This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.
"Perhaps most important from a psychological perspective, companies understand that some consumers idealize certain athletes as heroes, and they capitalize on this hero status through endorsement messages."
-Peetz, Parks, and Spencer 2004
Advertising is an important factor with regards to the economic activity of a country as it not only helps to stimulate consumption but also generate and present models for lifestyles and value orientation (Polly and Mittal, 1993). Celebrity ads are one effective way of doing so. Although the effectiveness is debatable, the trend seems to be largely in their favor nonetheless. Almost 25% of all American ads have celebrities in them (Stephens and Rice, 1998) and this percentage is more than double, up to 70% in Japan (Kilburn, 1998).
It is no surprise to see celebrity endorsements being used by marketers to voice their messages in the market to promote a huge spectrum of products and services. A large pool of money is devoted by firms to have athletes and other celebrities endorse their product, be it sports or non sports items. According to Boyd and Shank 2004 reports, 20% of all television commercials contain celebrities in one form or another. And, the situation should not be much different in print or radio ads. In a short span, the amount of moolah that goes to celebrities in advertising has skyrocketed. In 1986, the expenses on celebrities as payments for endorsement, stood at $100 million. This figure had grown more than tenfold in ten years. (Peetz, Parks and Spencer 2004)
According to McCracken (1989), celebrity endorser is a person who enjoys recognition by public and who can utilize this recognition and transfer it to the product by appearing with it. Similarly in words of Friedman and Friedman (1979), celebrity endorser is a person who is known to public not because of his association with the product, but because of his achievements in other field. In fact it is the other way round and a product is expected to be recognized by being associated with the celebrity. So, a celebrity endorser is brought on to the bandwagon to transfer his fame to the product, and give it a meaning that people can relate to him.
It is just another marketing strategy whose effectiveness has always been put under scanner. The primary reason is difficulty in subjective evaluation of results that can be attributed to the celebrity endorser. Also, it depends a lot on how the endorser behaves in his personal and professional life. This is particularly a serious concern while using athletes, as one bad performance may mar the entire image of the sportsperson, and along with, image of the product. Also, there are many factors related to celebrity that come into play like credibility, physical attractiveness, influence, familiarity and likability.
However there still remains much to be learned about how consumers perceive athlete endorsers and the variables that influence endorser effectiveness. More specifically, there is very little knowledge about how gender - of celebrity endorser and viewer/consumer- affects consumer's attitude towards the ad and celebrity.
Using the social learning theory as the foundation, we know that individuals learn behaviors and attitudes from previous experiences and through observation of other individuals and media in the social context (Bandura 1977, 1986). Purchase intentions and consumer behaviors are also learnt in a similar way. According to Escalas, Bettman, "People engage in consumption behavior in part to construct their self-concepts and to create their personal identities". Consumers use brands for this purpose and they appropriate the brand communication derived from celebrity endorsement.
Brand associations are important as consumers use brands and their symbolic meanings to construct self-images and express themselves. The 'symbolic meanings' of the celebrity are transferred to the brand then to the consumers when they favour brands which are congruent to their self-concept (McCraken 1989) Celebrities with aspirational images enhances this self-brand connection for the consumer. (Escalas, Bettman 2009). Consumers generally want to create favourable self-image and are highly influenced by the need to enhance self-esteem. (Crocker, Park 2004) When a consumer's self-concept gets linked to the brand, it positively affects their purchase intent and brand-loyalty in the long run. These effects are more pronounced when there is an inherent fit between the celebrity and the product. (McCraken 1989). This fit has been named as 'match-up hypothesis' (Kahle and Homer, 1985 ; Lynch and Schuler, 1994 ; Kamins 1990; Solomon et al 1990).
As a result of the pervasiveness of celebrity endorsement in advertising, much has been studied and written about this phenomenon over the past several years (for recent examples, see Bailey 2007; Biswas, Biswas, and Das 2006; Thomson 2006). The studies have explored a number of issues related to such factors as the relative effectiveness of celebrity endorsers in advertising, the factors that impact consumer response to celebrity endorsers, and, more recently, the consequences of negative information or publicity surrounding a celebrity endorser (for example, Bailey 2007; Louie and Obermiller 2002; Silvera and Austad 2004; Till and Shimp 1998).
Celebrity endorsement has been found to be effective due to the following reasons:
Capturing Attention: It increases the attention paid to an advertisement by the consumer (Buttle, Raymond, Danziger 2000).
Persuasive effect: Celebrities have significant persuasive effect when the consumer seeks social acceptance (DeBono, Harnish 1988) or when the product is attractiveness-related (Kamins 1990).
Increase credibility: Celebrities which are perceived to have expertise in their domain can make an ad more reliable for the consumer; for example: a good-looking celebrity endorsing cosmetics (Baker, Churchill: 1983) an athlete associated with sport shoes (Ratneshwat, Chiaken 1991). Kahle and Homer (1985) found that higher celebrity-product brand fit, higher is the perceived reliability of the ad.
Act as a peripheral cue: In dual-process models applied to psychology, where persuasion can lead to change in conscious processes or attitudes; celebrities have been found to have an peripheral cue in case of low-involvement product categories (Petty, Caciopo, Schuman 1983). For high-involvement product categories, better the brand fit between the celebrity and the product, higher is the consumer reliability on the ad
The impact of celebrities on adolescents' brand loyalty was studied by Lockwood, Kunda (1997). They found that "individuals of outstanding achievements can serve as role models to others", motivating young adults to adopt certain self images and lifestyles. Martin and Bush (2000) reported that vicarious (entertainers and sportspersons) can influence adolescents to change brands to a greater extent than keeping them loyal to a brand. The effect the gender of the celebrity has on adolescents depends on the culture.
Kahle and Homer (1985) found that women were more likely to buy a product endorsed by a attractive celebrity than were men. Daneshvary and Schwer (2000) also found that Individuals with lower level of education are more impressionable to endorsements than individuals with higher level of education. Till and Busler (1998) concluded that an endorser's expertise is more important than physical attractiveness in affecting consumer's attitude towards a brand.
According to Ohanian, her study of 'The impact of celebrity spokespersons' perceived image on consumers' intention to purchase' (1991) showed that there was no significant impact of celebrity spokesperson's gender on the perceived attractiveness, trustworthiness, or expertise of the celebrity, nor on the purchase intent of the endorsed product. She found that only a celebrity's perceived expertise had effect on purchase intention. In her study, she had used male celebrities John McEnroe (tennis player) and Tom Selleck (actor) alongside female celebrities: Linda Evans (Actor) and Madonna (then, relatively unknown singer). There was no matching of the domains of the celebrities, as in this study where we are using M S Dhoni (Cricketer) and Bipasha Basu (Actress).
Peetz, Parks, and Spencer (2004) tried to find, in their study of heroes as sport product endorsers, the effect of gender of the sportsperson on the recognition of the celebrity, consumer's perceived expertise of the endorsed product and the influence of the celebrity on purchase intent. They used male endorsers: Lance Armstrong and Michael Jordon and female endorsers: Mia Hamm and Jackie Joyner-Kersee. The results showed that male endorsers were identified more often as compared to the female counterparts; male endorsers had more influence on the purchase intentions and that the lesser known male celebrity was more favored by the male respondents.
Boyd and Shank in their study Athletes as product endorsers: The effect of gender and product relatedness. Sport Marketing Quarterly 13(2): 82-93, 2004 tried to find out the impact on gender matching between sport celebrity endorsers and consumers on consumer's perception of the celebrity's trustworthiness, attractiveness and expertise. The result of the research indicated that there was no significant effect of gender matching on the perceived expertise or attractiveness of the celebrity. However, on the item 'trustworthiness', it was concluded gender matching was strongly influencing it. Female endorsers were more favorable to women respondents than male endorsers and similarly, men were more favourable towards male endorsers. For their research they had used a Causian female swimmer Amy Van Dyken on one hand and on the other, Hispanic-American male boxer Oscar de la Hoya. They were of the opinion that there was potential for effect of alternative factors like (sport, race, recognition etc.) on the observed results.
Empirical finding support the fact that celebrities have positive effect on both attitude toward ad and brand (Ohanian, 1990). These results are in favor of celebrity endorsers because they are widely recognized, are perceived to be more credible and produce greater influence on evaluation of brand and its purchase intentions (Cohoi and Rifon, 2007; Atkin and Blok, 1983; Ohanian, 1990; Ohanian, 1991). The credibility components are interconnected with other traits of the celebrity and the image dimension which reflects good qualities e.g. pleasant, wise, educated etc. produce sincere and positive perception of the celebrity in consumer minds (Choi & Rifon, 2007). Generally the sequence of criteria that a firm follows goes like this; credibility is given top priority followed by familiarity and likeability and at last there is gender, and within credibility, expertise is most sought after which is followed by trustworthiness and attractiveness in that order (Knott & James, 2004). The celebrities that were found to be regular in news are trusted more than the celebrities who are not in the constant spotlight (Friedman, Santeramo, & Traina, 1978). Physical attractiveness of the endorser can play a pivotal role in determining the change in attitude of a consumer (Kahle & Homer 1985). In general attractive endorsers are more effective promoters than unattractive endorsers (Till & Busler, 1998). The target audience relates or at least prefers to relate to the attractive celebrities and hence, these celebrities are more likable (Miciak & Shanklin, 1994). Most advertisements use attractive celebrities and the consumers are habituated to see good looking people in ads. For this reason, the physical attractiveness and trustworthiness of a celebrity were not as significantly related to the purchase intentions as expertise is (Ohanian, 1991). Involvement enhances the recall of the product and it also enhances the recall of the brand (Petty & Cacioppo, 1980). There is a strong effect of celebrity attractiveness as well as interaction of likeability, involvement, and gender on the recall of the product.
In Indian context also, there have been many researchers who have debated the issue and relevance of celebrity endorsers. According to Agarwal and Kamakura (1995), approximately 20% of all television commercials in India feature popular celebrities. There are certain products that are more often associated and promoted by celebrities. They did a study of all endorsement contracts over the period 1980-1992, and observed that 60% of their sample endorsement deals involved either soft drinks or athletic shoes. They also observed that whenever a company made an announcement of a deal with a celebrity for any product's endorsement, its stocks showed a bullish trend in the market, thereby signaling a positive response from the market.
Mishra and Beauty (1990) have suggested that celebrity endorsements go a long way in enhancing brand recall. They found that people often assumed that a product would be of higher quality if it was endorsed by a congruent celebrity. Prachi Raturi (2005) in her research concluded that no one beats a celebrity when it comes to improving sales. Signing a celebrity helps the brand jump out of mess, and if there is coherence between the product and celebrity, it is bound to be a rip roaring success.
A study organized by Cyber media research published in Business World had a lot of different and unheard things to say about celebrity endorsements. The study was spread over different cities in India (Delhi, Meerut, Chennai, Coimbatore, Kolkata, Nasik, Mumbai), and 12 focus group interviews, 6 expert instruments and 8 expert ineterviews with ad agencies were conducted. Besides, there were 3375 respondents that helped provide various insights on celebrity endorsements. These are as follows:
More than 80% of the people remembered the celebrity but could not recall the brand.
Different celebs had varying degree of appeal in different geographical regions. For instance, Aishwarya Rai had high recall in South while Shahrukh Khan didn't
It also emphasized that ads without celebrities had as good chances of being successful as ads with them. A case in point was Hutch which had a better brand recall compared to Coke, in spite of Coke's association with many celebrities.
In spite of huge amount of literature on celebrity endorsements, there is no conclusive evidence of how effective they are and how demographics of consumer and celebrity can affect his/her attitude. This study is a step in that direction.
The objective of this research project is to assess the impact of gender - both gender of the consumer and gender of the endorser- on consumer's attitudes and perceptions of celebrity endorsements. The questions that this study aims at answering are:
How does gender of a consumer affect his/her attitude towards the celebrity endorser in the ad, as well as attitude towards the ad itself?
How does gender of a celebrity determine the consumer's response to the celebrity as well as the ad featuring him/her?
What are the interaction effects of gender of celebrity and gender of consumer with regards to celebrity endorsement in the ad?
There has been prior research on gender differences that suggests that, men and women react to persuasive messages differently. In general, women are more easily persuaded than men (Widgery and McGaugh 1993). Recently, a study done by Berney-Riddish and Areni (2006) also made a similar claim that women accept advertising claims more readily.
Wolin (2003) has done an extensive research on gender differences in advertising. Her research aimed at making certain generalizations regarding gender and its relation to advertising. She concluded that the research related to this hypothesis indicated that "under low risk condition, females show equally favorable response to objective and subjective claims" and that males showed no preference for objective claims. She also suggested that the research shows that women tend to be more persuadable than men. Hence, it is expected that, in general, women will respond more favorably to advertising than men. This also includes cases where celebrity endorsers are used. This expectation is contained in Hypothesis H1A.
H1A: Participant gender will impact attitude toward the ad featuring a celebrity endorser. Women will respond more favorably to the ad than men.
Boyd and Shank found in their study that male consumers were more receptive than female consumers to athlete endorsers (2004). Men saw athletes as trustworthy and reliable irrespective of their gender. However, the difference was not statistically significant (Boyd and Shank, 2004).
There has been lot of research that suggest that men respond more to visual cues than women, and they also respond less favorably to complex messages. Celebrities in ads represent visual, peripheral cues to which men are likely to respond more favorably than women.
The gender-based differences in advertising information processing have explored by Putrevu (2001) and Putrevu, Tan, and Lord (2004). Putrevu (2001) sought to explain the origins of these differences in information processing. Among the factors that he identified were social factors and biological factors. Based on his theory, under certain conditions, men are more likely to be driven by overall message themes or schemas and women are more likely to engage in detailed elaboration of message content. The theory also suggests that women are likely to be more comprehensive information processors than men. It seems obvious that message elaboration will lead to less favorable attitudes toward a celebrity endorser, given the peripheral nature of this cue (Petty, Cacioppo, and Schumann 1983). Thus, men and women will differ in response to a celebrity endorser in advertising. The expected difference in response to the celebrity endorser in an ad is contained in H1B.
H1B: Participant gender will impact attitude toward the celebrity endorser. Men will respond more favorably to the celebrity endorser than women.
In their study, Boyd and Shank (2004) found that, overall, male endorsers may be more effective than female endorsers. It was found that male endorsers were rated as being more trustworthy, having more expertise and high on attractiveness. However, there weren't many differences with regards to credibility and trustworthiness. According to recent research findings, men are generally more paid and more frequently used as endorsers (Thomaselli, 2004). Hence, male endorsers might enjoy familiarity with the audience. Ironically, this excess familiarity could have an adverse effect on the credibility of male endorsers. For example, prior research has shown that overexposure in the form of multiple product endorsements could hurt celebrities (Tripp, Jensen, and Carlson 1994). Women, being less often used as celebrity endorsers, may therefore enjoy more credibility. This leads to hypotheses H2A and H2B.
H2A: Gender of the celebrity will impact attitude toward the ad. Ads featuring female celebrities will be evaluated more favorably than ads featuring male celebrities.
H2B: Gender of the celebrity will impact attitude toward the celebrity endorser in the ad. Female celebrity endorsers will be evaluated more favorably than male celebrity endorsers.
Certain interaction effects are bound to occur. It is expected that there will be differences between men and women in their responses to the use of female and male celebrities in product endorsements. Boyd and Shank (2004) found cross-over interactions in relationship to gender matching in celebrity endorsements. On the matter of trustworthiness, women evaluated female endorsers more favorably than male endorsers, and men evaluated male celebrities more favorably than female celebrities. Finally, Boyd and Shank concluded that, overall, male endorsers may be more effective than females (2004).
According to Putrevu (2001), the socialization literature states that children tend to identify first with their own gender, and that they tend to validate their identification by aligning themselves with the attitudes of members of their gender.
There are also many streams of research that stress on similarity and congruence between the target audience member and communicator, and the impact these variables have on the response. Simpson (2000) found that racially congruent actors had a far greater impact on Black consumers. Morimoto and La Ferle (2006) stated similar results for Asian American consumers. Earlier research has also established, essentially, that people have a tendency to respond more favorably to others that they view as similar to themselves (Byrne 1971; Gaertner and Dovidio 1977; Shanteau and Nagy 1979). From this information, the following is hypothesized:
H3A: Female consumers will respond more favorably to ads featuring female celebrities than to ads featuring male celebrities. Male consumers will respond more favorably to ads featuring male sports celebrity endorsers than to ads featuring female sports celebrity endorsers.
H3B: Female consumers will respond more favorably to female sports celebrity endorsers than to male sports celebrity endorsers. Male consumers will respond more favorably to male sports celebrity endorsers than to female sports celebrity endorsers.
Overall, the idea of consumers responding more favorably means that participants in the study will have higher scores with regard to attitude toward the advertisement and attitude toward the endorser on measures of the basis of his/her gender.
Study Design and Participants
A questionnaire has been created to test the hypotheses stated above. The questionnaire had 2 print advertisements, and 10 questions based on each advertisement apart from demographic information of consumer like age and gender. Both print advertisements are for product "Reebok", and one has MS Dhoni as the endorser and another has Bipasha Basu as the endorser. Both ads are set in absolutely similar backgrounds, with similar poses of endorsers and depict similar theme and message. The only difference is the celebrity endorser. There are 5 questions that test that consumer's attitude towards the ad, and 5 that test his attitude towards the celebrity. This survey was floated online among college students and working professionals. It has been taken by 111 people, 84 males and 27 females.
Please refer to Appendix 1 to see the questionnaire.
As specified before, the aim of this study is to observe effect on consumer response to a celebrity endorser and an ad featuring them, based on gender differences. In order to discard other possible reasons for differences, an effort has been made to keep everything about the ad similar except the gender.
Please refer to Appendix 2 to see both the print ads used in the study.
The two independent variables of interest are:
Gender of the participant
Gender of the celebrity endorser.
We are chiefly concerned with studying consumers' attitudes in this study. Therefore, the dependent variables of interests are:
Attitude toward the ad
Attitude toward the celebrity endorser in the ad.
These dependent variables are valuable to advertisers as they give an indication of how effective a particular celebrity endorsed ad would be amid a target consumer segment. And, so they can formulate a winning strategy based on the findings of this study.
Attitude toward the ad was measured on 5-point Likert type scales (5 response categories: favorable/unfavorable, good/bad, positive/negative, unappealing/appealing, not likable/likable)
Attitude toward the celebrity was measured using 5-point Ohanian scale: (5 response categories: not sincere/sincere; not credible/credible; unreliable/reliable; not objective/ objective untrustworthy/trustworthy.
SPSS v.17 was used for data analysis. ANOVA was used to assess the effects of the independent variables on the dependent variables.
We did a reliability test of the results obtained from the survey using SPSS. The Cronbach's alpha has a value of 0.836 which is greater than 0.7. So the results are reliable. Please refer Table 1 in Section 1 in Appendix III.
Independent Variable: Participant Gender
Dependent Variable: Attitude towards the Ad
H1A states that the gender of the participant has an impact on attitude towards the ad. And it is expected that women would respond more favorably to the ad as compared to men. Our findings support this hypothesis. (F value = 21.961, sig = 0.00) The mean score for women on "attitude towards the ad" is 19.72 as compared to men' score for the same, 16.96. Also done is an Independent Sample T-Test. Please refer Table 1 in Section 2 and Table 5 in Section 3 in Appendix III, for detailed test results.
Dependent Variable: Attitude towards the Celebrity Endorser
H1B argues that the participant gender would have an impact on attitude towards the celebrity endorser. The expectation here is that men would respond more favorably to celebrity in the ad as compared to women. However, the results were against our expectations. The hypothesis is not supported. (F value = 3.141, sig = 0.078). Also, the mean score for women on attitude towards celebrity is found out to be 16.38 while for men, the score is 15.21. As can be seen, there is little difference between the means and they also tend in direction of women, which is unlike our prediction. Please refer Table 2 in Section 2 and Table 5 in Section 3 in Appendix III.
Independent Variable: Celebrity's Gender
Dependent Variable: Attitude towards the Ad
H2A contends that gender of celebrity has an impact on attitude towards the ad. It suggests that an ad with female celebrity will garner more favorable response compared to an ad with male celebrity in it. This hypothesis has been supported with our findings. (F value = 6.377, sig = 0.012). The mean score for attitude towards ad with female celeb is 18.36 compared to corresponding mean score for an ad with male celeb which is 17.00. Please refer Table 3 in Section 2 and Table 6 in Section 3 in Appendix III, for detailed test results.
Dependent Variable: Attitude towards the Celebrity Endorser
H2B states that attitude towards a celebrity endorser is affected by the gender of celebrity. It suggests that female celebrity endorsers would be favored more than male celebrity endorsers. The findings prove that attitude towards a celebrity is indeed affected by the celebrity's gender. (F value = 9.913, sig = 0.002). However, the mean scores suggest a different story. The mean score for attitude towards a female celebrity is 14.62 while for male celebrity, it is 16.41. So, we can say this hypothesis is partially supported. Please refer Table 4 in Section 2 and Table 6 in Section 3 in Appendix III, for detailed test results.
Interaction Effects: Independent Variables: Participant and Celebrity Gender
The existence of interaction effects has also been evaluated, for hypotheses H3A and H3B.
Dependent Variable: Attitude towards the Ad
H3A says that women should respond favorably to ads with female celebs, and men should respond favorably to ads with male celebs. From univariate analysis, the results are not significant (F value = 2.478, sig = 0.117). From mean score analysis, there is no significant difference among females to the ad based on celebrities gender (female celeb ad = 19.68; male celeb ad = 19.78). However the men favored the ad with female celebs more than ad with male celebs (female celeb ad = 17.85; male celeb ad = 16.12). As is evident from these findings, there is no support for first part of H3A, and second part of H3A is out rightly rejected. Please refer Table 7 and Table 9 in Section 4 in Appendix III, for detailed test results.
Dependent Variable: Attitude towards the Celebrity Endorser
H3B argues that women would appraise female celebrity endorsers and men would appraise male celebrity endorsers. Here also, the univariate test yielded no significant results (F value = 0.858, sig = 0.355). From mean score analysis, there are some differences in women's response towards celebrity. They however tended towards male celebrities (female celeb = 15.94, male celeb = 16.89). Among men, there is considerable difference between his responses towards celebs. They have given high ratings to male celeb as compared to female celeb (female celeb = 14.11, male celeb = 16.26), as predicted in our hypothesis. So we can say that H3B is partially supported. Please refer Table 8 and Table 10 in Section 4 in Appendix III, for detailed test results.
The purpose of this study was to examine attitude towards the ad and endorser, based on gender of participant and celebrity. The study generated mixed outcomes, but we addressed certain issues that were ignored in previous researches on similar themes. These issues range from use of similar product, similar background, similar themes and same set of people. This was done to avoid any bias and rule out any other deciding factor other than celeb's gender.
Of the six hypotheses, two have been fully supported; two have been partially supported, while two could not be supported with our findings. Both the hypotheses that are supported suggest that gender does play a role in consumer's attitude towards the ad, be it participant's gender or celebrity's gender.
From H1A, it was concluded that women respond more favorably to the ad compared to men. This finding is not a new one, but it adds weight to the prior research that has generated similar result.
H2A has been fully supported while H2B has been partially celebrated. The findings have suggested that consumers respond differently to male and female celebrity endorsers. The ads with female endorsers have been favored more by participants, as was expected. This could be the result of excess use of male celebrities in ads that has probably led to contempt. Also, there is evidence that attitude towards male celebrities is more favorable compared to female celebrities. This is against our hypothesis, but possible reason could be lack of credibility and expertise that people associate with Bipasha Basu compared to MS Dhoni, for a product like sports shoes.
There is no significant difference between males and females in their attitude towards a celebrity. And hence we could not support H1B.
We could not garner enough support for H3A from our findings. However we found that there were hardly any differences in the way women perceive ads based on celebrity gender. Also, men were found to be more responsive towards ad with female celebrity, although not by a huge margin. This could be the result of better visual appeal that an ad with female celeb might have. H3B has been supported partially with both women and men appraising male celebrity. This could be because we tested the celebrity only on Ohanian's trustworthiness scale, and MS Dhoni being a sportsperson was found to be more credible for a sports product.
Limitations and Future Research
In spite of our best efforts to make this study comprehensive and beneficial to discussion on celebrity endorsements and gender impact on them, there are a few shortcomings that can be looked into and worked upon to improvise future research.
First, we have used a sample of 111 people, most of whom are college students and young working professionals. Although for a college project, this sample size should suffice, but for a more generalized analysis, a wider spectrum of sample can be used.
Another limitation is with respect to the celebrities used. We have made efforts to use two very similar print ads. But the celebrities come from different backgrounds. MS Dhoni is a cricketer while Bipasha is an actress. The ad revolved around the theme, "Doing thing in their own way", and both suited the bill perfectly. But the ad being for sports shoes, people might be inclined towards MS Dhoni as he is a sportsperson. This bias could be eliminated. Also, certain people might have a prejudice towards one of the celebrities and this would hamper neutrality if his answers.
Another interesting study related to this research could be to assess whether the results would hold in case of other products. Here, for this product the match up hypothesis was supported, but in case of a different product, would there be any gender impact? This question needs to be answered, by a more varied study. Additionally, there could be study of other dependent variables like attitude towards brand and purchase intentions.
In case of sports celebrities, there is a great risk of negative publicity adversely impacting the brand's image. For instance, Tiger Woods and Marion Jones both have been widely used as endorsers. But recent infamous case of Tiger Woods' and Marion Jones being stripped of her 2000 Olympic medal because of steroid use, are two prominent example of publicity gone bad. It would be interesting to know how consumers perceive the transgressions of these celebrities in a different light, depending on their gender. If women are given a harsher treatment in such misdemeanor, is yet to be seen.
Finally, there are many ads that focus on using older celebrity endorsers to target older audience. The impact on their attitude can also be studied by using an older aged sample. These studies can help provide more insights on the interesting phenomenon of using iconic celebs to sway consumer's attitudes.