Only Men Can Do This and Only Women Can Do That
An Experiment on The Effectiveness of Female and Male Athletes in Advertising
Yasmin Saud Abdulkadir Student number: 11754834
Master’s Thesis
Graduate School of Communication Master’s programme of Communication Science
Supervisor: Dr. Saar Mollen 1st February 2019
Abstract
The discrepancy between male and female endorsers in advertising was addressed in this study by focusing on what makes an endorser credible and, therefore, effective. To be credible, an endorser should possess the following three characteristics; attractiveness, trustworthiness and expertise. This research focuses on the effects of the gender of the athlete, the appropriateness of sport and the product fit on the credibility variables. To asses this a 2 (Endorser: female vs male) x 2 (Appropriateness: Appropriate vs
Inappropriate) x 2 (Product fit: High product fit vs low product fit) online experiment was conducted assigning the participants randomly to one of the eight experimental groups. Findings indicated that the type of sport had no influence on the credibility, whereas the gender of the athlete had a significant difference on one credibility variable; attractiveness. Further, female athletes were perceived to be more attractive than male athletes. Results revealed that the product fit was key to the effectiveness of an endorser.
In conclusion, the gender appropriateness of sports and the use of male or female endorsers was an important aspect relative to endorser effectiveness, which may be explained by a change in gender norms.
Sports coverage has for years focused more heavily on men than women (Koivula, 1999; Kian & Hardin, 2009). In 1999, 86.7% of sports coverage was of men, 11.7% was of women and the residual 1.7% sports coverage included both (Koivula, 1999). This difference in sport coverage of both genders is also salient in advertising (Fink Parker, Cunningham, & Cuneen, 2012). The discrepancy between male and female athletes in advertising and endorsement is still present today. As of last year, the top 10 highest paid male athletes earned $65 million on endorsement deals (Statista, 2018a), whereas for female athletes it was only $18 million (Statista, 2018b). Individuals are exposed to 3000-5000 advertisements daily, making it important for advertisers to create a memorable advertisement (Petrecca, 2006). To do so, companies use endorsers1 (Belch & Belch, 2013).
When choosing an endorser, several characteristics need to be considered to make the endorser credible and, thus, effective. Numerous studies have demonstrated that trustworthiness, attractiveness and expertise are characteristics that have a positive impact on the credibility of an endorser. Furthermore, it is found that the endorser’s gender influence their effectiveness (Barner, 1999; Ganahl, Prinsen, & Netzley, 2003; Eisend, 2010). The difference in the use and effectiveness of men and women as endorsers may stem from gender stereotyping (Bem, 1981), which is an idea of how men and women are expected to behave and dress. These stereotypes are also present in sports, as most sports have been studied to be more appropriate for men (Riemer & Visio, 2003). The gender stereotypes could lead to a higher expectation of men participating in a broad variation of sports, which could also help explain why certain sports have traditionally been categorized as masculine over feminine (Van Riper & Badenhausen, 2008; Hardin & Greer, 2009; Fink et al., 2012). For example, different types of sports have been identified as fitting into different categories
1
An endorser is a person, often a celebrity, that recommends and supports a product. When athletes are used as endorsers for a brand of specific products they help to both gain and maintain consumers’ attention (Koernig & Boyd, 2009).
determined by the gender stereotypes: feminine and masculine (Csizma, Wittig, & Schurr, 1988; Koivula, 1999; Hardin & Greer, 2009). For instance, if a woman participates in a sport, which is categorized as masculine, such as wrestling, the sport would be perceived as gender-inappropriate (Fink et al., 2012). In examination of what makes a sport appropriate and inappropriate for females (Metheny, 1965; Csizma et al., 1988; Riemer & Visio, 2003; Fink et al., 2012), studies has found that when a female athlete participated in a gender appropriate sport, she was given a higher social status than those participating in a gender inappropriate sport (Kane, 1988). It appears that when an athlete participate in gender appropriate sports, they are more popular among the public, due to these being more in line with gender stereotypes and, therefore, more pleasant to perceivers (Holland & Andre, 1994; Eagly & Wood, 2012). Despite this, studies (Erdogan, 1999; Till & Busler, 2000; Fink et al., 2012) about how and if the perceived gender appropriateness of sports affects the perception and effectiveness of athlete endorsers in advertising have been limited and, thus, remains relatively unknown.
Until now, the gender appropriateness of sports when using an athlete endorser seems to affect their effectiveness (Fink et al., 2012). Besides sport appropriateness’
influence on endorser effectiveness, literature has shown that a so-called match-up hypothesis also plays a role (Kamins, 1990; Till & Busler, 2000). It suggests that there is a need for a match between the endorser and the product, i.e. that the products needs to support the same values as those of the endorser (Till & Busler, 2000). For example, if an athlete endorses a sports drink, it seems like a natural match, as the purpose of sports drinks is to enhance athlete performance. However, this would not be the case with a Coca Cola as its purpose is not to enhance athletes’ performance and, thus, there would be a mismatch. Therefore, a good product fit can influence how effective an endorser is (Kamins, 1990; Till & Busler, 2000; Fink et al., 2012). Similarly, a mismatch can cause the endorser to be less effective.
The match-up hypothesis suggests that the product fit can influence whether the endorser is going to be more or less effective. However, a combination of product fit and appropriateness of sport with regard to endorsers’ gender may affect the effectiveness. A good product match with an endorser in a gender appropriate sport may produce a more credible endorser compared to an endorser in a gender inappropriate sport with bad product fit. The combination of product fit and appropriateness of sport for females and males is a neglected area in the advertising field, since the majority of research has either focused on female and male endorsers that are models and not athletes, or on only female athletes
without considering male athletes (Kanungo & Pang, 1973; Lynch & Schuler, 1994, Erdogan, 1999; Till & Busler, 2000; Fink et al., 2012). Therefore, this present study will focus on addressing female and male athlete endorsers and the effectiveness between both genders through this research question:
1) How does the gender of athlete endorsers and the classification of the sport as masculine or feminine impact the credibility of the endorser and the purchase intention? And does this depend on whether there is a high or a low product-fit?
Theoretical framework Endorsers
Brands often use endorsers who are both attractive and highly relevant to the product that is endorsed (Belch & Belch, 2013). Endorsers must match the values of the brand to be effective and credible in the eyes of the consumers (Belch & Belch, 2013; Erdogan, 1999; Fink et al., 2012; Belch & Belch, 2018). Using endorsers that are either not well-liked or have credibility issues of any kind can result in the consumer rejecting the message. As mentioned in the introduction, trustworthiness, attractiveness and expertise are found to have a positive influence on endorser effectiveness (Erdogan, 1999; Fink et al., 2012). An analysis by Amos, Holmes, and Strutton (2008) also found that these characteristics were the most
important ones, as they significantly enhanced endorser effectiveness. Furthermore, do literature suggest that an effective endorser will influence to an increase in purchase intention (Fink et al., 2012).
Gender of endorsers
Characteristics of an endorser can influence the way he or she will be perceived by an audience. Besides finding an endorser who is recognised as credible (i.e., trustworthy,
attractive and an expert), the gender should be considered when deciding on an endorser for a certain product, as it may influence the perceived credibility (Fink et al., 2012).
Boyd and Shank (2004) examined the effects of the gender of an endorser on trustworthiness, attractiveness and expertise. They found that male athletes were significantly more attractive than the female endorsers, but there was no difference on their
trustworthiness and expertise (Boyd & Shank, 2004). Boyd and Shank (2004) also suggest that male athletes may be more effective as endorsers than their female counterparts. The stereotypes that lies within gender may be a reason to why the Boyd and Shank (2004) found differences between a female and male athlete endorser. Previous work (Kane, 1988,
Cunningham et al., 2008, Fink et al., 2012) has manily focus on the credibility of female athlete endorsers and, thus, limited to study the influence of the gender of an athlete endorser in advertising. This lead to the second research question of this paper:
2) Which differences are there in using a female athlete compared to a male in terms of their credibility?
Gender stereotyping
When clarifying which gender an endorser should have when used for a sports advertisement, it is important to note the differences between the two concepts of sex and gender. The concept sex refers to the biological grouping of human beings in two groups – male and female (Eagly, 1987). Whereas, according to Eagly (1987), using the term gender is socially
constructed by society and individuals to assign meanings to either a female or male
category. For example women should wear dresses and men should wear suits. These gender stereotypes are carried out when different social roles are enacted, e.g. when being a parent or an employee (Eagly, 1987; Eagly & Wood, 2012; Fink et al., 2012).
The perceptions of gender interfere with how male and female behaviour is defined, which means that certain characteristics are associated with each gender (Bem, 1981). Both genders are expected to obtain sex-specific skills, but also personal attributes to fit in the perception of being either masculine or feminine (Bem, 1981). This mostly stems back from childhood when children are taught content-specific information about the link between behaviour and sex, also known as sex-typing (Bem, 1981).
The gender-stereotypes, which are social constructs, can be explained by the gender schema theory (Bem, 1981). The theory suggests that the phenomenon of sex-typing is determined and derived in two parts; from a gender-based schematic processing and general readiness to process information on the grounds of the associations created with the two different sexes. The sex-typing also originates from the fact that the self-concept gets assimilated into the gender schema. The gender schema theory also proposes that humans have a cognitive structure consisting of a network of associations. These associations are organised and lead an individuals’ perception (Bem, 1981). The function of the schema theory is to assimilate new information and understand how it fits with already existing information. This process is called schematic processing and is not only highly selective but also allows the individuals to establish structure and an extensive array of incoming stimuli (Bem, 1981). Bem (1981) also highlights that the schema theory interprets perceptions as being a process in which the perception is a product of the interaction of the new information and the information that is already in an existing schema within the perceiver.
The schema is formed on the basis of associations, for example when a girl is told not to play with toy cars because these are only for boys, and boys are told not to play with dolls because these are only for girls. This information is then mentally stored. When new information is being processed, it will first be associated and encoded with existing information, which will then create a schema. When growing up children are taught content-specific information, such as certain behaviours being appropriate for either boys or girls. This invokes gender-related associations which are already stored from prior experience, that get evaluated and assimilated (Bem, 1981). The result of the process leads to certain gender expectations, the gender schema. The schema leads to what is known as gender stereotypes (Bem, 1981). Therefore, when considering whether to use a female or male athlete endorser, it should be considered that the type of sport may be influenced by stereotypes. Derived from gender schema theory a female endorser would be more effective, if she was to operate in a gender stereotypical role, such as wearing a dress (compared to if she did not appear in a gender stereotypical role). If a male endorser was to wear a dress, he would not align with the stereotypical man, which would mean that he would appear as a less effective endorser than if he were to wear pants. When an endorser acts in line with gender stereotypes, they become more pleasant to the perceiever and, thus, more effective endorsers (Holland & Andre, 1994; Eagly & Wood, 2012;).
Classification of sport
Stereotypes with regard to the role of men and women should take on in society span many areas (Eagly & Wood, 2012). Examples are the labour market, the household, choosing which study to enroll in or which sport to engage in. It has been found that sports can be determined to either be gender appropriate for a male or a female (Metheny, 1965; Csizma et al., 1988; Riemer & Visio, 2003). For example, gymnastics is typically considered to be a feminine sport, while football is typically considered to be a masculine sport (Csizma et al.,
1988). Metheny (1965) found that the majority of sports were perceived as being more appropriate for men than women. The sports that were found to be more appropriate for women were more asthetically pleasing with no body contact with the opponent ( Metheny, 1965; Fink et al., 2012), for example figure skating, golf and diving. Sports that were
perceived as more appropriate for men were those sports in which the participant propels the body into or through space over long distances (e.g. pole-vaulting) or applies force to heavy objects (e.g. weight lifting). The different sports that are appropriate for males or females lead to the classification of athletes participating in their respective gender appropriate sports.
In summary, the way the aforementioned gender roles are socially constructed is reflected in sports contexts. Previous results have found people participating in gender inappropriate sports to have negative outcomes on their social status compared to people participating in gender appropriate sports (Kane, 1988). Contrary, Fink and colleagues (2012) who found that there was no difference between females participating in both gender
inaproppriate and appropriate sport. However, Fink et al. (2012) only demonstrated evidence for females and not males. Kanungo and Pang (1973) studied the effects of both female and male models, but this was not with regard to athletes and sports. Also, previously mentioned studies (Kane, 1988; Koivula, 1995) were based on female endorsers, thus, a comparative study regarding credibility between female and male athlete endorsers has not yet been established. Therefore, the following hypotheses are proposed:
H1. Endorsers in gender appropriate sports are perceived as more attractive (H1a), trustworthy (H1b) and to have higher expertise (H1c) compared to endorsers in gender-inappropriate sports and will result in higher purchase intentions (H1d).
The match-up hypothesis and product fit
When optimising the effectiveness of an athlete endorser in advertising, the gender and the type of sport (s)he plays, are only two out of multiple decisions that have to be made when
creating effective advertisements. The effectiveness of an endorser is not only determined by his or her credibility, but also by the type of product that is being endorsed (Kanungo & Pang, 1973; Lynch & Schuler, 1994; Fink et al., 2012). This is called the match-up
hypothesis (Kamins, 1990). Kanungo and Pang (1973) suggested that an advertisement using endorsers would be more effective when there is a fit between the model used and the
product being advertised (as in Lynch & Schuler, 1994). In comparing male and female models with different products their analysis revealed that the effectiveness was dependent on how the endorser and the product fit together. For example, a bad fit could be Lebron James (professional basketball player) endorsing a Mcdonald’s Cheeseburger, whereas a good fit could be James and a granola bar (e.g. KIND) as one would associate an athlete with healthy food and not junkfood. The ‘fit’ between an endorser and a product characterises how well the product suits the endorser and his/her representation. In other words, the fit of an endorser with a product influences the overall effectiveness.
As mentioned above, athletes participating in gender appropriate sports are important to obtain a positive effect on endorser credibility (i.e. attractiveness,
trustworthiness and expertise). Also, when there is a high product fit, the source credibility should be even greater (Riemer & Visio, 2003; Fink et al. 2012). Fink et al. (2012) found that a sports drink and an athlete were perceived to have a significantly higher fit (i.e. matches) compared to when the athlete was endorsing a non-sports drink. Their results are in line with previous results regarding the match-up hypothesis (Kamins, 1990; Till & Busler, 2000; Koernig & Boyd, 2009). Research shows (Till & Busler, 2000; Fink et al., 2012) that the match-up hypothesis can be explained by the associative learning theory, meaning that a relationship between concepts can be developed. Brands and endorsers can each represent different and specific components and values to consumers. However, these will be based on consumers’ own previous experiences (Till & Busler, 2000), which means that they will
create their own associations when seeing a brand or an endorser. These associations are set for both brand and endorser. For instance, when a certain endorser or products brings specific associations to mind, e.g. when seeing Serena Williams (professional Tennis-player), one might think of Nike. However, whether or not the associations become stronger is often determined by several factors. The more similarities there are between the product and the endorser, the more likely it is that they will be part of the same associative network (Till & Busler, 2000). This means that a person’s attitude and evaluation of a brand constitute what is called an association set. When an endorser and a product repeatedly appear together, they become part of each others’ association set. The speed with which an associative link between two concepts are built is highly influenced by the fit (how well they are related to each other). When brands succeed in creating associative links, it can result in positive reactions from consumers (Belch & Belch, 2018), which could mean that an endorser was effective. This is exactly why understanding the association process is important.
As reported earlier in the section about gender stereotypes, the gender appropriateness of the sport needs to align with the endorser to benefit from an expected positive evaluation on credibility. Similarly, the product fit is also key in the credibility, i.e. the perceived attractiveness, trustworthiness and expertise. This would mean that the combination of a high product fit and an athlete in a gender appropriate sport should have a greater effect on the source credibility. Therefore, following hypotheses have been
formulated:
H2. Endorsers in gender appropriate sports are perceived as more attractive (H2a), more trustworthy (H2b) and have a higher expertise level (H2c) compared to endorsers in gender-inappropriate sports and will result in higher purchase intentions (H3d) when there is a misfit between the endorser and the product, but not when there is fit.
Participants and design
To assess the research questions in an online experiment was conducted with a total of 502 participants. Participants were randomly assigned to one of eight conditions in a 2 (Endorser: female vs male) x 2 (Appropriateness: Appropriate vs Inappropriate) x 2 (Product fit: High product fit vs low product fit) between-subject design. The participants were mainly recruited through Facebook and LinkedIn. This led to a broad sample, where the mean age was 29.52 (SD = 10.60), and a nationality variation of several countries (n = 48) such as Denmark (n = 117), USA (n = 80), The Netherlands (n = 52) and India (n = 38). Participants under 18 (n = 5) were excluded from the sample due to ethical reasons. Forty people were removed from the sample because they did not fill out the entire questionnaire. Another seven participants were removed primarily due to choosing the answer option far to the right throughout the entire questionnaire making their answers invalid and unreliable. This research had no other criteria, and the final sample (n = 450) consisted of 178 males and 272 females.
Pre-test
A pre-test (n = 36) was conducted with two different purposes and consisted of 18 females and 18 males with the mean age 24.72 (SD =3.53). First, to ensure that the two sports that were going to be used in the experiment would be perceived as feminine and masculine, respectively. This was determined by asking participants to rate the following 10 sports, swimming, soccer, ballet, gymnastics, running, athletics, boxing, tennis, yoga and figure skating on a scale from 1 (feminine) to 7 (masculine). The results revealed that ballet was considered to be the most feminine sport (M = 1.89, SD = .79), and boxing to be the most masculine (M=5.92, SD = .81). A paired samples t-test with ballet and boxing showed that the sports were significantly different from each other t (35) =16.08, p <.001. It was
concluded that ballet would be used as the female appropriate sport, and boxing as the male appropriate.
Second, the pre-test was to test two drinks (i.e. Gatorade and Lipton Ice tea) to determine whether or not the drinks were a good or bad fit. To measure the product-fit a five-item scale was used with a nine-point answering scale from 1 (a good fit) to 9 (a bad fit). After they had been transformed into a new variable, the product fit of the Gatorade was higher (M = 3.24, SD = 1.99) than the product fit of the Ice tea (M = 6.24, SD = 1.71). A paired samples t-test also revealed that the products were significantly different with respect to a good or bad fit for an athlete t (35) = -6.67, p <.001. Gatorade was, therefore, used as the product with a high fit and the Ice tea was used as the product with a low fit. Participant in the pre-test did not participate in the main experiment.
Stimuli
Eight different manipulations were created for this study (Refer to Table 1 for conditions). They consisted of a picture of a female and male participating in boxing or ballet with either a logo of Gatorade and Lipton Ice tea (Refer to Image 1 and 2 for examples)2. The purpose of the logo in the pictures was to manipulate whether the product fit. Moreover, the photos had the same tagline, “The drink that helps me on my journey to become the best”, to ensure that the manipulation of products was successful.
Table 1.
Overview of experimental conditions
Condition
1 Female, inappropriate, high fit
2 Female, inappropriate, low fit
3 Female, appropriate, high fit
4 Female, appropriate, low fit
5 Male, appropriate, high fit
6 Male, appropriate, low fit
7 Male, inappropriate, high fit
8 Male, inappropriate, low fit
Procedure
When participants clicked on the online link, the participants were first briefed and informing them that the study was about advertising, and that the study was conducted in the context of a thesis research project. Furthermore, they were informed about their rights to refuse participation and that their responses would be anonymized, by not asking for any personal data. After giving their consent, participants could start the online experiment. First, they were asked about background characteristics including demographics (e.g., educational level and age). Next, participants were asked to look at an advertisement and to carefully pay attention to the text on the picture before moving on to the questions which would be about the advertisement just seen. Afterwards, the participants were asked questions on their perception of the athlete’s trustworthiness, attractiveness and expertise. Fourth, they were asked about their intention to purchase the product. Finally, participants were then guided to the final questions. First, there was a question on how appropriate it was for the athlete to
Image 2. Manipulation number 7. Image 1. Manipulation number 4.
participate in the sport shown in the advertisement to check whether the manipulation was successful. This was followed by two questions; one that asked how often the product in the advertisement was consumed, and one on how the participants would rate themselves as a fan of the sport shown in the respective advertisement. The third and last question pertained to the fit between the athlete and the product. Lastly, the participants received a debriefing.
Measures Dependent variables
Endorser credibility. The endorser credibility was measured by using Ohanian's (1990) scale, which consists of three concepts; expertise, attractiveness and
trustworthiness. All items started with the phrase “The athlete in the advertisement is…” with a nine-point semantic differential answer. For expertise, the endpoints were “not an expert-an expert”, “inexperienced-experienced”, “unknowledgeable - knowledgeable”, “unqualified-qualified and “unskilled-skilled” (M= 6.44, SD= 1.84), and Cronbach’s alpha was .92. For trustworthiness the endpoints were “undependable–dependable,” “dishonest–honest,” “unreliable–reliable,” “insincere–sincere,” and “trustworthy–untrustworthy (M= 5.37, SD= 1.99), and Cronbach’s alpha was .95. Lastly, the attractiveness scale consisted of the end-points “unattractive–attractive,’’ ‘‘not classy–classy,” “ugly–beautiful,” “plain–elegant,” and “not sexy–sexy” (M= 6.35, SD= 1.68), and Cronbach’s alpha was .88.
Purchase Intention. To measure purchase intention, Till and Busler's (2000) nine-point and three-item semantical differential scale was used, asking “how likely is it that you would consider purchasing the product in the advertisement?”. The endpoints were “definitely would not – definitely would,” “extremely unlikely – extremely likely” and “improbable – probable” (M = 3.46, SD = 1.91), and Cronbach’s alpha was .97. Control variables
Involvement. Parker and Fink's (2008) involvement scale was used to measure participants’ interest in the sport they were exposed to. A nine-point, three-item scale was used with the question “How would you rate yourself as a fan of the sport in the
advertisement”. Following endpoint were used “not a fan at all – big fan,” “do not follow closely – closely follow,” “not at all involved – very involved” (M = 3.25, SD = 2.20) and Cronbach’s alpha was .94.
Drink frequency. The drink frequency was included to ensure participants use the products. A nine-point Likert scale was used with the item “How often do you drink the product in the ad” ranking from “never” to “very often” (M = 3.45, SD = 2.71).
Manipulation checks
Product fit. Till and Busler’s (2000) athlete-product fit scale was used to measure whether or not the manipulation of the type of product was successful. A five-item with a nine-point semantic differential. With items consisting of statements: “I think the athlete and the drink…” and answer options “do not belong together – belong together,” “do not go together – go together,” “do not fit together – fit together”. The last item was “I think the athlete is an ___ endorser of the advertised product with the responses “Inappropriate – appropriate,” “Ineffective – effective”.
Classification of sport. The manipulation check ensures whether or not the sports were perceived appropriate for men and women. Therefore, participants were asked “please rate how appropriate you think it is for the athlete to participate in the sport shown in the advertisement” using the one item scaling from “not appropriate – very appropriate”.
Results Randomisation check
The purpose of the randomisation check was to ensure that participants were comparable across the experimental groups. First, a one-way ANOVA was conducted with all the
conditions as the independent variable and drinking frequency as the dependent variable. This revealed that there was a significant difference in drink frequency between the eight groups, F(7, 398) = 3.14, p = .003. Drinking frequency, therefore, served as a control variable in the main analysis. It was also tested whether there was a difference between the groups in terms of their involvement with the sports, which was also checked with a one-way ANOVA. It should be noted that Levene’s test of homogeneity of variances was not significant F(7,398) = .38, p = .917. The ANOVA showed that there was no significant difference between the experimental groups F(7, 398) = .375, p = .917, and, therefore, the involvement of sport was not served as a control variable. A chi-square test was conducted to determine if the
participants were equally distributed between the eight conditions with regard in their gender. The chi-square did not find a difference between the conditions in the distribution of gender, X2(8, N =450) = .992, p = .995. As there was no relation between the variables, gender did not serve as a control variable in this analysis.
Manipulation check
To check whether the manipulation of the product fit and the classification of sports were successful, two independent t-tests were conducted.
Product fit. The purpose of the product fit was to ensure that the athletes endorsing the Ice tea were perceived as a low fit between the endorser and the product and the athletes endorsing Gatorade perceived as a high fit. The independent variable was the type of product and the dependent variable the product fit. The t-test showed a significant effect on the product fit t(404) = 4.91, p <. 001, between the participants who were exposed to the Gatorade (M= 5.11, SD = 2.39) and the ones exposed to the Ice tea (M = 3.98, SD = 2.21). Therefore, the manipulation check of the product was successful.
Classification of sport. To assure that ballet and boxing were perceived as appropriate or inappropriate depending on gender, an independent t-test was conducted. The
type of sports (ballet or boxing) were the independent variable, whereas the dependent variable was the perceived gender appropriateness between the gender of the athlete and the sport. There was no significant difference, t(404) = -.351, p = .726, between the groups that were exposed to a ballet dancer (M = 5.00, SD = 1.90) and the ones exposed to boxers (M = 4.94, SD = 1.87).
Main analyses Attractiveness
The hypothesis predicted that endorsers in gender appropriate sports would be perceived to be more attractive (H1a), compared to endorsers in inappropriate sports, only when there is a misfit between the endorser and the product, but not when there is fit (H2a). This was tested with a three-way ANCOVA with endorsers’ gender, product fit and sport as the independent variables and attractiveness as the dependent variable. The Levene’s test revealed that the assumption of equal variances was not violated F(7,398) = 1.82, p = .271. The ANCOVA showed a significant main effect of the gender of endorser F(1,397) = 12.63, p <.001, η2 = .03. Overall, female endorsers (M = 6.66, SD = 1.57) were perceived to be more attractive than male endorsers (M = 6.08, SD = 1.72). Refer to Table 2 to see means and standard deviations.
Table 2.
Means and standard deviations of the gender of the athlete
Attractiveness Trustworthiness Expertise Purchase intention Female endorser 6.65 (1.57) 5.45 (1.92) 6.43 (1.79) 3.28 (1.95) Male endorser 6.08 (1.72) 5.34 (2.03) 6.49 (1.83) 3.63 (1.86)
The main effect of appropriateness of sport was not significant F(1,397) = .943, p = .332, η2 =.00. Refer to Table 3, to see means and standard deviations.
Table 3.
Means and standard deviations of the sport appropriateness
Attractiveness Trustworthiness Expertise Purchase intention Appropriate 6.43 (1.57) 5.49 (1.91) 6.60 (1.75) 3.42 (1.96) Inappropriate 6.30 (1.76) 5.33 (2.03) 6.32 (1.86) 3.49 (1.86)
The main effect of product fit also showed a significant effect on how attractive the athlete was perceived to be F(1,397) = 10.19, p = .047, η2 =.01. When there was high product fit (M = 6.47, SD = 1.58), the athlete was perceived as more attractive than when there was low product fit (M = 6.26, SD = 1.74).
The interaction effect between the gender of the endorser and the
appropriateness of sport on the attractiveness was found to be significant, F(1,397) = 15.29, p <.001, η2
=.04. The female endorsers participating in a gender appropriate sport (M = 6.30, SD = 1.55) were perceived to be more attractive than male endorsers participating in gender appropriate sport (M = 6.29, SD =1.95). Similarly, the female endorser participating in a gender inappropriate sport (M = 6.95, SD = 1.52) was also perceived to be more attractive than the male endorser participating in a gender inappropriate sport (M = 5.88, SD = 1.43). Refer to table 4 to see which groups differed significantly from each other.
Table 4.
Interaction effect of the genders of endorser and appropriateness of sport
Appropriate Inappropriate
Male endorser 6.29 (1.95)a 5.88 (1.43)c Note. ap <.001,bp =.002, cp=.005.
The two-way interaction between gender of the endorser and product fit was found not to be significant, F(1,397) = .02, p = .866, η2 =.00. The results indicate that there was no difference between a male endorser with a high fit (M = 5.43, SD = 1.65), male endorser with a low fit (M = 5.73, SD = 1.80), female endorser with a high fit (M = 7.05, SD = 1.35). and a female endorser with a low fit (M = 6.28, SD = 1.66).
The two-way interaction between the sport appropriateness and the product fit was not significant, F(1,397) = 2.12, p = .146, η2 =.00. Refer to Table 5 for means and standard deviations.
Table 5.
Interaction effect of appropriateness of sport and product fit.
Appropriate Inappropriate
High fit 6.81 (1.60) 6.65 (1.59)
Low fit 5.78 (1.77) 6.23 (1.62)
The three-way ANCOVA also revealed that the interaction effect between the gender of endorser, appropriateness of sport and product fit on the attractiveness was not significant, F(1,397) = .017, p = .866, η2 =.00. The means and standard deviations of all the eight groups on attractiveness are provided in Table 6.
Table 6.
Means and standard deviations on dependent variables
Attractiveness Trustworthiness Expertise Purchase Intention Female, inappropriate, 6.54 (1.92) 5.84 (2.01) 7.18 (1.61) 4.65(1.76)
high fit
Female, inappropriate, low fit
6.04 (1.97) 5.06 (2.00) 6.60 (1.93) 2.50 (1.47)
Female, appropriate, high fit
5.96 (1.29) 5.32 (1.99) 6.09 (1.79) 3.38 (2.11)
Female, appropriate, low fit
5.81 (1.33) 5.29 (2.02) 6.10 (1.80) 2.36 (1.51)
Male, appropriate, high fit
6.92 (1.29) 5.67 (1.69) 6.79 (1.63) 4.41 (1.90)
Male, appropriate, low fit 6.97(1.73) 5.67 (1.97) 7.36 (1.48) 2.99 (1.68) Male, inappropriate, high
fit
6.44 (1.42) 5.48 (1.99) 5.94 (1.81) 3.98 (1.94)
Male, inappropriate, low fit
6.13 (1.16) 4.91 (1.98 5.49 (1.67) 3.14 (1.57)
Trustworthiness
The hypotheses predicted that endorsers in gender appropriate sports are perceived to be more trustworthy (H1b) compared to endorsers in inappropriate sports, and only when there is a misfit between the endorser and the product and not when there is fit (H2b). Similar to the attractiveness, a three-way ANCOVA was carried out the with the gender of the endorser, product fit and sport appropriateness as independent variables, and the trustworthiness as dependent variable. The assumptions of equal variances were not violated F(7,398) = 1.460, p = .180.
The ANCOVA showed that there was no significant main effect of the gender of endorser on trustworthiness, F(1,397) = .708, p = .400 η2 = .002. See Table 2 for means and standard deviations.
The main effect of appropriateness of sport was also found not to be significant, F(1,397) = 1.80, p = .181, η2 = .005. Overall, there was no differences between the endorsers that participated in a gender appropriate sport (M = 5.46, SD =1.94) and endorsers
participating in gender inappropriate sports (M = 5.29, SD = 2.05). See Table 3 for means and standard deviations.
The main effect of the product fit differed significantly F(1,397) = 10.78, p = .001, η2 = .03, which indicated that the endorsers were perceived to be more trustworthy when there was a high product fit (M = 5.58, SD = 1.93) relative to a low product fit (M = 5.25, SD = 2.00).
The two-way interaction between the gender of the endorser and the
appropriateness of sport was not significant F(1,397) = 2.33, p = .128, η2 = .00. This result means that there was no difference in their effect on trustworthiness between female endorsers in appropriate sports (M = 5.71, SD = 1.75), female endorsers in inappropriate sports (M = 7.06, SD = 1.58), male endorsers in appropriate sports (M = 6.90 SD = 1.79) and male endorsers in inappropriate sports (M = 6.10, SD = 1.79).
The ANCOVA showed that there was no interaction effect between the gender of the endorser and product fit, F(1,397) = 9.36, p = .334, η2 = .00. This groups that did not differ from each other was male endorsers with a high fit (M = 6.27 SD = 1.73), male endorsers with a low fit (M = 4.56, SD = 1.90), female endorsers with a high fit (M = 6.32, SD = 1.73) and female endorsers with a low fit (M = 4.66, SD = 1.73).
There was no significant interaction effect of the appropriateness of sport and the product fit, F(1,397) = 1.754, p = .186, η2 = .00, which means that there was no
difference between the following groups, endorsers in an appropriate sport with a high fit (M = 6.32, SD = 1.81), endorsers in an appropriate sport with a low fit (M = 4.34, SD = 1.75), endorsers in an inappropriate sport a high fit (M = 6.28, SD = 1.66) and endorsers in an inappropriate sport a with a low fit (M = 4.77, SD = 1.86).
The ANCOVA indicated that the three-way interaction effect between the gender of endorser, the appropriateness of sport and product fit was not significant, F(1,397) = .191, p = .662, η2 = .000. The means and standard deviations of the experimental groups on attractiveness are provided in Table 6.
Expertise
For H1c and H2c, a three-way ANCOVA was conducted. They predicted that endorsers in gender appropriate sports are perceived to have higher expertise compared to endorsers in gender inappropriate sports, when there is a misfit, but not when there is fit between the endorser and the product (H2c).
The main effect of the gender of the endorser on expertise was not significant, F(1,397) = .292, p = .589 η2 = .00. See Table 2 for means and standard deviations.
The main effect of appropriateness of sport was also not significant, F(1,397) = 3.83, p = .05, η2 = .01. See Table 3 for means and standard deviations. However, the main effect of the product fit was again found to be significant, F(1,397) = 30.38, p <.001, η2 = .07. This means that the endorser was perceived to have a higher expertise when there was a high product fit (M = 6.30, SD = 1.73) compared to when there was a low product fit (M = 6.26, SD =1.75).
The two-way interaction effect between the gender of the endorser and the appropriateness of the sport was statistically significant F(1,397) = 52.42, p <.001, η2 = .11. Female endorsers participating in a gender appropriate sport (M = 7.07, SD =1.58) were perceived to have a higher expertise level than females in a gender inappropriate sport (M =
5.71, SD = 1.75) than male endorsers participating gender appropriate sport (M =6.10, SD =1.79) and inappropriate sport (M= 6.90, SD = 1.79). However, female endorsers
participating in a gender inappropriate sport were perceived to have a lower expertise level than male endorsers in an inappropriate sport. Moreover, in gender inappropriate sports, male endorsers were perceived to have a higher expertise level than female endorsers. There was no difference between female endorsers in a gender inappropriate sport and male endorsers in gender appropriate sports. Refer to Table 7 to see which groups differed significantly from each other.
Table 7.
Means and standard deviations for interaction effect of gender of the endorser and appropriateness.
Appropriate Inappropriate
Female endorser 7.07 (1.58)abc 5.71 (1.75)ad
Male endorser 6.10 (1.79)be 6.90 (1.79)cde
Note. abcde p <.001
The interaction effect between the gender of the endorser and product fit was found to not be significant, F(1,397) = .04, p = .840, η2 = .00. The result showed that there was no difference between the male endorsers with a high fit (M = 6.95, SD = 1.66), male endorsers with a low fit (M = 6.02, SD = 1.88), female endorsers with a high fit (M = 6.90, SD = 1.52) and female endorser with a low fit (M = 6.02, SD = 1.91).
The interaction between the appropriateness of sport and product fit had a significant effect, F(1,397) = 5.06 p = .025, η2 = .01. Endorsers who endorsed a product with a low fit were perceived to have a higher expertise when participating in a gender appropriate sport (M = 6.70, SD =1.87) compared to endorsers participating in gender inappropriate sport
(M = 6.07, SD =1.89). Refer to Table 8 to see which groups differed significantly from each other.
Table 8.
Means and standard deviations interaction effect between product fit and appropriateness
Appropriate Inappropriate
High product fit 6.45 (1.74) 6.52 (1.88)
Low product fit 6.70 (1.87) a 6.07 (1.89) a
Note. a p =.016
The three-way interaction between gender of the endorser, appropriateness of sport and product fit did not yield a significant interaction effect F(1,397) = 1.91, p = .168, η2 = .00. The means and standard deviations of all the eight groups on attractiveness are
provided in Table 6. Purchase Intention
The last hypothesis (H3) predicted that using endorsers in gender appropriate sports will lead to a higher purchase intention compared to endorsers in gender-inappropriate sports, when there is a misfit between the endorser and the product, but not when there is fit. To do so, a three-way ANCOVA was conducted with the independent variables, endorser’s gender, sports appropriateness and product fit, while the purchase intention served as the dependent variable. The Levene’s test specified that the assumption of homogeneity of variances was not violated F(7,398) = 1.50, p = .166.
The main effect of the gender of the endorser was not significant F(1,397) = 1.48, p = .224, η2
= .01. Refer to Table 2 for means and standard deviations.
The main effect of the sport appropriateness, F(1,397) = 3.25, p = .072, η2 = .00, was also not significant. Refer to Table 3 for means and standard deviations. However, a
significant main effect of product fit was found, F(1,397) = 41.58, p <.001, η2 = .10. The high fit (M = 4.19, SD = 1.95) had a stronger effect on purchase intention than the low fit (M = 2.74, SD = 1.58).
The ANCOVA revealed that there was no significant interaction effect between the gender of endorser and sport appropriateness F(1,397) = 1.48, p =.225, η2 = .00. Refer to Table 9 for means and standard deviations.
Table 9
Means and standard deviations interaction effect between gender of the endorser and appropriateness
Appropriate Inappropriate
Female 3.19 (2.00) 3.37 (1.91)
Male 3.62 (1.91) 3.62 (1.82)
The interaction effect between the gender and the product fit was also found not to be significant, F(1,397) = .225, p =.635, η2 = .00. Refer to Table 10 for means and
standard deviations. Table 10.
Means and standard deviations interaction effect between product fit and appropriateness
High fit Low fit
Female 3.37 (1.91) 3.90 (1.78)
Male 3.13 (2.03) 3.44 (1.86)
There was no interaction effect between the sport appropriateness and the product fit, F(1,397) = .00, p =.949, η2 = .00, which means that there was no difference
between a high and low product fit amongst endorsers appropriate sports and inappropriate sports. Refer to Table 11 for means and standard deviations.
Table 11.
Means and standard deviations interaction effect between product fit and appropriateness
Appropriate Inappropriate
High product fit 3.16 (1.89) 3.33 (2.06)
Low product fit 3.82 (1.79) 3.50 (1.87)
Lastly, the three-way ANCOVA revealed that no significant interaction was found between the gender of endorser, sport appropriateness and product fit F(1,397) = .00, p = .949, η2 = .00. Means and standard deviations are presented in Table 6.
Discussion and implications
The purpose of this research study was to examine the gap between female and male athletes in advertising. Hence, the effectiveness between male and female athletes as endorsers was investigated, in particular the influence of the gender of endorsers, with sport appropriateness and product fit on both the credibility of the endorser and the purchase intention.
Furthermore, the purpose was to investigate the difference between male and female athletes as endorsers, and their effectiveness.
The first hypothesis predicted that athlete endorsers that appeared in gender appropriate sports (boxing for males and ballet for females) would be perceived as more credible (i.e. attractive, trustworthy, with a higher expertise level), resulting in an increased purchase intention of the product advertised. It was expected that endorsers in gender
appropriate sports would be perceived with more overall credibility then endorsers in gender inappropriate sports. It was found that there was no difference between the endorsers in
gender appropriate sports compared to inappropriate on the credibility variables, nor the purchase intention.
The second hypothesis predicted that endorsers in gender appropriate sports would be perceived as more credible when there was a product fit compared to endorsers in gender inappropriate sports. The endorsers in gender appropriate sports were found not to be more attractive, trustworthy or have a higher purchase intention of the product compared to endorsers in gender inappropriate sports. However, a significant difference on the interaction effect between the sport appropriateness and the product fit was found. More specifically, the experimental group exposed to an endorser in a gender appropriate sport perceived the endorser to have a higher expertise level than an endorser in a gender inappropriate sport, but only there was a low fit.
The results indicated that the gender of the endorser, in general, only had a significant difference on the credibility variable, attractiveness. Female athletes were perceived to be more attractive than male athletes. Moreover, the interaction effect of the gender of the athlete and appropriateness of sport was significant on expertise and
attractiveness. When female endorsers participated in an appropriate sport, they were perceived as more attractive than females in an inappropriate sport. Additionally, males in both appropriate and inappropriate sports were perceived to be less attractive than the female endorsers in a gender appropriate sport. Female endorsers in gender inappropriate sports were perceived to have a higher expertise level than male endorsers participating in a gender appropriate sport and both female and male endorsers in gender inappropriate sport.
The results of this study have implications for potential change in gender norms and its influence on effectiveness of athlete endorsers in following areas; appropriateness of sport, combinations of appropriateness of sport with product, and the gender of an endorser.
It was hypothesized that an athlete participating in a gender appropriate sport would be perceived as more attractive, trustworthy and with more expertise than athletes in a gender inappropriate sport (Kane, 1988; Amos et al., 2008; Fink et al., 2012). However, no difference was found between endorsers participating in perceived gender inappropriateness and gender appropriate sports. Contrary to the findings of this study, findings in previous research suggest that athlete endorsers that participate in a gender inappropriate sport are less attractive, trustworthy and less of an expert than athlete endorsers who participate in gender appropriate sports (Kane, 1988; Amos et al., 2008; Fink et al., 2012). Part of the findings in this study are in line with the findings of Fink et al. (2012), who also found that the perceived expertise level did not differ amongst endorsers in inappropriate compared with appropriate sports. This means that H1c was partially accepted, however, not enough evidence was found to support H1ab.
In the combination of appropriateness and product fit, the results indicated a gender appropriate sport and a high product fit would result in an endorser as more attractive, trustworthy and as an expert compared to the combination of an endorser in a gender
inappropriate sport and a low product fit. Contrary to expectations, an endorser in a gender appropriate sport with a high product fit did not result in the endorser being more attractive, trustworthy, have a higher level of expertise or a higher purchase intention. Till and Busler (2000) found that with a good fit between product and endorser and, thus, a better endorser than if a bad fit was present. This results in the rejection of H2a, H2b and Hc H2d.
The findings of female endorsers in gender appropriate sports being perceived as more attractive than male endorsers are in line with previous research (Till & Busler, 2000; Fink et al., 2012). Although, literature predicts that there should be a difference between male endorsers in gender appropriate and inappropriate sports, the difference was not found in this study. However, the investigation of the effectiveness of male endorsers
participating in gender inappropriate sport is limited, and further studies should examine this. Thus, these findings contribute to the theoretical gap of the effectiveness of male athletes and appropriateness.
However, the results of this study might also have implication for the
effectiveness of an endorser. For the majority of the analysis, there was no difference found between female and male endorsers, also when other factors such as product fit and sport appropriateness are considered. The limited findings could be explained by the change in gender norms (Koivula, 2001; Fink et al., 2012). The act of females and males is bound by social and historical constructions. These can be influenced and developed by social
interaction between a female and male and their relations to each other. The change in gender norms may also explain why the appropriateness of sport for females and males did not influence the credibility heavily. More research, however, is necessary to examine the salience of gender stereotypes in athlete endorsers. The cut between whether a sport is appropriate or inappropriate for a woman or man may not be as salient anymore. Limitations
Current study also included various limitations. First, to rule out any pre-existing opinions, both athletes were fictional, and the advertisement was created for this purpose only. This, however, could have affected the study in such way that the athletes could have been perceived as not being professional athletes. This could also be an explanation for why no significant difference was found between the male and female athlete.
Second, to obtain a large sample under time and resource constraints, a convenience sampling was used. This could result in a bias such as an under- or
overrepresentation of a group in the population. The largest group was Danes (27%), whereas the next largest was Americans (17%), whereas some countries are represented with less than
1%. This could affect the external validity of this research, as it may struggle to represent the general population.
Third, a nine-point semantic differential scale was used to measure several concepts e.g. product fit, attractiveness, trustworthiness and expertise. The more answering options there are, the lower scores will be produced (Dawes, 2008). This may also cause confusion for participants struggling to understand the difference between, e.g. four and five.
Conclusion
When choosing endorsers for a specific advertisement, it is of crucial importance that the fit between the product and the endorser is good to obtain greatest amount of effectiveness. This study shows that the product fit is one of the most important variables in the effectiveness of an endorser, as the product fit lead the endorsers to be more attractive, trustworthy, have a higher expertise level and the intention to purchase the endorsed product. The gender of the athlete was also an important aspect of this study. However, no difference was found on the effectiveness of whether the athlete endorser was a man or a woman. When a female athlete was participating in a gender appropriate sport, she was found to be perceived as the most attractive endorser, and the highest expertise level compared to male athletes in gender appropriate sports and in gender inappropriate sports both female and male athletes. Moreover, the product fit in combination with the appropriateness was significant, but this only occurred to the low product fit. Overall, this research point towards a change in gender stereotypes since there was no difference between males and females participating in gender inappropriate sports on their credibility.
References
Amos, C., Holmes, G., & Strutton, D. (2008). Exploring the relationship between celebrity endorser effects and advertising effectiveness: A quantitative synthesis of effect size. International Journal of Advertising. https://doi.org/10.1080/02650487.2008.11073052 Barner, M. R. (1999). Sex-role stereotyping in FCC-mandated children’s educational
television. Journal of Broadcasting and Electronic Media, 43(4), 551–564. https://doi.org/10.1080/08838159909364509
Belch, G. E., & Belch, M. A. (2018). Advertising and Promotion: An Integrated Marketing Communications Perspective (11th Editi, Vol. 65). New York: McGraw Hill Education. Belch, G. E., & Belch, M. A. (2013). A content analysis study of the use of celebrity
endorsers in magazine advertising. International Journal of Advertising. 32(3), 369-389. Bem, S. L. (1981). Gender schema theory: A cognitive account of sex typing. Psychological
Review. https://doi.org/10.1037/0033-295X.88.4.354
Boyd, T. C., & Shank, M. D. (2004). Athletes as product endorsers: The effect of gender and product relatedness. Sport Marketing Quarterly, 13(2), 82–93.
Csizma, K. A., Wittig, A. F., & Schurr, K. T. (1988). Sport Stereotypes and Gender. Journal of Sport & Exercise Psychology. https://doi.org/10.1123/jsep.10.1.62
Cunningham, G. B., Fink, J. S., & Kenix, L. J. (2008). Choosing an endorser for a women’s sporting event: The interaction of attractiveness and expertise. Sex Roles.
https://doi.org/10.1007/s11199-007-9340-z
Dawes. J. (2008). Do Data Characteristics Change According to the Number of Scale Points Used? An Experiment Using 5-Point, 7-Point and 10-Point Scales. International Journal of Market Research, 50(1), 61–104. https://doi.org/10.1177/147078530805000106 Eagly, A. H. (1987). Sex differences in social behavior: A social-role interpretation. Eagly, A. H., & Wood, W. (2012). Social role theory. In Handbook of Theories of Social
Eisend, M. (2010). A meta-analysis of gender roles in advertising. Journal of the Academy of Marketing Science. https://doi.org/10.1007/s11747-009-0181-x
Eisend, M., & Langner, T. (2010). Immediate and delayed advertising effects of celebrity endorsers attractiveness and expertise. International Journal of Advertising.
https://doi.org/10.2501/s0265048710201336
Erdogan, B. Z. (1999). Celebrity Endorsement: A Literature Review. Journal of Marketing Management. https://doi.org/10.1362/026725799784870379
Fink, J. S., Cunningham, G. B., & Kensicki, L. J. (2004). Using Athletes as Endorsers to Sell Women’s Sport: Attractiveness vs. Expertise. Journal of Sport Management, 18(4), 350–367. https://doi.org/10.1123/jsm.18.4.350
Fink, J. S., Parker, H. M., Cunningham, G. B., & Cuneen, J. (2012). Female athlete endorsers: Determinants of effectiveness. Sport Management Review, 15(1), 13–22. https://doi.org/10.1016/j.smr.2011.01.003
Ganahl, D. J., Prinsen, T. J., & Netzley, S. B. (2003). A Content Analysis of Prime Time Commercials: A Contextual Framework of Gender Representation. Sex Roles. https://doi.org/10.1023/A:1025893025658
Hardin, M., & Greer, J. D. (2009). The Influence of Gender-role Socialization , Media Use and Sports Participation on Perceptions of Gender-Appropriate Sports. Journal of Sport Behavior.
Holland, A., & Andre, T. (1994). Athletic Participation and the Social Status of Adolescent Males and Females. Youth & Society. https://doi.org/10.1177/0044118X94025003005 Kamins, M. A. (1990). An investigation into the “match-up” hypothesis in celebrity
advertising: When beauty may be only skin deep. Journal of Advertising. https://doi.org/10.1080/00913367.1990.10673175
of high school adolescents. Adolescence.
Kanungo, R. N., & Pang, S. (1973). Effects of human models on perceived product quality. Journal of Applied Psychology. https://doi.org/10.1037/h0037042
Kian, E. M., & Hardin, M. (2009). Framing of Sport Coverage Based on the Sex of Sports Writers: Female Journalists Counter the Traditional Gendering of Media Coverage. International Journal of Sport Communication.
Koernig, S. K., & Boyd, T. C. (2009). To catch a tiger or let him go: The match-up effect and athlete endorsers for sport and non-sport brands. Sports Marketing Quarterly.
Koivula, N. (1995). Ratings of gender appropriateness of sports participation: Effects of gender-based schematic processing. Sex Roles. https://doi.org/10.1007/BF01544679 Koivula, N. (1999). Gender stereotyping in televised media sport coverage. Sex Roles.
https://doi.org/10.1023/A:1018899522353
Koivula, N. (2001). Perceived characteristics of sports categorized as gender-neutral,
feminine and masculine. Journal of Sport Behavior. Stockholm, Sweden.: University of South Alabama.
Lynch, J., & Schuler, D. (1994). The matchup effect of spokesperson and product congruency: A schema theory interpretation. Psychology & Marketing. https://doi.org/10.1002/mar.4220110502
Metheny, E. (1965). Connotations of movement in sport and dance: a collection of speeches about sport and dance as significant forms of human behavior.
Ohanian, R. (1990). Construction and validation of a scale to measure celebrity endorsers’ perceived expertise, trustworthiness, and attractiveness. Journal of Advertising. https://doi.org/10.1080/00913367.1990.10673191
Parker, H. M., & Fink, J. S. (2008). The effect of sport commentator framing on viewer attitudes. Sex Roles. https://doi.org/10.1007/s11199-007-9344-8
Petrecca, L. (2006). Product placement you can’t escape it. USA TODAY.
Riemer, B. A., & Visio, M. E. (2003). Gender typing of sports: An investigation of metheny’s classification. Research Quarterly for Exercise and Sport.
https://doi.org/10.1080/02701367.2003.10609081
Statista. (2018a). Athletes ranked by earnings from salary and endorsements worldwide in 2017/18 (in million U.S. dollars)*. Retrieved 12 January 2019, from
https://www.statista.com/statistics/250295/highest-paid-athletes-worldwide/ Statista. (2018b). Female athletes with the highest income from prize money and
endorsements in 2018 (in million U.S. dollars).
Till, B. D., & Busler, M. (2000). The match-up hypothesis: Physical attractiveness, expertise, and the role of fit on brand attitude, purchase intent and brand beliefs. Journal of
Advertising. https://doi.org/10.1080/00913367.2000.10673613
Van Riper, T., & Badenhausen, K. (2008). Top-Earning Female Athletes. Forbes. Wright, S. A. (2016). Reinvestigating the endorser by product matchup hypothesis in
Appendix A Stimuli 1: Female, inappropriate, high fit
Stimuli 4: Female, appropriate, low fit
Stimuli 6: Male, appropriate, low fit
