on the news websites (yes/no) and write down what they did not perceive as professional. Lastly, in order to check the positive/negative perception of the news they were asked to indicate how they perceived the article when it came to the representation of Zuzana Hejnová.

Based on the statistical analysis (for descriptive statistics of all pre-test variables see Appendix B), one positive and one negative article were selected. The news message

reporting about Zuzana Hejnová’s victory at the 2013 World Championships in Moscow was

selected as it was perceived as very successful. Furthermore, the article was considered to be very positive and the performance of Zuzana Hejnová was also seen as very positive. The

news message reporting about Zuzana Hejnová’s underperformance at the 2010 European Championships in Barcelona was selected as it was perceived as unsuccessful and fairly

9_{Age (M = 26.75, SD = 3.45, min = 24, max = 35), gender (37.5% males, 62.5% females), members of}

(29)

UNDERPERFORMANCE ON CONSUMER ATTITUDE TOWARD A BRAND 29 negative. Respondents indicated that the article portrayed the performance of Zuzana Hejnová as fairly negative. Furthermore, both of the articles were considered to be highly

professional without respondents negatively commenting on the credibility of the articles. In order to exclude that a negative attitude toward the ad could influence a consumer’s evaluation of the brand, the professionality and credibility of the advertising stimuli had to be checked. Therefore, the measure of attitude toward the ad was used in the last part of the second pre-test. The overall attitude toward the ad was neutral and almost 90% of the

respondents perceived it as professional and comparable with other professional sports drinks advertisements.

Stimulus Material

Three different articles (see Appendix C) were developed to manipulate sports celebrity on-field performance, and framing of the news. In order to be as realistic as possible10 each of the articles was inspired by online news published on the most popular Czech online news website. The news used (about a victory of a sports celebrity positive framed by the media) in the experiment, reports about the first gold medal win of Zuzana Hejnová at the 2013 World Championships in Moscow which portrayed the athlete as a

phenomenal and experienced hurdler who was very confident and in great shape, and did not give her rivals any chance to stop her from winning the gold medal during the competition. The news about an underperformance of a sports celebrity negatively framed by media reports about Zuzana Hejnová’s performance at the 2010 European Championships in Barcelona. Hejnová was portrayed as inexperienced competitor who could not handle the pressure and give the performance expected of her, and that she used it as an excuse for her underperformance. The neutral article reports about Zuzana Hejnová’s participation at a

10_{Which is especially important as part of our study participants are (former) athletes who would}

(30)

Czech meeting during her preparation for upcoming season and it portrays her only as the Czech hurdler. All three articles used a picture of Zuzana Hejnová from the respective event or competition11.

The second stimulus material used in this study was the advertisement (see Appendix D) of Zuzana Hejnová endorsing the sports drinks. In order to eliminate any consumer pre-attitude toward the brand, familiarity with the brand, preferences, and user experiences with existing brands, a new fictitious brand name and logo of the sports drink were created. Afterwards a picture of Zuzana Hejnová captured during a running scene at an empty stadium was selected. The advertisement was inspired by advertisements of the famous sports drink brands (Gatorade, Powerade). Similar to these advertisements, we used darker colours and a dramatic filter. In the bottom right corner a blue bottle of the endorsed sport drink Sportic was placed. Additionally, in the right corner, the logo of the brand Sportic was placed. The text in the middle of the advertisement stated: "Complex sports drink" and "Used by Zuzana Hejnová ". In order to provide more detailed information for highly involved consumers the text “Helps to replenish all four electrolytes lost in sweat. Sodium / Potassium / Calcium / Magnesium” was placed in the right bottom of the advertisement.

Measures

Dependent variable

Attitude toward the brand. This dependent variable was assessed by a 3-item, bipolar adjective, scale (unfavorable – favorable, bad – good, negative – positive) used by McDaniel and Kinney (1996 in Koo et al., 2012). The participants were asked to indicate their overall impression of the brand. For the purpose of this study, the response categories ranged from 1 (unfavorable, bad, negative) to 7 (favorable, good, positive). A scale was

11

(31)

UNDERPERFORMANCE ON CONSUMER ATTITUDE TOWARD A BRAND 31 created by averaging the items, resulting in a Cronbach’s alpha of .97 (M = 12.70, SD =

3.00).

Mediators and Moderators

Sports celebrity attractiveness. As proposed by McGuire (1985), this mediator consists of three constructs: likeability, familiarity and similarity. One scale assessing the overall attractiveness of sports celebrity was created by incorporating and averaging four12 independent scales of in total 12 items, resulting in a Cronbach’s alpha of .88 (M = 48.82, SD = 8.60).

Similarity. To measure the participants’ perception of similarity with the sports celebrity, a 3-item scale used by Taylor, Halstead and Haynes13 (2010; = .93) was incorporated. The participants were asked to indicate to what extent they agreed with the proposed statements. The response categories ranged from 1 (strongly disagree) to 7 (strongly agree). The items included: "Zuzana Hejnová is a lot like me"; "Zuzana Hejnová holds beliefs that are similar to my own"; "Zuzana Hejnová has attitudes that are similar to my own". The overall level of similarity was assessed by averaging the items, resulting in a Cronbach’s alpha of .86 (M = 10.19, SD = 3.25).

Likeability. To measure the likeability of Zuzana Hejnová, a 6-item, bipolar, scale (friendly – unfriendly, sincere – insincere, likeable – unlikeable, warm – unkind,

charismatic – uncharismatic, appealing – unappealing) used by Yilmaz et al.14 (2011; AVE15

12

The principal component analysis (PCA, see Appendix E) of the original single uni-dimensional scale measuring likeability revealed two components with an eigenvalue above 1 (eigenvalue 3.89 and 1.02).

13

The scale was originally proposed by Byrne (1971 in Taylor et al., 2010) and Pilkington and Lydon (1997 in Taylor et al., 2010).

14_{These items were adapted from two previously used scales developed by W}_{hittler and Meo (}_{1991 in}

Yilmaz et al., 2011) and Zhang and Buda (1999 in Yilmaz et al., 2011) 15

(32)

= 0.75) was adopted. For the purpose of this study, the response categories ranged from 1 (unfriendly, insincere, unlikeable, unkind, uncharismatic, unappealing) to 7 (friendly,

sincere, likeable, warm, charismatic, appealing). The principal component analysis of the

original single uni-dimensional scale revealed two components with an eigenvalue above 1 (component 1 is 3.89 and component 2 is 1.02). Therefore, it appears that these two

components measure "likeability of the character" and "perception of celebrity charisma". An overall level of likeability of the character was assessed by averaging 4 items, resulting in a Cronbach’s alpha of .92 (M = 22.02, SD = 4.27). The overall perception of celebrity

charisma was assessed by averaging 2 items, resulting in a Cronbach’s alpha of .76 (M = 9.48, SD = 2.19).

Familiarity. This variable was measured by a 3-item bipolar scale (not at all familiar

– extremely familiar, definitely do not recognize – definitely recognize, definitely have not heard of her before – definitely have heard of her before) developed by Simonin and Ruth

(1998 in Bruner, 2009; = .80) and originally measured familiarity of the object. The

participants were supposed to indicate how familiar they are with the athlete Zuzana Hejnová. For the purpose of this study, the response categories ranged from 1 (not at all familiar,

definitely do not recognize, definitely have not heard of her before) to 7 (extremely familiar, definitely recognize, definitely have heard of her before). An overall familiarity with Zuzana

Hejnová was assessed by averaging the items, resulting in a Cronbach’s alpha of .77 (M =

16.57, SD = 3.37).

Sports celebrity credibility. This second mediator was measured by the scale proposed by Ohanian (1990) which consists of three subscales of 15 items measuring the following constructs: attractiveness, trustworthiness and expertise. A scale assessing overall credibility of the sports celebrity was created by averaging all three subscales, resulting in a Cronbach’s alpha of .90 (M = 75.65, SD = 11.75).

(33)

UNDERPERFORMANCE ON CONSUMER ATTITUDE TOWARD A BRAND 33 Attractiveness. To measure the attractiveness of the sports celebrity a 5-item bipolar scale (attractive – unattractive, classy – not classy, beautiful – ugly, elegant – plain, sexy –

not sexy) proposed by Ohanian (1990; = .89) was used. For the purpose of this study, the

response categories ranged from 1 (unattractive, not classy, beautiful, plain16, not sexy) to 7

(attractive, classy, ugly, elegant, sexy). The item with the response categories beautiful – ugly was reversed. The overall attractiveness was assessed by averaging the items resulting in a Cronbach’s alpha of .82 (M = 19.87, SD = 4.46).

Expertise. To measure the expertise of the sports celebrity a 5-item bipolar scale (expert – not expert, experienced – inexperienced, knowledgeable – unknowledgeable,

qualified – unqualified, skilled – unskilled) proposed by Ohanian (1990; = .89) was used.

For the purpose of this study, the response categories ranged from 1 (not expert,

inexperienced, unknowledgeable, unqualified, unskilled) to 7 (expert, experienced,

knowledgeable, qualified, skilled). The overall expertise was assessed by averaging the items

resulting in a Cronbach’s alpha of .94 (M = 29.25, SD = 5.84)

Trustworthiness. To measure the trustworthiness of the sports celebrity a 5-item bipolar scale (dependable – undependable, honest – dishonest, reliable – unreliable, sincere

– insincere, trustworthy – untrustworthy) proposed by Ohanian (1990; = .90) was used. For

the purpose of this study, the response categories ranged from 1 (undependable, dishonest,

unreliable, insincere, untrustworthy) to 7 (dependable, honest, unreliable, sincere). The

overall trustworthiness was assessed by averaging the items resulting in a Cronbach’s alpha

of .92 (M = 28.30, SD = 5.29).

Level of identification with the sports celebrity. This moderator was measured by a 5-item scale used by Um (2013) which combined several previously used scales (Lammie, 2007; Rubin, & McHugh, 1987; Rubin, Perse, & Powell, 1985 all in Um, 2013). The

(34)

participants were asked to indicate to what extent they agree with the proposed statements. The response categories ranged from 1 (strongly disagree) to 7 (strongly agree). The items included: "I like Zuzana Hejnová"; "I can easily relate to Zuzana Hejnová"; "I think of Zuzana Hejnová as of a good friend"; "I have no doubt Zuazana Hejnová and I would work well together"; "Zuzana Hejnová is a personal role model". The level of identification with the sports celebrity was measured by averaging the items, resulting in a Cronbach’s alpha of .78 (M = 20.29, SD = 4.56) and then, by median, split divided into two groups (low vs. high).

Product involvement. This moderator was measured by a 6-item scale proposed by Taylor and Joseph (1984; = .92). The participants were asked to indicate to what extent they agree with the proposed statements. The response categories ranged from 1 (strongly

disagree) to 7 (strongly agree). Examples of items included: "When other people see me

using this product, they form an opinion of me", "You can tell a lot about a person by seeing what brand of this product he uses"; "This product helps me express who I am"; "When I use this product, others see me the way I want them to see me". The level of product involvement was measured by averaging the items, resulting in a Cronbach’s alpha of .87 (M = 17.20, SD

= 6.41) and then, by median, split divided into two groups (low vs. high). Control Variables and Pre-test/Post-test Measures

Age. This control variable was operationalized with one question: ‘How old are you?’ Each participant had to write down his or her age (M = 26.43, SD = 3.50).

Athletic identification. This variable was measured by a 4-item scale used by Carrillat17 et al. (2014). The response categories ranged from 1 (strongly disagree) to 7 (strongly agree). The participants were asked to indicate how much they agree with the listed statements: "I regularly watch athletic competitions"; "I regularly check athletic results"; "I regularly track athletic statistics"; "I consider myself an athletic fan". The level of athletic

17

The scale was originaly adapted from Trail and James (2001 in Carrillat et. al, 2014) and Trail, Robinson, Dick and Gillentine (2003 in Carrillat et al., 2014).

(35)

UNDERPERFORMANCE ON CONSUMER ATTITUDE TOWARD A BRAND 35 identification was measured by averaging the items, resulting in a Cronbach’s alpha of .93 (M

= 17.54, SD = 6.30)

Athletic membership. This variable was operationalized with one question: ‘Have you been a member of athletic club in last 15 years?’ Each participant was asked to choose

one of the response categories: "Yes, I have"; "No, I have not" (41.90% yes, 58.10% no). Attitude toward the ad. This variable was measured by a 4-item scale developed by Lee (2000). The response categories ranged from 1 (strongly disagree) to 7 (strongly agree). The participants were asked to indicate how much they agree with the listed statements: "I like the advertisement that I saw"; "The advertisement that I saw is attractive to me"; "The advertisement that I saw is appealing to me"; "The advertisement that I saw is interesting to me". The overall attitude toward the ad was assessed by averaging the items, resulting in a Cronbach’s alpha of .89 (M = 19.09, SD = 5.29).

Education. This control variable was operationalized with one question: "What is the highest educational level you have obtained?" Each participant had to choose one of the 7 response categories: "No schooling completed"; "Primary/elementary school",

"Secondary school/high school"; "Higher vocational school"; "Bachelor‘s degree"; "Master‘s degree"; "PhD". (67.8% university education completed, 32.2% non-university education completed).

Gender. This variable was operationalized with one question: "What is your gender?" Each participant was asked to choose one of the response categories: "female" or "male" (55.9% male and 44.1% female).

Perception of success. This manipulation check variable was operationalized with one question: "How do you perceive the performance of Zuzana Hejnová which was

described in the article?" For the purpose of this study, the response categories ranged from 1 (very successful) to 7 (very unsuccessful).