Arrest record or openly gay: the impact of athletes' personal lives on endorser effectiveness.

Parker, Heidi M. ; Fink, Janet S.

Introduction

Companies often use celebrity endorsers in an effort to help sell their products and promote their brands. Athletes have benefited tremendously from these relationships, and many make more money from their endorsement deals than their player salaries or winnings. For instance, Phil Mickelson earned $52,000,000 from his endorsement deals from July 2009 to July 2010 while only earning 18% of that amount ($9,660,757) from winnings (Freedman, 2010). Similarly, LeBron James earned nearly twice the amount of money through endorsement deals, $30,000,000, during the 2009-2010 season relative to what he was paid in his player contract, $15,779,912 (Freedman, 2010).

The financial investments companies make when using high-profile athletes to endorse their products do not come without risk. In 2003, many companies were left contemplating their endorsement relationship with Kobe Bryant after he was charged with sexual assault. A similar situation occurred in the summer of 2007 when the popular Michael Vick was indicted on federal dog fighting charges. Recently, Tiger Woods, who earned $105,000,000 in endorsement money in 2008 (Freedman, 2008), saw his image tarnished as multiple accounts of infidelity became public and shed light onto a less than perfect personal life, shattering his carefully crafted "family man" image. In light of the accounts and fearing negative repercussions, several of Woods' sponsors cut ties with the golfer, including Accenture, Gatorade, and AT&T. Nike, however, chose to keep Woods as an endorser and, according to a recent report by the Tepper School of Business (2010), lost an estimated $1.3 million in profits due to the Woods scandal. Woods himself lost an estimated $35 million dollars in endorsement revenue between 2008 and 2010 (Freedman, 2010).

There are other factors, beyond an athlete's criminal history, that also influence endorsement decisions. One of these is the athlete's sexual orientation. Although attitudes have generally improved, many Americans still express opposition to homosexuality, sexual minorities, and lesbian, gay, bisexual, and transgender (LGBT) communities (see Herek, 2009). Heterosexism is especially evident in sport settings and among athletes (Cunningham, Sartore, & McCullough, 2010; Sartore & Cunningham, 2009). Thus, lesbian and gay athletes are often left with the difficult choice of living life openly and honestly with potential financial consequence or remaining closeted in an effort to protect their marketability and financial viability. Marketers must also decide if partnering with a gay athlete to endorse their product will have a negative impact on the product brand image or company bottom line.

Therefore, the purpose of this study was to examine the ramifications to an athlete's image and/or marketability experienced from engaging in illegal behaviors or choosing to be an "out" gay or lesbian athlete (which many still classify as "immoral" and harbor prejudices against). Specifically, this study examined the consequence of having an arrest record or identifying as a gay or lesbian athlete on the perceived endorser characteristics of expertise, attractiveness, trustworthiness, and the impact on endorser-product fit. In other words, how might an openly gay athlete or one with a prior arrest record be received as an endorser of a particular product? Does being openly gay or being involved in an illegal activity diminish the athlete's source credibility? Is what some view as "immoral" behavior viewed differently than an illegal behavior? The following section will further discuss source credibility and the match-up hypothesis in relation to endorsers as well as present specific hypotheses of the study.

Theoretical Framework

Endorser Effectiveness

Much of the endorser literature has involved application of the match-up hypothesis (Kahle & Homer, 1985; Kamins, 1990) and examination of endorser, or source, credibility (Dholakia & Sternthal, 1977; Hovland & Weiss, 1951; Ohanian, 1991). The match-up hypothesis suggests endorsements are most successful when there is a natural fit between the product and the endorser and that such a fit is essential for endorser effectiveness (Boyd & Shank, 2004; Till & Busler, 2000).

Source credibility has also been shown to be an important aspect of endorser effectiveness, and the more credible the perception of the endorser, the more persuasive their message (Amos, Holmes, & Strutton, 2008; Dholakia & Sternthal, 1977; Hovland & Weiss, 1951; Ohanian, 1991). Specifically, trustworthiness, expertise, and attractiveness are characteristics of source credibility that have been shown to have the most positive influence on endorser effectiveness (Amos et al., 2008; Cunningham, Fink, & Kenix, 2008; Fink, Cunningham, & Kensicki, 2004).

Trustworthiness, defined as "the listener's degree of confidence in, and level of acceptance of, the speaker and the message" has long been viewed as a key component in persuasive communication (Ohanian, 1990, p. 41). Expertise refers to knowledge, experience, competence, or skills that an endorser possesses (Erodgan, 1999; Ohanian, 1990), while attractiveness, a widely studied construct, has been shown to increase the persuasiveness of a communicator (Ohanian, 1990).

Credibility is typically the result of the consumers' perception of the endorser's combined expertise, attractiveness, and trustworthiness. For instance, an endorser may be seen as an expert (i.e., an athlete endorsing a sport related product) but they may not have a reputation of being trustworthy and/or they may not be that attractive. Likewise, an endorser may be very attractive and yet not be considered an expert (i.e., an athlete endorsing an automotive brand). Thus, each of the three characteristics provides an important component of the overall credibility and the more credible the source is deemed to be, the more influential and persuasive their message (Ohanian, 1991). Obviously, the ideal endorser would be viewed as simultaneously attractive, trustworthy, and expert. However, an endorser may still be effective, even if they fall short in one or more of those areas, if consumers view them as credible through the combined effect of those characteristics.

Personal Information

While the use of celebrity endorsers can have great reward--research has shown the announcement of celebrity endorsement contracts to positively increase stock returns (Agrawal & Kamakura, 1995; Louie, Kulik, & Jacobson, 2001)--the risk of the endorser engaging in an "undesirable event" or scandal clearly exists. Negativity effect suggests individuals pay more attention and place more value on negative information than positive information (Fiske, 1980; Klein, 1996) while Louie et al. (2001) found a company can experience a decrease in stock returns as a result of an undesirable act committed by a company endorser.

This rubbing off of attitudes from one entity to another, known as the spillover effect, can occur when an organization joins forces with a spokesperson (Simonin & Ruth, 1998; Votolato & Unnava, 2006). When one of the two entities is involved in an undesirable event and shrouded with negative press, a spillover can occur and consequently tarnish the image of the partnering organization.

For instance, Till and Shimp (1998) examined the spillover from a celebrity endorser to the endorsed brand. Results revealed negative information about a fictitious endorser resulted in a negative spillover to the endorsed brand when the endorser was perceived as unfavorable. Likewise, Votolato and Unnava (2006) considered various types of reported negative information on the spillover of attitudes between associations. Specifically, Votolato and Unnava contrasted negative information perceived as immoral with negative information which highlighted incompetence. Their results found that if the partnering alliance was a spokesperson a greater negative spillover occurred when the negative information involved an issue of immorality. Kuzma et al. (2003) examined negative spillover from sponsor to team and found decreased attitudes toward the sponsored team, the related university, and additionally decreased purchase intentions of team/university gear.

The above results suggest that involvement in an undesirable event can tarnish the reputation of an endorser as well as spillover onto the product or brand being endorsed, subsequently reducing the effectiveness of the endorser. Given the potential impact of a personal scandal to damage the perceived trustworthiness of an endorser and the importance of this source credibility characteristic on endorser effectiveness, (Amos et al., 2008), we propose the following hypothesis:

Hypothesis 1: An athlete endorser who participates in an undesirable behavior will be rated less favorably on measures of trustworthiness than an athlete with identical credentials who has not engaged in undesirable acts.

The question then becomes, is being an openly gay athlete currently viewed by consumers as undesirable, as it relates to their potential endorsement opportunities? There is anecdotal evidence to suggest gay athletes are reluctant to "come out" due to fear of losing endorsement opportunities. In 2008, Majendie reported only 10 of more than 10,000 Olympic athletes in Beijing were openly gay, which many speculate is, in part, due to fear of losing sponsorships and endorsements. Dean Bonham, a sports marketing executive, stated, "The question isn't whether coming out would have a negative impact on an athlete as an endorser. The question is, how much of a negative impact" (as cited in Wertheim, 2005).

In 2005, Penn, Schoen, & Berland Associates conducted a poll to gauge public opinion of gay athletes. While 86% of the participants felt it was OK for openly gay males to participate in sports, almost 25% of participants felt having an openly gay player would hurt a team (Wertheim, 2005). Additionally, 68% of the participants felt being openly gay would hurt a player's career, 64% believed gay athletes were less likely to be selected as endorsers of brands or products, and 18% said they would be less likely to purchase footwear endorsed by a gay athlete (Wertheim, 2005).

Empirical data also supports the notion that gay athletes may be perceived as less credible. Knight and Giuliano (2003) conducted a study in which participants read a fictitious newspaper article about an Olympic athlete which portrayed the athlete as "clearly heterosexual" (i.e., mentions athlete's spouse) or as having an ambiguous sexual orientation (i.e., mention of a roommate rather than of a spouse). Results indicated that participants had more favorable impressions of the athlete who was described as clearly heterosexual than of the athlete whose sexual orientation was ambiguous (Knight & Giuliano, 2003).

Additionally, Cunningham et al. (2010) examined perceptions of applicants for personal training positions at a fitness club. Results found that men rated lesbian and gay applicants less favorably in terms of their morality, trustworthiness, and ethicality values than their straight counterparts. They also found these ratings to be linked with hiring intentions. Researchers also found athletes less willing to play for a gay or lesbian coach and parents less willing to allow their child to play for a gay or lesbian coach (Sartore & Cunningham, 2009).

Given the above information regarding openly gay and lesbian athletes and employees, we propose our second hypothesis:

Hypothesis 2: An openly gay athlete endorser will be rated less favorably on measures of trustworthiness than a straight athlete with identical credentials.

Additionally, the match-up hypothesis suggests congruence, or "fit," between the product and the endorser is also imperative to endorser effectiveness. The relevance of the match-up hypothesis can be further understood when examined through the lens of the associative learning theory. Associative learning theory suggests that individual and/or distinct bits of information and concepts can become linked with one another when repeatedly paired together such that the recollection of one of the concepts automatically recalls the associated other (Till & Busler, 2000). For instance, baseball and hotdogs are two distinct products, which, due to repeated pairing, have become associated with one another. The thought of attending an afternoon baseball game automatically brings about the thought of hotdogs.

The same linked association can be established between an endorser and a product. However, the ease in building the link is heavily dependent on how intuitively the two entities fit together (Kamins, 1990; Lynch & Schuler, 1994). For instance, athletes were found to be more effective endorsers than actors in advertisements for an energy bar (Till & Busler, 2000). A model posing as an unidentified famous athlete was rated as a more effective endorser when paired with a sport-related brand than when paired with a non-sport-related brand (Koernig & Boyd, 2009). And, women rated athlete endorsers as more effective when they endorsed athletic shoes than when they endorsed a non-athletic-related product (Boyd & Shank, 2004).

There have been a few studies that examined how source credibility characteristics relate to the fit of the athlete as an endorser for a product. Fink et al. (2004) contended that, according to the match-up hypothesis, athlete expertise should be more important to perceived fit than athlete attractiveness. Their experimental study of female athlete endorsers found that expertise was more important to perceived fit than athlete attractiveness. Cunningham et al. (2008) found an interaction between the two source credibility variables in their study of female athlete endorsers. When athlete expertise was high, attractiveness was not a significant predictor of fit, but when the athlete was perceived as having less expertise, attractiveness did significantly contribute to fit. Fink, Parker, Cunningham, and Cuneen (in press), however, found that only perceived trustworthiness, not attractiveness or expertise, contributed significantly to endorser fit in a study of female athletes endorsing a sport drink. Further, in a study of four New Zealand athletes (two males, two females) Charbonneau and Garland (2006) found that both of the female athletes were deemed more attractive than their male counterparts, none of the athletes received significantly higher scores on expertise compared to one another, and one female athlete was considered significantly more trustworthy than the male athletes. This female athlete had the highest total score (attractiveness, expertise, trustworthiness combined) and she was deemed the best "fit" for endorsing a sport-related product (a sport drink).

Thus, the literature related to source credibility characteristics and athlete endorser fit has been somewhat mixed in terms of which characteristic is most important to fit. A recent meta-analysis of the general endorser effectiveness literature (not just athlete endorsers) examined a variety of endorser traits and found trustworthiness, expertise, and attractiveness (in that order) had the most positive influence on endorser effectiveness (Amos et al., 2008). Obviously, higher levels of each characteristic should result in higher levels of perceived fit. However, in alignment with the match-up hypothesis, we contend that trustworthiness and expertise should be more important predictors of fit than attractiveness when the athlete is endorsing a sport product.

Hypothesis 3: Perceptions of the athlete's trustworthiness, attractiveness, and expertise will positively affect measures of fit with athlete expertise and trustworthiness contributing the most to perceptions of fit.

Finally, endorser product fit appears to be the most important variable relative to purchase intentions, and it explains a significant amount of variance in purchase intentions beyond the source credibility characteristics (Cunningham et al., 2008; Fink et al., 2004; Till & Busler, 2000; Veltri, Kuzman, Stotlar, Viswanathan, & Miller, 2003). Thus, rather than expertise, trustworthiness, and attractiveness being the sole influences regarding attitudes toward a product, the fit of the endorser to the product is vital and, conceptually, goes beyond these three characteristics. For example, Cunningham et al. (2008) suggested:

... an athlete's persona, familiarity, or likeability may influence fit with various events. Lindsay Davenport and Serena Williams often vied for the number one ranking in women's tennis making them both "experts" as endorsers of their sporting events; however, they exhibited very different personas both on and off the court that could influence consumers' perceived fit at different events (e.g., Wimbledon versus the United States Open). (p. 308)

Indeed, the Meaning Transfer Model (McCracken, 1989) suggests that the "cultural meanings" endorsers possess (e.g., lifestyle, status) can be transferred to the endorsed product, but meaning of the endorsement can take place more effectively when there are greater levels of congruency, or fit. Thus, we propose the last hypothesis:

Hypothesis 4: After controlling for trustworthiness, attractiveness, and expertise, measures of fit will positively influence purchase intentions for the endorsed product.

Methods

Participants

Data were collected from undergraduate sport management students (N = 183) at a university in the Northeast United States. Sixty two percent of the participants identified as male while 82% identified as White, 6% as African American, 4% Hispanic, and 3% Asian. Ninety eight percent of the participants were between the ages of 18-22. Two percent of the participants identified as lesbian/gay/bisexual. Participation in the study was voluntary and participants did not receive any compensation for their participation.

Design and Materials

A 2 (Male/Female) x 3 (Married/Gay/Arrested) between-subjects experimental design was used to examine perceptions of endorser source credibility measures, fit, and effectiveness. Across all six conditions, an athlete profile was created to resemble the form and content of an actual athlete profile found on the USA Bobsled and Skeleton Federation team website. The sport of Skeleton was chosen due to the relative anonymity of the athletes, which made manipulation of the profiles believable to the study participants. The profile included two pictures, statistics from recent races, and a section with biographical information. We controlled for athlete expertise by using identical statistics and race information in all of the profiles. We also controlled for attractiveness by using the same two pictures in all six of the created profiles. The pictures included a headshot of a US Skeleton athlete in a full, face covering, helmet as well as a picture of a US Skeleton athlete during a competition sliding down the icy course. The pictures were such that gender and appearance were undetectable, which allowed for control of the attractiveness variable. However, gender was evident by the first name of the athletes as well as the gendered pronouns (e.g., she, he) that were used in the profile. The only difference in the athlete profiles concerned the biographical information.

We used six treatment conditions that manipulated the athlete's biographical information and gender. All six treatment conditions included the same first paragraph under the biographical section, which read: Four-time National Champion Jen Davenport began the sport of skeleton in 2002, quickly climbing national rank by winning both the National and Junior National Championships in her first season. She dominated the World Cup tour during the 2006 season, winning five gold medals and claiming the overall World Cup crown. Davenport made the Winter Olympic Team in 2006 and slid to a sixth place finish in Torino. She is a favorite to medal at the 2010 Winter Olympic Games in Vancouver.

The second paragraph of the biographical section included personal information, which was manipulated to create the three treatments. The first three sentences of this paragraph were exactly the same in all treatment conditions and read: Davenport was born in Arizona and moved to Colorado with her family as a teenager. She has one brother and one sister and her hobbies include playing beach volleyball, hiking, and reading. Davenport graduated in 2006from Denver University with a degree in communication.

The final two sentences in the personal information paragraph were altered with the treatments. The first treatment, in which the athlete was portrayed as married and not involved in undesirable behavior, read: She resides in Breckenridge with her husband, Tom, and their two dogs, Ice and Slide. Davenport is a spokesperson for Amnesty International.

In the second treatment, the athlete was portrayed as openly gay. The paragraph read:

She resides in Breckenridge with her partner, Tina, and their two dogs, Ice and Slide. Davenport is a spokesperson for the Gay and Lesbian Alliance Against Defamation (GLAAD).

In the third treatment, Davenport was said to have been recently arrested for drunk driving. It read: Her career was derailed briefly in 2007 after an arrest for driving under the influence and reckless misconduct. Davenport has since become a spokesperson for RADD.

The name Jen Davenport was used in the three female conditions while the name Jon Davenport and male pronouns were used in the male conditions.

Participants were randomly assigned to one of the six treatment conditions: male/married (n = 33), male/gay (n = 32), male/arrested (n = 29), female/married (n = 28), female/lesbian (n = 30), and female/arrested (n = 31) and handed a packet that contained a coversheet, the athlete profile, and a questionnaire. The coversheet, which was also read out loud to all of the participants, included directions for the participants to read and study the attached athlete profile and complete the questionnaire, which sought their opinions about the athlete's recent endorsement of the SoBe beverage No Fear. The coversheet also introduced the athlete as a member of the US Skeleton Team and clearly stated the athlete's relationship with SoBe No Fear energy drink including a picture of the beverage.

Measures

We used Ohanian's (1990) scale to assess the perceived trustworthiness, attractiveness, and expertise of the athlete. Reliability coefficients for the trustworthiness ([alpha] = .83), attractiveness ([alpha] = .78), and expertise ([alpha] = .86) measures were high. The fit of the athlete with the product was measured using Till and Busler's five-item scale (2000). Items were measured on a 7-point Likert-type scale ranging from 1 (strongly disagree) to 7 (strongly agree), and the reliability estimate was high ([alpha] = .94). Till and Busler's (2000) three-item semantic differential scale was used to measure intentions to purchase the endorsed product. The reliability estimate for the measure was high ([alpha] = .95). To serve as a control variable relative to participants' interest in the sport, we adopted Parker and Fink's (2008) involvement scale. The reliability estimate for this measure was high ([alpha] = .84). Finally, a single item 7-point scale was used to measure participants' involvement with SoBe No Fear energy drink. The stem read, "How often do you drink SoBe No Fear energy drink?" with endpoints ranging from "never" to "always."

Results

The means and standard deviations of variables of interest can be found in Table 1. A one-way, between groups analysis of covariance (ANCOVA) was used to examine hypotheses 1 and 2, which proposed an athlete endorser who had participated in undesirable acts and an openly gay athlete endorser would both be rated less favorably on measures of endorser trustworthiness, than a straight, non-criminal athlete endorser with identical credentials.

Treatment group served as the independent variable, trustworthiness served as the dependent variable, and participant gender, race, sexual orientation, and involvement served as covariates. The means and standard deviations for variables of interest by treatment group can be found in Table 2. Preliminary assumption testing was conducted and no violations were found. The overall model was not significant F(5, 167) = 1.38; p = .23; [[eta].sup.2] = .04. Additionally, none of the covariates were significant. Thus, hypotheses 1 and 2 were not supported.

Hypothesis 3, which stated perceptions of the athlete's trustworthiness, attractiveness, and expertise would positively affect measures of fit; and, athlete expertise and trustworthiness would contribute the most to perceptions of fit, were examined using multiple regression analysis. The results of the regression are shown in Table 3.

The overall model was significant F(3, 171) = 28.52, p < .001 and explained 33.4% of variance in perceptions of fit. Only attractiveness ([beta] = .33; p < .001) and trustworthiness ([beta] =. 24; p < .05) were significant.

Hypothesis 4, which predicted that, after controlling for trustworthiness, attractiveness, and expertise, measures of fit would positively impact purchase intentions for the endorsed product, was examined using hierarchical regression analysis. Step one included the controls of attractiveness, expertise, and trustworthiness; this step was significant [F(3, 171) = 6.07, p < .001) and accounted for 8.6% of the total variance in purchase intentions. However, only trustworthiness was significant ([beta] = .231, p < .05). Fit was entered into step two, which was significant [F (4, 170) = 13.30, p <.001) and ([beta] = .43; p <.001) and accounted for an additional 13.9% of variance in purchase intentions. Results can be seen in Table 4. Thus, endorser-product fit explained differences in purchase intentions beyond the source credibility measures. Thus, Hypothesis 4 was supported.

Discussion

Hypothesis 1 predicted that undesirable behavior by an athlete endorser would negatively impact participants' perceptions of endorser trustworthiness. However, our results indicated that an endorser's prior arrest had no negative impact on participants' perceptions of endorser characteristics or endorser fit. This could partly be a function of our study design as the treatment indicated the athlete had been arrested for drunk driving and reckless misconduct; however, it was also further mentioned the athlete was now a spokesperson for RADD--Recording Artists, Actors, and Athletes Against Drunk Driving. Therefore, the treatment informed the participants that the athlete was making amends for his/her undesirable behavior.

It is not uncommon for athletes who fall out of favor with fans and consumers, due to public missteps, to regain the confidence of the consumers after they "make amends" for their actions. For instance, Kobe Bryant was accused of rape and publicly admitted to infidelity. While his stock took a hit for a period of time, since 2008 his jersey has been the number one seller in the United States, Europe, and China (Associated Press, 2010). Likewise, Michael Vick, who was convicted of dog fighting and spent two years in prison before returning to the NFL, was tops in jersey sales for the weeks near the end of the 2010 NFL season (Badenhausen, 2010).

Our findings could also be due to the type of undesirable behavior in which the athlete participated. Research has shown that individuals react differently to various types of scandalous events. Hughes and Shank (2005) found participants defined an athletic scandal as one that was unethical or illegal, ongoing, and which affected the integrity of the sport. The most common scandal cited by the participants was the BALCO steroid scandal, which covered all three parts of their definition. Certainly the example of driving while intoxicated and reckless misconduct used in the current study is an illegal act, but perhaps it was not egregious enough for participants to harbor ill will toward the athlete, particularly given their current affiliation with RADD. Future research should reexamine endorser scandals and the impact various types of scandals have on consumer perceptions as well as the process involved in consumer forgiveness.

Hypothesis 2 predicted that openly gay/lesbian athlete endorsers would be perceived as less attractive, expert, and trustworthy than straight athlete endorsers. However, our results indicated that identifying as a gay athlete did not negatively impact participants' perceptions of the athlete's endorser characteristics. These results were unexpected and, according to the research by Penn, Schoen, & Berland Associates (Wertheim, 2005) appear to be contrary to public perception. Perhaps, given our sample of college students (discussed in greater detail below under limitations), our results are an indication of the changing views toward gays and lesbians in our society expressed by younger generations. Or, perhaps our participants exhibited some sort of response bias--that is, they did not want to appear to be homophobic in their responses. However, given the anonymous nature of the survey, this is unlikely.

These results could also be partially explained through the idea of pluralistic ignorance (Miller & McFarland, 1987), which refers to the false belief that one's personal beliefs and attitudes are vastly different than what others believe and feel. A study by Bowen and Bourgeois (2001) sampled college students and asked them to report their personal attitudes toward lesbian, gay, and bisexual (LGB) students as well as their friends' attitudes and typical students' attitudes. The results revealed the students' personal attitudes were significantly more positive toward LGB individuals than how they perceived their friends' attitudes and typical students' attitudes toward LGB individuals. That is, the students believed the world (i.e., their friends and other typical students) was more homophobic than they were. What is interesting about pluralistic ignorance is that these beliefs about others can alter behaviors and actions as individuals look to conform to the way they believe others see the world--rather than acting on their own beliefs (Bowen & Bourgeois, 2001).

If applied to our study, we might surmise that the belief that other people are not accepting or that society and the sport world is not ready for an openly gay athlete endorser may be born out of pluralistic ignorance. In other words, there is a belief that gay athletes and gay athlete endorsers would be hard pressed to exist in the current sport culture--and yet 86% of people believe it is okay for an openly gay individual to play sports and only 18% would be dissuaded by a gay athlete endorser of footwear (Wertheim, 2005). Thus, there seems to be a disconnect. Unfortunately, these "widespread" beliefs may lead athletes to remain closeted out of fear for their careers and marketers may avoid signing gay athletes. This pluralistic ignorance may be enough to limit gay athletes' comfort in being "out," even though, potentially, many consumers in society may personally be very accepting of a gay athlete endorser.

Additionally, the use of an unknown athlete in a somewhat unknown sport may have contributed to these results. It is one thing to be told that an unknown athlete is gay/lesbian, it may be quite different to learn the same information about a well-known athlete. If an individual identifies highly with a celebrity athlete, and the athlete comes out, the image the fan held of the athlete may be incongruent with this new information. Indeed, some research suggests using a non-celebrity endorser is more effective as you can create a public persona to fit your brand and target market rather than being at the mercy of the already created public personality of the well-known celebrity (Tom, Clark, Elmer, Grech, Masetti, & Sandhar, 1992). Further, the unknown endorser is not linked to other brands or products in the way a celebrity endorser may be through multiple product endorsement deals. In the case of this study, our athlete was an unknown persona, whose primary characteristics included being very good at riding a sled, head first, down an icy tube, and was linked with only one product, an energy drink named No Fear. The strong congruency between the product and the endorser is what seemed to matter most--not the piece of their created persona which labeled them as gay. Had the endorser been a well-known athlete with a previously created persona, our study results may have been different.

Though the results of our study cannot be generalized to all consumers, they are tentatively promising for gay/lesbian athletes. Being labeled as gay/lesbian did not significantly impact this (younger) sample's impressions of attractiveness, expertise, or trustworthiness, three key characteristics relative to endorser effectiveness. Although further studies are needed, this suggests that, all other things being equal, gay/lesbian athletes can be effective endorsers and need not keep their identity hidden, especially when dealing with younger consumers.

Hypotheses 3 and 3a predicted that expertise, trustworthiness, and attractiveness would significantly impact perceptions of fit, with expertise and trustworthiness being the more important predictors. However, the results indicated attractiveness and trustworthiness were the significant predictors of endorser-product fit. While we anticipated the experimental conditions to significantly impact perceptions of all three variables, this was not the case. This suggests expertise may be the most "controlled" variable; that is, participants were provided overt information regarding the athlete's expertise, they were not provided explicit information relative to the athlete's attractiveness or trustworthiness. Still, given the tenants of the match-up hypothesis, one would anticipate that being an expert in one's sport would have the greatest impact on endorser-product fit. However, perhaps it is a "given" that athlete endorsers be experts in their sport. That is, an athlete would not be chosen as an endorser unless he/she was somewhat of an expert. Thus, as perceptions of attractiveness and trustworthiness of the athlete increase, they serve to further enhance perceptions of fit. Trustworthiness certainly makes intuitive sense for increased perceptions of fit as consumers would want to believe the athlete has actually used the product, and also trust what the athlete says about the product. Still, according to the match-up hypothesis, attractiveness should have little to do with a sport product, unless it is one designed to make an individual more physically appealing. However, past research has indicated that people, in general, have more positive attitudes about attractive persons compared to those who are less attractive (Eagly, Ashmore, Makhijani, & Longo, 1991; Jawahar & Mattsson, 2005). Further, Cunningham et al. (2008) contended, "... athlete attractiveness has become increasingly utilized in promoting and selling many women's sports ..." thus consumers come to expect athlete endorsers to be attractive as well (p. 373). While they referred to female athletes specifically, perhaps this is becoming true for male athletes as well. For example, Real Madrid soccer star Ronaldo has recently posed shirtless for Time Force watches as well as Armani (Parace, 2010), and Tom Brady's sex appeal is often mentioned as one of his greatest qualities as an endorser (Hill, 2008; Sandomir, 2007).

Hypothesis 4 predicted endorser-product fit would determine purchase intentions beyond the source credibility measures and the results supported this hypothesis. In fact, fit predicted nearly 14% of the variance in purchase intentions. The product used in this experiment was No Fear energy drink, and the company that produces it is known as the "original action sports company" ("About No Fear," para. 1). The company philosophy, in part, states, "Since 1990, No Fear has been there to help develop, define, and promote forms of numerous individual sports that celebrate both the human and competitive spirit" ("About No Fear," para. 2). Obviously, the skeleton is an individual sport fraught with danger and one would need to have "no fear" to participate in the sport, thus, the fit for the athlete and the product is excellent. Similar to other studies (e.g., Cunnigham et al., 2008; Fink et al., 2004), our results indicate that fit goes beyond the typically studied source credibility measures of expertise, attractiveness, and trustworthiness to impact endorser effectiveness. Thus, marketers must be very deliberate in choosing athlete endorsers for particular products. Interestingly, the treatment had no effect on perceptions of endorser-product fit, thus, the congruency between the skeleton athlete and No Fear energy drink seems to have transcended the athlete's legal transgression and his/her being labeled as gay/lesbian. We would suggest that these characteristics might even enhance endorser fit; however, our results indicated all treatment conditions were similar in their effect on perceived fit.

Conclusions, Limitations, and Recommendations

In conclusion, this study found ratings of endorser expertise, attractiveness, and trustworthiness were not statistically different from those of a straight athlete with identical credentials. Thus, an openly gay athlete endorser was viewed similarly effective as an endorser. Likewise, an athlete previously arrested for drunk driving and reckless misconduct who was now serving as a spokesperson for RADD was also viewed as an equally effective endorser compared to an athlete with no prior criminal background. Additionally, we found attractiveness and trustworthiness to be key characteristics in endorser effectiveness while the congruent fit between endorser and product was instrumental to purchase intentions.

As with any study, there are also limitations to this research that should be considered. First, we used a fictional athlete in a somewhat obscure sport. While this allowed us to control for some extraneous variables, it is also possible that those variables could be impactful in other situations. For instance, the use of a well-known athlete with an established public persona who plays in the National Football League might produce different results. Second, our participant sample was comprised of students attending a Northeastern university. We feel the student sample is justified as this age demographic is one that is highly coveted by marketers due to their loyalty and persuasiveness (Broughton, 2010; Knight, 2007). Further, the energy drink we used in the experiment targets this age range ("About No Fear," para. 1). However, we realize it is possible if the study were replicated with participants of different ages as well as different regional locations, the results might vary. In fact, future research should replicate this study in an effort to examine various generational and regional effects.

Finally, our results lead to additional questions which should be examined in future research efforts. For instance, it appears that views of gay athletes may be changing. Is the "locker room culture" in athletics a greater barrier to athletes coming out than the perceptions of the general public? How does athlete gender expression impact perceptions of endorser characteristics and endorser effectiveness? Additionally, the concept of "fan forgiveness" should be further explored including its antecedents and consequences.

References

About No Fear. (2010). Retrieved from http://www.nofear.com

Agrawal, J., & Kamakura, W. A. (1995). The economic worth of celebrity endorsers: An event study analysis. Journal of Marketing, 59, 56-62.

Amos, C., Holmes, G., & Strutton, D. (2008). Exploring the relationship between celebrity endorser effects and advertising effectiveness. International Journal of Advertising, 27, 209.

Associated Press. (2010). Kobe still #1 in NBA jersey sales, LeBron 2nd. Retrieved from http://nbcsports.msnbc.com/id/34961309/ns/sports-nba/

Badenhausen, K. (2010). Michael Vick jerseys flying off shelves. Forbes. Retrieved from http://www.forbes.com/sites/kurtbadenhausen/ 2010/12/20/michael-vick-jerseys-flying-off-shelves/

Bowen, A. M., & Bourgeois, M. (2001). Attitudes toward lesbian, gay, and bisexual college students: The contribution of pluralistic ignorance, dynamic social impact, and contact theories. Journal of American College Health, 50(2), 91-96.

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.

Broughton, D. (2010, February 15). Sweet spots in sponsor loyalty. Retrieved from http://www.sportsbusinessdaily.com/Journal/Issues/2010/02/20100215/ SBJ-In-Depth/Sweet-Spots-In-Sponsor-Loyalty.aspx

Charbonneau, J., & Garland, R. (2006). The use of celebrity athletes as endorsers: Views of the New Zealand general public. International Journal of Sports Marketing and Sponsorship, 7(4), 326-333.

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, 58, 371-378.

Cunningham, G. B., Sartore, M. L., & McCullough, B. P. (2010). The influence of applicant sexual orientation and rater sex on ascribed attributions and hiring recommendations of personal trainers. Journal of Sport Management, 24, 400-415.

Dholakia, R., & Sternthal, B. (1977). Highly credible source: Persuasive facilitator or persuasive liabilities? Journal of Consumer Research, 3(4), 223-232.

Eagly, A. H., Ashmore, R. D., Makhijani, M. G., & Longo, L. C. (1991). What is beautiful is good, but ... A meta-analytic review of research on the physical attractive stereotype. Psychological Bulletin, 110(1), 109-128.

Erdogan, B. Z. (1999). Celebrity endorsement: A literature review. Journal of Marketing Management, 15, 291-314.

Fink, J. S., Cunningham, G. B., & Kensicki, L. J. (2004). Using athletes as endorsers to sell women's sport: Attractiveness versus expertise. Journal of Sport management, 18, 350-367.

Fink, J. S., Parker, H. M., Cunningham, G. B., & Cuneen, J. (in press). Female athlete endorsers: Determinants of effectiveness. Sport Management Review.

Fiske, S. T. (1980). Attention and weight in person perception: The impact of negative and extreme behavior. Journal of Personality and Social Psychology, 38(6), 889-906.

Freedman, J. (2008). Ranking the 50 highest-earning athletes in the U.S. Sports Illustrated. Retrieved from http://sportsillustrated.cnn.com/ specials/fortunate50-2011/index.html

Freedman, J. (2010). The 50 highest-earning American athletes. Sports Illustrated. Retrieved from http://sportsillustrated.cnn.com/specials/ fortunate50-2010/index.html

Hill, J. (2008). Ladies love Teflon Tom--No matter what. Retrieved from http://sports.espn.go.com/espn/page2/story?page=hill/080129

Herek, G. M. (2009). Sexual stigma and sexual prejudice in the United States: A conceptual framework. In D. A. Hope (Ed.), Contemporary perspectives on lesbian, gay, and bisexual identities (pp. 65-111). New York: Springer.

Hovland, C. J., & Weiss, W. (1951). The influence of source credibility on communication effectiveness. Public Opinion Quarterly, 15, 635-650.

Hughes, S., & Shank, M. (2005). Defining scandal in sports: Media and corporate sponsor perspectives. Sport Marketing Quarterly, 14, 207-216.

Jawahar, I. M., & Mattsson, J. (2005). Sexism and beautyism effects in selection as a function of self-monitoring level of decision maker. Journal of Applied Psychology, 90(3), 563-573.

Kahle, L. R., & Homer, P. (1985). Physical attractiveness of the celebrity endorser: A social adaptation perspective. Journal of Consumer Research, 11, 954-961.

Kamins, M. A. (1990). An investigation into the match-up hypothesis in celebrity advertising: When beauty may only be skin deep. Journal of Advertising, 19(1), 4-13.

Klein, J. G. (1996). Negativity in impressions of presidential candidates revisited: The 1992 election. Personality and Social Psychology Bulletin, 22, 289-296.

Knight, C. (2007). Study shows targeting 18-34 year olds can pay off. Retrieved from http://www.bizreport.com/2007/03/ study_target_youngwithmoney_demographic.html

Knight, J. L., & Giuliano, T. A. (2003). Blood, sweat, and jeers: The impact of the media's heterosexist portrayals on perceptions of male and female athletes. Journal of Sport Behavior, 26(3), 272-285.

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. Sport Marketing Quarterly, 18, 15-37.

Louie, T. A., Kulik, R. L., & Jacobson, R. (2001). When bad things happen to the endorsers of good products. Marketing Letters, 12(1), 13-23.

Lynch, J., & Schuler, D. (1994). The matchup effect of spokesperson and product congruency: A schema theory interpretation. Psychology and Marketing, 11, 417-445.

Majendie, P. (2008). Olympics--Only 10 games athletes say they are gay--survey. Reuters. Retrieved from http://www.smh.com.au/news/off-thefield/ only-gay-in-the-village/2008/08/18/1218911546077.html

McCracken, G. (1989). Who is the celebrity endorser? Cultural foundations of the endorsement process. Journal of Consumer Research, 16(3), 310-321.

Miller, D. T., & McFarland, C. (1987). Pluralistic ignorance: When similarity is interpreted as dissimilarity. Journal of Personality and Social Psychology, 53(2), 298-305.

Ohanian, R. (1990). Construction and validation of a scale to measure celebrity endorsers' perceived expertise, trustworthiness, and attractiveness. Journal of Advertising, 19(3), 39-52.

Ohanian, R. (1991). The impact of celebrity spokespersons' perceived image on intention to purchase. Journal of Advertising Research, 31(1), 46-54.

Parace, J. (2010). Real Madrid star Critiano Ronaldo is fashionable, stylish, and untouchable. Retrieved from http://www.examiner.com/athletestyle-in-national/ real-madrid-star-cristiano-ronaldo-is-fashionable-stylish-and-untouchable

Parker, H. M., & Fink, J. S. (2008). The effect of sport commentator framing on viewer attitudes. Sex Roles, 58, 116-126.

Ragins, B. R., & Cornwell, J. M. (2001). Pink triangles: Antecedents and consequences of perceived workplace discrimination against gay and lesbian employees. Journal of Applied Psychology, 86, 1244-1261.

Ragins, B. R., Singh, R., & Cornwell, J. M. (2007). Making the invisible visible: Fear and disclosure of sexual orientation at work. Journal of Applied Psychology, 4, 1103-1118.

Sandomir, R. (2007). Brady and Manning meet in battle of the brands. Retrieved from http://www.nytimes.com/2007/11/04/sports/football/ 04qbs.html?pagewanted=print

Sartore, M. L., & Cunningham, G. B. (2009). Sexual prejudice, participatory decisions, and panoptic control: Implications for sexual minorities in sport. Sex Roles, 60, 100-113.

Simonin, B. L., & Ruth, J. A. (1998). Is a company known by the company it keeps? Assessing the spillover effects of brand alliances on consumer brand attitudes. Journal of Marketing Research, 35, 30-42.

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, 29, 1-14.

Till, B. D., & Shimp, T. A. (1998). Endorsers in advertising: The case of negative celebrity information. Journal of Advertising, 27(1), 67-82.

Tepper School of Business, Carnegie Mellon University. (2010). Economic value of celebrity endorsement: Tiger Woods' impact on sales of Nike golf ball. Retrieved from http://www.andrew.cmu.edu/user/derdenge/ TWExecutiveSummary.pdf

Tom, G., Clark, R., Elmer, L., Grech, E., Masetti, J., & Sandhar, H. (1992). The use of created versus celebrity spokesperson in advertisements. The Journal of Consumer Marketing, 9(4), 45-51.

Veltri, F. R., Kuzman, A. T., Stotlar, D. K., Viswanathan, R., & Miller, J. J. (2003). Athlete-endorsers: Do they affect young consumer purchasing decisions? International Journal of Sport Management, 4, 145-160.

Votolato, N. L., & Unnava, H. R. (2006). Spillover of negative information on brand alliances. Journal of Consumer Psychology, 16(2), 196-202.

Wertheim, J. L. (2005). Gays in sports: A poll. Sports Illustrated, 102(16).

Heidi M. Parker, PhD, is an assistant professor of sport management in the School of Business at the University of Southern Maine. Her research centers on sport consumer behavior, focusing on factors that influence fan attitudes and perceptions.

Janet S. Fink, PhD, is an associate professor in the Neag School of Education at the University of Connecticut. Her research interests include diversity issues in sport, sport media and the female athlete, and sport fan behavior. Table 1. Means, Standard Deviations, and Correlations of Variables of Interest Variable N M SD Attractive Expert Trust Fit Attractiveness 177 4.19 .84 1.00 -- -- -- Expertise 181 5.74 1.02 .54 1.00 -- -- Trustworthiness 179 4.84 .94 .69 .66 1.00 -- Fit 182 5.27 1.26 .52 .45 .52 1.00 Table 2. Mean Scores of Variables of Interest by Treatment Group Treatment Trustworthy M SD Male/Married 4.94 1.11 Male/Gay 5.03 .85 Male/Arrested 4.61 .71 Female/Married 4.99 1.03 Female/Lesbian 4.96 .95 Female/Arrested 4.52 .86 Table 3. Results of Regression on Perceptions of Fit Variable B SE [beta] Adjusted [R.sup.2] Attractiveness .446 .131 .303 *** Expertise .171 .106 .135 Trustworthiness .311 .132 .223 * .334 * p<.05, ** p<.01, *** p<.001 Table 4. Results of Hierarchical Regression on Purchase Intentions Variable B SE [beta] Adjusted [R.sup.2] [R.sup.2] Change Step 1 .08 ** Attractiveness .316 .186 .169 Expertise -.156 .150 -.101 Trustworthiness .393 .187 .231 * Step 2 Fit .555 .100 .*** .216 *** 13.9 *** * p<.05, ** p<.01, *** p<.001