Content-Type: text/html

INTRODUCTION


The emphasis on sponsorship is growing as an element of marketing communications (Madrigal 2000; Javalgi et al. 2001), and celebrity endorsement in advertising has been widely used by advertisers and marketers (Dean 1999; McCracken 1989).  However, the scope of research on celebrity and sponsorship has not been wide enough.
Most celebrity endorsement research has focused on how source credibility and attractiveness influence consumers' overall attitudes toward a product or brand (McCracken 1989).  This initial attempt to examine celebrity effects on advertising has been developed into the so-called "match-up" hypothesis, which indicates that a consumer's perception of a celebrity endorsement heavily depends on a perceived match between the image of the celebrity and the product or brand that is being promoted (McDaniel 1999).
The match-up hypothesis of celebrity endorsement has been adapted to the research on sponsorship, resulting in a reasonable amount of literature on the appropriate linkage between the sponsor and the object of sponsorship (Cornwell et al. 2001).  Also, much literature has focused on the effects of event sponsorship in sports.  For instance, the Olympic or Soccer World Cup sponsors were examined for the effects of underwriting a sponsor's brand in events delivered through mass media (Javalgi et al. 2001), and the impact of Olympic sponsors on the corporate image was examined through the 1992 Olympic survey data (Stipp and Schiavone 1996).
Nonetheless, relatively little attention was paid to the impact of sponsorship on individual athletes, of athletes endorsing a sponsoring company's brand, and an endorser's ethnicity.
Questions to pursue include: "Do athletes as celebrity endorsers have a positive impact on consumers' perception of brand image and/or intention to purchase a sponsor's product?" and "Do consumers' perceptions of advertising differ, depending on the ethnic background of the endorser?"  This study uses these questions to examine the effects of sponsorship, celebrity endorsement, and the ethnic background of the endorser in advertising.
The object of sponsorship in this study is an individual athlete.  Sponsoring an individual athlete refers to corporate efforts in providing sports stars with apparel or equipment in expectation of future benefits to be generated by the athlete's use of the brand, namely by enhancing the brand image by continuously exposing it in the mass media.  Michael Jordan's contract with NIKE and Martina Hingis's use of Adidas tennis wear and shoes illustrate such sponsorship activity.
Literature on celebrity, sponsorship and the ethnic aspects of an endorser is first covered.  This leads to the hypotheses of this study.  Then, the method of the experiment conducted is specified, followed by findings, conclusions and discussion.









LITERATURE REVIEW


The function of advertising is to guide consumers' decision-making process in product consumption through language, images, and the beliefs of the consumers (Sandage 1993).  Thus, it is critical for marketers and advertisers to select appropriate target consumers and to decide how to deliver a persuasive message that fits the target consumers.
Because celebrity and sponsorship in advertising aim to persuade target consumers (Cunningham and Taylor, 1995; Gwinner 1997; Keller, 1993; Dean 1999), and because target consumers' ethnic background influences the determination of the endorser's ethnicity (Armstrong 1999), three factors - celebrity, sponsorship and the ethnic background of an endorser - are important strategies to developing an effective communication message in advertising.

Sponsorship/Celebrity Endorsement in Advertising

The rationales of sponsorship and celebrity endorsement are applied together in this study because an athlete endorser in advertising has both the characteristics of celebrity and of sponsorship.  Furthermore, sponsorship and celebrity endorsements have similar theoretical roots, in the sense that "association" (e.g., the association between the sponsor and the object of the sponsor, and the association between the endorser and the advertised brand) is a key determinant of the effectiveness of the message.
Sponsorship.  Sponsorship is a commercial association of a brand with a sports team or sporting event (Andrews 1996) in order to access exploitable commercial potential (Meenaghan 1996).  The major purposes of sponsorship include increasing awareness of the company or its brand, enhancing brand image, and/or increasing corporate sales (Cornwell et al. 2001; Javalgi et al. 1994).
Balance theory (Dean 1999; Crimmins and Horn 1996) and transfer of association (Keller 1993) explain why sponsorship and celebrity endorsements can have a positive impact on consumers' perception of and response to the advertising.
Dean (1999), and Crimmins and Horn (1996) draw upon Heider's (1946; 1958) balance theory to argue that sponsorship enhances the positive beliefs of consumers with respect to a sponsor by balancing the perceived value of the sponsors and the object of sponsorship when consumers change their attitudes to achieve consistency.  As Crimmins and Horn (1996) illustrated, by linking a highly valued sports event (i.e., the Olympics) with a less-valued sponsor (i.e., less-valued, mid-sized oil company), consumers can alter their attitudes toward the less-valued sponsor through the perceived value of the sports event, even at the risk of downgrading the perceived value of the sports event.  McDaniel's (1999) study examining the match-up hypothesis is in line with the balance theory, in which the perceived match between brand attributes and the sponsor's attributes influences consumers' attitude toward an ad.
        From a theoretical perspective, Keller (1993) suggested brand association in which a pre-existing consumer's memory regarding the sports event (or celebrity) becomes linked to the memory of the brand when that brand sponsors a sports event.  Therefore, a positive image of a sports event can positively influence the image of a brand.
Celebrity Endorsement.  The celebrity endorser is defined as "an individual who is known to the public (i.e., actor, sports figure, entertainer) for his or her achievements in areas other than that of the product class endorsed" (Friedman and Friedman 1979, p. 63).  Celebrities use and enjoy appearing in advertising for consumer goods (McCracken 1989).  Marketers and advertisers expect that celebrity endorsement will enhance the impact of advertising, because a celebrity endorser appeals to a target consumer group, and the life experiences of the endorsers can easily fit the advertising message (Mathur, Mathur, and Rangan 1997).
The notion that celebrity endorsement has a positive impact on consumers' perception of and response to advertising is explained in terms of source attractiveness, source credibility, and meaning transfer.
McGuire (1985) suggested the source attractiveness model.  This model indicates that the effectiveness of a message is determined by the characteristics of the source such as familiarity, likeability and similarity.  People gain familiarity with the source through repeated exposure to media, likeability of the source in terms of physical attributes and behavior, and the perception of similarity between the consumers and the source.
Another approach to celebrity endorsement is source credibility, suggested by scholars such as Ohanian (1990; 1991), and Gotlieb and Sarel (1991).  According to this approach, advertising is enhanced when the endorsers of advertisements are highly evaluated in terms of their expertise, trustworthiness, and perceived social value.
McCracken (1989) suggested meaning transfer theory to explain consumers' perceptions of celebrity endorsement.  Meaning indicates an overall evaluation of a celebrity determined by the consumer's interpretation of the public images of the celebrity conveyed by several routes (e.g., television, movies).  The image of a certain celebrity in a consumer's memory transfers to the product when the two are paired in an advertisement.  Gwinner and Eaton (1999) further suggested that the meaning not only transfers from the celebrity to the product but also transfers from the product to the consumers when consumers acquire or consume the product.

Previous Research & Hypotheses

Gwinner and Eaton (1999) suggested similarities between celebrity endorsement and sport sponsorship in advertising.  They suggest that consumers can draw particular meanings associated with both celebrities and sports teams.  For example, the meaning associated with a celebrity is drawn from their public activities, while consumers derive the meaning of a sports event from its type or size.  Javalgi et al. (2001) support this view.  They indicate that a company can share the sports event image with its product or brand in similar ways that a product shares the image of a celebrity in an advertisement.
By the association between the object of the sponsor (i.e., sports event) and the sponsor, and between the brand and the celebrity, the images of the object of the sponsor and/or the celebrity can be transferred to the sponsoring brand (Keller 1993; McCracken 1989).  Therefore, a positive attitude toward the object of the sponsor and/or the celebrity is likely to transfer to the advertised brand, when the object of the sponsor or the celebrity are highly recognized by people.
For example, based on the results of an exploratory study, Javalgi et al. (1994) suggest that, despite mixed findings, a corporate image can be improved by corporate sponsorship.  They found that companies that use corporate sponsorship have better public images than those that do not.  Dean (1999) examined the effects of third-party endorsement and event sponsorship as extrinsic advertising cues and found that both advertising cues have statistically significant effects on perceived quality, uniqueness and esteem of a brand.  Fireworker and Friedman (1977) found that expert, celebrity, and typical consumer endorsements had positive effects on the overall attitude toward a product and its expected price, although purchase intention was not affected.  Friedman and Friedman (1979) further developed one of their own studies (1977) and discovered that celebrity endorsement was more effective than expert or typical consumer endorsement in the ability to recall brand name and the ad.  A
lso, Misra and Beatty (1990) found that the effect of a spokesperson in advertising is easily transferred to the brand, which enhances a positive attitude toward the brand when the brand and the spokesperson are matched.
Given that sports stars are celebrities with well-developed and well-known public images, sports stars as endorsers are expected to generate positive attitudes toward a brand and a substantial degree of purchase intention.  Also, if consumers are given stimuli that the advertised brand is an official sponsor of the sport star, such stimuli act as a sponsorship cue, enhancing positive attitudes toward the brand and purchase intention.
Based on the theoretical and empirical evidence discussed above, the first set of hypotheses was drawn:
H1:

Advertising with celebrity endorsement will have more advertising effects in terms of attitude toward the brand and purchase intention than advertising without a celebrity endorsement cues.
H2:

Advertising with a sponsorship cues will have more advertising effects in terms of attitude toward the brand and purchase intention than advertising without a sponsorship cues.
H3:


Advertising with both sponsorship and celebrity cues will have an additive effects in terms of attitude toward the brand and purchase intention, compared to the advertising with either just sponsorship or celebrity cues (sponsorship x celebrity interaction effect).


Ethnic Background of Endorser in Advertising

Testing the previous three hypotheses does not answer whether consumers of a certain ethnic group equally perceive advertisements regardless of the endorser's ethnic background.  That is, is the Anglo-American consumers' perception of advertising with an Anglo-American endorser as great as the perception of advertising with a non-Anglo-American endorser?   The following section forms a theoretical and empirical basis to answer this question, which leads to additional hypotheses.
Sense of Self.  According to Belk (1988), people regard their possessions as parts of themselves.  Possessions not only include visible objects but intangibles such as events, sports teams, or acquaintances, by knowing them.  People build up emotional attachments to possessions, which become what Belk calls an extended-self.  The concept of the extended-self is applicable to both the individual and collective levels.  For instance, family group, cultural, and national identities illustrate collective levels of the extended-self.  Also, these levels of self were premised to stereotype certain behaviors, as Belk (1988) noted that people could share a part of their extended-self with others.  This explains civic responsibility, charity, and patriotism.
Belk's (1988) extended-self is in line with the concept of social alliance, which implies that people behave in prototypical ways by accepting the norms, values or goals of a group to which they belong.   This happens when discrete social categories such as age, religion or sex become incorporated with a person's sense of self (Hogg and Abrams 1988; Tajfel 1982).
Therefore, it is reasonable that belonging to a certain ethnic group forms certain shared values followed by the members of the group.  Also, such shared values can include the inherent attributes of celebrities such as ethnic background.
Social Identification.  Identification is viewed as the perceived belongingness to a group of which the person is a member, and with which a person identifies (Bhattacharya 1995).  Social identity consists of salient group classifications, usually based on gender, race or demographic categories (Bhattacharya 1995).
When a person identifies him/herself with a certain group, a sense of connectedness with the group is generated (Mael and Ashforth 1992).  With such connectedness, people act in a congruent way that appeals to the important aspects of their identity and the group they identify with (Bhattacharya 1995).  This is because people follow group norms that regulate people's expected behavior in a given situation (Madrigal 2000).  According to Turner (1984), sharing the same social identity results in a psychological group with a common mentality, thus, creating loyal citizens or sports fans within the group.
Oliver (1999) noted that consumers immerse their identities in the social setting to which a brand belongs when individual fortitude and social support are available.  Thus, consumers eagerly desire products affiliated with the social system (i.e. religious institutions and fan clubs) supporting the group norm and offer their patronage.
Parasocial Relationships.  Individuals build a sense of intimacy with a media personality by repeated exposure to the medium that features the particular celebrity.  This phenomenon is called a parasocial relationship (Horton and Wohl 1956; Basil 1996).  As a result, individuals develop identification with the media personality.  If a person has an ongoing relationship with the celebrity, he/she is likely to have strong identification with such a celebrity, which leads to the adoption of thoughts and behavior advocated by that celebrity.
Given that sports stars are some of the celebrities most frequently exposed on mass media, it is common for most people to have a sense of identification with sports stars.  Thus, if a sports star endorses an advertisement and advocates a specific brand, such endorsement and advocacy would influence viewers to follow the celebrity's "opinion," which can ultimately change consumers' previously existing attitude or behavior patterns.

Previous Research and Hypotheses

Applying the three concepts - self, social identification and parasocial relationship - to celebrity endorsement and sponsorship advertisement, it is expected that the ethnic background of endorsers would influence consumers' perception of advertising.
Specifically, by sharing the same ethnic background, consumers are likely to feel that the celebrities in the advertisements are a part of themselves, in terms of Belk's (1988) extended-self, which leads to a strong identification with the celebrities who belong to the same ethnic classification.  Also, consumers are more likely to form a strong parasocial relationship with celebrities of the same ethnic background because of the enhanced media availability.  Therefore, people are more likely to build a feeling of connectedness with the celebrity endorser of the advertising when the endorser and the consumers share the same ethnic background.  This, in turn, leads to a prototypical behavior such as becoming a loyal consumer for a certain brand that appears in the advertisement.
One example of such a study was conducted by Armstrong (1999).  She argues that marketing communications such as advertising have to appeal to target consumers in a culturally appropriate manner.  In an analysis of five NIKE commercials targeting black consumers, Armstrong (1999) found that NIKE apparently sought to appeal to African American consumers by intentionally incorporating African-American culture that emotionally linked the advertisements with African-Americans.  Also, Armstrong (1999) noted that the effects of advertising could be increased by allowing consumers to maintain a sense of association with products or services endorsed in the advertisement.
Roslow and Nicholls (1996) got similar results from their experiment, which examined viewers' purchase intentions when exposed to commercials manipulated by various cultures.  The findings suggest that Hispanic viewers are more persuaded when commercials are delivered in Spanish, and embedded in Spanish television programs, than commercials in English, and embedded in English television programs.
Madrigal (2000) found that respondents were more likely to have favorable intentions toward purchasing a product from a corporate sponsor if they believed that this intention was in line with the norms of other fans.  Also, a higher level of team identification generated stronger intentions to purchase a sponsor's products.  Madrigal (2000) argues, "psychological connectedness to a sports team represents an important aspect of self identity that contributes to a group norm, which in turn prescribes certain behavioral intentions that are considered to be supportive of the team."
Shimp (1997) also argues that people exposed to advertisements containing elements drawn from the target consumers' culturally-constituted world are more active in processing information in the advertising and in assigning meaning to the advertised product.
Therefore, when looking at the same advertisement containing endorsers of two different ethnicities, the advertising messages would be more effective when the consumers and the endorsers share the same ethnicity by building a cultural identity.
Based on the theoretical and empirical evidence discussed above, the second set of hypotheses was drawn:
H4:
Subjects will feel stronger identification with an endorser of their own ethnicity than with an endorser of a different ethnicity in advertising.
H5:
Endorsement advertising will have greater effects in terms of attitude toward the brand and purchase intention when the endorser shares the same ethnicity as the subjects.



METHOD


Overview of Research Design

The purpose of this study is to examine the effects of sponsorship, celebrity and the endorser's ethnicity on consumers' perception of advertising.
A total of 200 undergraduate students participated in this study to receive extra credit.  Only Anglo-Americans were included in the study because one of the purposes of this study was to examine whether the ethnic differences of the ad endorsers would have an impact on a particular ethnic group.

[Table 1] Experimental Design
Celebrity: Yes
Celebrity: No
Anglo-American Endorser
Asian
Endorser
Anglo-American Endorser
Asian
Endorser
Sponsorship: Yes
n = 25
Sponsorship: No


As shown in Table 1, subjects were randomly assigned to each of the eight cells in a 2 (sponsorship: yes or no) by 2 (celebrity: yes or no) by 2 (ethnicity of endorser: Anglo-American or Asian) factorial design resulting in 25 subjects per cell.  Therefore, three external advertising cues - celebrity, sponsor, and the ethnicity of endorser - served as independent variables in this study.
David Duval of the PGA Tour and Chan Ho Park of Major League Baseball were chosen as celebrities, and NIKE was chosen as the sponsoring brand.  This is because Duval and Park share the same sponsoring brand, NIKE, and both are similar in popularity among college students.  Also, one (Duval) is Anglo-American, while the other (Park) is an Asian sports athlete.  The choice of celebrities and the sponsoring brand was to make it possible to examine the effects of celebrity, sponsorship and the ethnic background of the endorser in a single experiment.
As McDaniel (1999) said, it was important to choose a well-known brand because sponsorship tends to be a strategy applied to mature brands.  Likewise, it was necessary to choose real celebrities because the subjects needed well-developed schematic memories to match the sponsoring brand and the celebrities.  Therefore, applying a fictional brand and an unknown celebrity to sponsorship research would have had little applied value (McDaniel 1999).
To control the confounding effects caused by the subjects' personal preferences toward NIKE, the pre-test included items to measure the subjects' pre-existing attitudes toward NIKE.  This was used as a covariate in the data analysis stage, which is explained specifically in the latter part of this paper.


Stimulus Material

There were eight versions of the NIKE advertisements (See Appendix A).  Each advertisement was identical except for the manipulations of the independent variables: sponsorship (yes or no), celebrity (yes or no) and the endorser's ethnicity (Anglo-American or Asian).  All advertisements contained one NIKE logo, NIKE's slogan "Just do it" and the endorser's photo (whether the endorser was a celebrity or non-celebrity).
In the sponsorship condition, textual stimuli, "NIKE is proud to be a sponsor of (athlete name)," were stated in the middle of the advertisement, whereas, in no sponsorship condition, there were no such textual stimuli.  This method has been adopted from Dean's (1999) study of sponsorship.
In the celebrity condition, Duval and Park endorsed the advertisements, whereas, in the non-celebrity condition, an average-looking Anglo-American golfer and Asian baseball player endorsed the advertisement.  To ensure that Duval and Park were highly recognized, a short description of each athlete accompanied their photos: "David Duval of the PGA Tour" and "Chan Ho Park of the Los Angeles Dodgers."  The official logos of the PGA Tour and the Major League Baseball were also displayed.
In the Anglo-American endorser conditions, Duval or the average-looking Anglo-American golfer endorsed the advertisement, whereas, in the Asian endorser conditions, Park or an average-looking Asian baseball player endorsed the advertisement.
All stimulus advertisements, along with the corresponding questionnaires, were posted on the Web, and the URLs of the experimental sites were delivered by the researcher via emails.  The subjects accessed the survey site individually and participated in the experiment as described in the "Procedure" section below.

Procedure
The present study consists of two parts: pre-test and main study.  For the pre-test, the researcher sent out emails to subjects to inform them of the URL to the Web site where a questionnaire measuring the subjects' pre-existing attitudes toward NIKE was posted.  Pre-existing attitudes toward the brand were measured as a covariate in order to control the confounding effects generated by the subjects' pre-existing attitudes toward the brand.  The subjects were asked to click on the link and complete the questionnaire on-line.  The pre-test not only included the measure of pre-existing attitudes toward NIKE, but also the measures of attitudes toward five other commercial brands, in order to desensitize the subjects' response to NIKE in the main study.  Once the subjects completed answering the pre-test, they clicked on the "Submit" button at the end of questionnaire, then the data were transferred to a Web space accessible to the researcher.
After completing data collection for the pre-test, an interval of seven days was given before the main study was conducted, in order to ensure that the pre-test did not affect the way the subjects answered questions in the main study.  The URL of the Web site containing the advertisements and questionnaires for the main study was delivered by emails.
The first page of the main study on the Web site contained instructions necessary to complete the study.  Subjects were instructed not to click or surf the Web except for when answering questions in the questionnaire.  Each subject was exposed to a total of three advertisements (one target advertisement and two filler ads).  The site was designed to automatically refresh in the order of the first filler ad, the target ad, and the second filler ad, as intended by the researcher.  Each subject was exposed to each of the three ads (two filler ads and one target ad) for 20 seconds, enough time for subjects to view the advertising.
The instruction page was linked to the web page containing the first filler ad.  The first filler ad was linked to the target advertisement (one of the eight cells of the experimental design).  The target advertisement was linked to the second filler ad.  The two filler ads were inserted before and after the target ad in order to remove any primacy or recency effect.
The second filler ad was linked to the questionnaire site.  Upon completion of the questionnaire, subjects were instructed to click on the "Submit" button, thereby sending data to a Web space designated by the researcher.

Dependent Variables and Measurement

        There were three dependent variables: attitudes toward the brand, purchase intention and identification.  To measure the attitudes toward the brand, Till and Shimp's (1998) scale was used.  The subjects were asked to rate  "How would you describe your overall feeling on NIKE?" on three 7-point semantic differential scales, anchored by favorable/unfavorable, positive/negative, and strongly dislike/strongly like.  Subjects were also asked to rate "What do you think about the product NIKE?" on three scales - bad/good, low quality/high quality, and inferior/superior.  The ratings of the six scales were averaged to represent the attitudes toward the brand.
Purchase intention was measured by applying the scales used by Machleit and Wilson (1988), and Petty, Cacioppo and Schumann (1983).  Subjects were asked to rate "How likely is it that you would purchase a product of NIKE the next time you need a product of this nature?" for the following three items: probable/improbable, likely/unlikely and possible/impossible.
Identification was measured by using Bhattacharya's (2001) five-point Likert-type scale, anchored by strongly disagree and strongly agree.  The actual statements were: "When someone criticizes (the celebrity's name), it feels like a personal insult"; "I am very interested in what others think about (celebrity's name)"; "(celebrity's name)'s success is my success"; "When someone praises (celebrity's name), it feels like a personal compliment"; "If a story in the media criticized (celebrity's name), I would feel embarrassed."
The coefficient alpha for each dependent variable was as follows: pre-existing attitude toward the brand (.93), attitude toward the brand (.96), purchase intention (.97), and identification (.91).








FINDINGS


Overview.  A total of 200 subjects (each of the eight cells contained 25 subjects) were analyzed.  Ninety-two subjects (46 %) were male and 108 subjects (54 %) were female.  In order to test the hypotheses, a series of two-way and three-way ANCOVAs and ANOVAs were run.  Two dependent variables - attitudes toward the brand, and purchase intention - were independently entered into the ANCOVA, with sponsorship, celebrity, and endorser's ethnicity as independent variables; and pre-existing attitude toward the brand as a covariate.  The covariate was included in the analysis to preclude any confounding effects caused by the subjects' response to the brand NIKE instead of to the experimental manipulation.  And the other dependent variable, identification, was entered into ANOVA with the same independent variables but no covariate.  All of the Levene's tests were non-significant at p > .05, indicating equality of error variances across the treatment groups on each dependent variable.
Preliminary Analysis.  The means and standard deviations of each dependent variable for the eight experimental cells are shown in Table 2.  The familiarity between the Anglo-American celebrity and the Asian celebrity showed no significant difference at p < .05, assuring that the effects from the two celebrities were kept equal.  In addition, the correlations of each pair of the three dependent variables were significant at p < .001, suggesting linear relationships.  Pearson's correlation coefficient between purchase intention and attitudes toward the brand was .64, the correlation between attitudes toward brand and identification was .22, and the correlation between identification and purchase intention was .26.

[Table 2]
Mean and Standard Deviations for Each Experimental Cell
On Attitudes Toward Brand, Purchase Intention, and Identification.

Yes Celebrity
No Celebrity
Anglo-American Endorser
Asian Endorser
Anglo-American Endorser
Asian Endorser
Yes Sponsor
Attitude Toward Brand
6.00 (.65)
5.52 (.96)
5.17 (1.01)
4.51 (1.00)
Purchase Intention
5.27 (1.17)
4.64 (1.70)
4.40 (1.52)
3.57 (1.70)
Identification
1.92 (.81)
1.96 (.78)
1.85 (.82)
1.31 (.48)
No Sponsor
Attitude Toward Brand
5.1 (.95)
5.05 (.97)
5.28 (1.39)
4.59 (1.30)
Purchase Intention
4.33 (1.54)
4.23 (1.49)
4.67 (1.65)
4.20 (1.93)
Identification
1.24 (1.51)
1.85 (.81)
2.61 (.76)
1.60 (.71)
Note: Standard deviations are in parentheses.

Effects of Sponsorship.  The main effect of sponsorship was detected in the attitudes toward the brand.  Subjects exposed to an advertisement accompanying a sponsorship cues (M = 5.30) generated a significantly greater degree of positive attitudes toward the brand than did the subjects exposed to an advertisement without a sponsorship cues (M = 5.01; F (1, 197) = 4.17, p < .05).  However, no main effect was detected on the purchase intention, between the subjects exposed to an advertisement accompanying the sponsorship cues (M = 4.58) and subjects exposed to an advertisement without the sponsorship cues (M = 4.36; F (1, 197) = .78, p > .05).  Therefore, Hypothesis 1, which proposed that sponsorship would significantly enhance both attitudes toward the brand and purchase intention, was partly supported.
Effects of Celebrity.  The main effect of celebrity was detected in both attitudes toward the brand and purchase intention.  Subjects exposed to an advertisement
[Figure 1]
Interaction Effects of Sponsorship and Celebrity on Attitude Toward Brand and Purchase Intention


  [--- WMF  Graphic Goes Here  ---]



  [--- WMF  Graphic Goes Here  ---]


Note: Both interaction effects are statistically significant at p < .05


accompanied by a celebrity cues generated a significantly greater degree of positive attitude toward the brand (M = 5.42) than did subjects exposed to an advertisement without a celebrity cues (M = 4.89; F (1, 197) = 16.18, p < .01).  A similar effect was detected on the subjects' purchase intention.  The subjects in the celebrity condition generated a significantly greater degree of positive attitude toward the brand (M = 4.73), than did subjects in the no celebrity condition (M = 4.21; F (1, 197) = 5.88, p < .05).
Therefore, Hypothesis 2, which proposed that celebrity would significantly enhance both attitudes toward the brand and purchase intention, was upheld.
Interaction Effect.  As shown in Figure 1, the interaction effect between sponsorship and celebrity on both attitudes toward the brand (F (1, 195) = 4.13, p < .05) and purchase intention (F (1, 195) = 6.18, p < .05) was statistically significant.  Therefore, the effect of the celebrity on attitude toward the brand and purchase intention differs, depending on whether a sponsorship cue is provided in the advertising.
While the subjects of the celebrity condition generated a greater degree of positive attitude towards the brand when the sponsorship cue was provided than when there was no sponsorship cue, the subjects of the no celebrity condition generated a weaker degree of positive attitude toward the brand when a sponsorship cue accompanied the advertisement than when there was no sponsorship cue.
The same pattern occurs with purchase intention.  While the subjects of the celebrity condition generated a greater degree of purchase intention when a sponsorship cue was provided (M = 5.18) than when there was no sponsorship cue (M = 4.28), the subjects of the no celebrity condition generated a weaker degree of purchase intention when a sponsorship cue accompanied the advertisement (M = 3.99) than when there was no sponsorship cue (M = 4.43).
Therefore, Hypothesis 3 predicting the interaction effect between sponsorship and celebrity on both attitudes toward the brand and purchase intention, was supported.
Identification.  The main effect of the endorser's ethnicity on identification was detected.  The subjects (all Anglo-American) felt a stronger identity with the Anglo-American endorser (M = 1.91) than with the Asian endorser (M = 1.68; F (1, 198) = 3.92, p < .05).  Therefore, Hypothesis 4, which proposed that the subjects would feel a stronger identity with the endorser of the subjects' own ethnicity than with an endorser of a different ethnicity, was supported.
Effect of Endorser's Ethnicity.  The main effect of the endorser's ethnicity on the attitudes toward the brand was detected.  Subjects (all Anglo-American) exposed to an advertisement with American endorser generated a significantly greater degree of positive attitude toward the brand (M = 5.39) than did the subjects exposed to an advertisement with the Asian endorser (M = 4.92, F (1, 197) = 5.73, p < .05).  However, no main effect was detected on purchase intention at p < .05, although the difference in purchase intention between the subjects exposed to an advertisement with an American endorser (M = 4.78) versus with an Asian endorser (M = 4.16; F (1, 157) = 3.85, p = .051) was marginally significant.
Therefore, Hypothesis 5, which proposed that subjects would have more of a positive attitude toward the brand and a stronger purchase intention when the endorser of the advertisement had the same ethnic background as the subjects, was partly supported.


























CONCLUSIONS


This study investigated the effects of three advertising cues - sponsorship, celebrity endorsement, and the ethnicity of the endorser - on the consumers' perception of advertising, in terms of attitudes toward the brand, purchase intention and identification.
The overall results are noteworthy, in that subjects were more likely to have a positive attitude toward the brand when the advertisement included a celebrity endorsement, sponsorship cue, or an endorser of the same ethnicity as the subjects.  These findings are congruent with the conclusions of various sponsorship and celebrity researchers such as Javalgi et al. (1994) and Dean (1999).
However, interestingly enough, only celebrity endorsement was found to be a statistically effective factor that influenced a consumer's purchase intention (albeit the endorser's ethnicity also had some marginal effect on the consumer's purchase intention).  This indicates that having a positive attitude toward the brand does not necessarily enhance the consumer's purchase intention.
Such a conclusion might be explained by Petty and Cacioppo's (1983; 1986) Elaboration Likelihood Model (ELM), which suggests two routes in the consumer's consideration of a message: the central or the peripheral route.  According to ELM, processing information via the central route involves a person's diligent consideration of information.  This results in a relatively enduring attitude change, which is predictive of behavior.  However, messages processed via the peripheral route are made by a person's simple inferences about the message, based on various simple cues, such as attractiveness of the endorser in an advertisement.  Attitude changes induced via the peripheral route are presumed not to be predictive of behavior.
Therefore, according to ELM, if information on sponsorship and/or the ethnicity of the endorsers was processed via the peripheral route, the attitude changes do not necessarily affect the consumer's purchasing behavior.  Similarly, the celebrity may induce persuasion through the central route by providing product-relevant arguments (Petty and Cacioppo 1983), even though celebrity endorsement in advertising is often perceived as a peripheral cue.  In other words, two athletes endorsing the advertisements of a sports brand, as in the present study, might offer a product-relevant argument, so that the subjects are induced into processing the message via the central route, which influences the subjects' attitudes as well as their behavior patterns.  A similar interpretation was made by Petty, Cacioppo and Schumann (1983) when they reviewed the results and reassessed their conclusions from the 1980 study.  They said that an attractive celebrity provided a product-relevant argument for
a beauty product, and that such information could be processed via the central route.
As hypothesized, a statistically significant interaction effect between sponsorship and celebrity on both attitudes toward the brand and purchase intention was detected.  This result reflects an additive effect of having two advertising cues, over having only one of the two advertising cues (sponsorship or celebrity cues).  In other words, having two advertising cues together better persuades consumers.  On a practical level, this suggests that advertising sports-related brands would be more effective when using sponsored athletes as endorsers rather than using merely well-known public figures.  Interestingly, the attitudes toward the brand and the purchase intention were weaker when no celebrity endorsement was provided with a sponsorship cue, than when no celebrity endorsement was provided without a sponsorship cue.  One possible explanation is that consumers might have felt that the association between sponsorship and a non-celebrity endorser is illegitimate, which, therefore,
resulted in a negative impact on the attitude toward the brand and the purchase intention.
It is also noteworthy that subjects felt a stronger identification with endorsers of their own ethnicity, which supports various studies on targeting consumers by including elements in advertisements that culturally appeal to the potential consumer (i.e., Shimp 1997; Armstrong 1999; and Nicholls 1996).  Therefore, the result of this study reconfirms the fact that marketers and advertisers targeting a specific consumer group should be aware of the consumer group's specific characteristics (Armstrong 1999), including the major ethnicity of the target group members.
Interestingly, although not hypothesized, a significant interaction effect between sponsorship and celebrity on identification was also detected.  This interaction effect revealed that the subjects exposed to advertising with a sponsorship cue felt a stronger identification with the endorser when the endorser was a celebrity, than when the endorser was a non-celebrity.  In reverse, subjects exposed to advertising without a sponsorship cue felt a relatively weaker identification when the endorser was a celebrity, than when the endorser was a non-celebrity.  This result indirectly suggests that a sponsorship cue might aid in generating a stronger degree of identification when used in a celebrity endorsement advertisement, thus enhancing the effectiveness of the advertising message.
There are several limitations to this study.  Although this study was purposely conducted on-line to offer a natural situation for subjects (i.e., the subjects were surfing the Web as in a real, everyday environment), the power to control the experiment process was not as strong as it would have been in a laboratory experimental situation.  For example, there was no way to control whether the subjects were answering the questions with their own opinion or with others' aid, although all of the subjects were instructed to answer based on their own thinking.  Additionally, the purpose of sponsorship in this study was presumed to enhance the attitude toward the brand and the purchase intention, but, as Dean (1999) noted, brand awareness may be the primary goal of sponsorship, which was not tested in this study.
Other shortcomings include that, first, this study did not differentiate the level of the athletes' prestige.  It is necessary to examine the difference (or similarity) in advertising effects caused by selecting different athletes with different levels of prestige.  Secondly, in terms of the ethnicity of the endorser, this study should be replicated in an Asian country to see if similar results are produced.  In addition, besides exposing two different ethnically manipulated advertisements to subjects of one ethnic group, it would be useful to compare, for instance, the effects of the same advertisement when exposed to Anglo-Americans in the U.S. where they are a majority and those in other countries where they are a minority.  This could be hypothesized and tested under McGuire's (1984) distinctiveness theory suggesting that "the lower the proportion of minority group members in the overall population, the more likely that ethnically targeted stimuli (such as the use of a ethnic
spokesperson in an ad) will be effective" (Deshpande and Stayman 1994, p. 57).  Finally, for sponsorship research purposes, it is still hard to draw a line between sponsorship and advertising, although several scholars such as McDaniel (1999) put effort to describing the difference between the two.  It is necessary to first define what sponsorship is in order to thoroughly investigate its effects and not to confuse it with other similar concepts in academia.
        Nonetheless, the present study was worthwhile in that it confirmed previous celebrity endorsement and sponsorship studies, and as an initial effort examining the effects of sponsorship on individual athletes, but there is more to be studied in this field.
































REFERENCE


Adler, Peter and Patricia A. Adler (1987), "Role Conflict and Identity Salience: College
Athletics and the Academic Role," Social Science Journal, 24, 443-455.

Armstrong, Ketra L. (1999), "Nike's Communication With Black Audiences,"  Journal of Sport & Social Issues, 23(3: August), 266-286.

Basil, Michael D. (1996), "Identification as a Mediator of Celebrity Effects," Journal of Broadcasting & Electronic Media, 40, 478-495.

Belk, Russel W. (1988), "Possessions and the Extended Self," Journal of Consumer Research, 15(Septempber), 139-168.

Bhattacharya, C. B., Hayagreeva Rao and Mary Ann Glynn (1995), "Understanding the Bond of Identification: An Investigation of Its Correlates Among Art Museum Members," Journal of Marketing, 59(October), 46-57.

Cornwell, T. Bettina, Stephen W. Pruitt, and Robert Van Ness (2001), "The Value of Winning in Motorsports: Sponsorship-linked Marketing," Journal of Advertising Research, 41(January/Feburary), 17-31.

Crimmins, James, and Martin Horn (1996), "Sponsorship: From Management Ego Trip to Marketing Success," Journal of Advertising Research, 36(July/August), 11-21.

Cunningham, M. H., and S. F. Taylor (1995), "Event Marketing: State of the Industry and Research Agenda," Festival Management & Event Tourism, 2, 123-137.

Dean, Dwane Hal (1999), "Brand Endorsement, Popularity, and Event Sponsorship as Advertising Cues Affecting Consumer Pre-Purchase Attitudes," Journal of Advertising, 28(3), 2-12.

Deshpande, Rohit and Douglas M. Stayman (1994), "A Tale of Two Cities: Distinctiveness Theory and Advertising Effectiveness," Journal of Marketing Research, 31(February), 57-64.

Fireworker, Robert B. and Hershey H. Friedman (1977), "The Effects of Endorsements on Product Evaluation," Decision Science, 8(July), 576-583.

Friedman, Hershey H. and Linda Friedman (1979), "Endorser Effectiveness by Product Type," Journal of Advertising Research, 19(October), 63-71.

Gotlieb, Jerry B. and Dan Sarel (1991), "Comparative Advertising Effectiveness: The Role of Involvement and Source Credibility," Journal of Advertising, 20(1), 38-45.

Gwinner, K. (1997), "A Model of Image Creation and Image Transfer in Event Sponsorship," International Marketing Review, 14, 145-158.

Gwinner, Kevin P. and John Eaton (1999), "Building Brand Image Through Event Sponsorship: The Role of Image Transfer," Journal of Advertising, 27(4), 47-57.

Heider, Fritz (1946), "Attitudes and Cognitive Organization," Journal of Psychology, 21, 107-120.

Heider, Firtz (1958), The Psychology of Interpersonal Relations, NY: John Wiley & Sons, Inc.

Hogg, Michael A. and Dominic Abrams (1988), Social Identifications: A Social Psychology of Intergroup Relations and Group Processes, London: Routledge.

Horton, D. and Wohl, R. R. (1956), "Mass Communication and Parasocial Interaction: Observations on Intimacy at a Distance," Psychiatry, 19, 215-229.

Javalgi, Rajshekhar G., Mark B. Traylor, Andrew C. Gross, and Edward Lampman (1994), "Awareness of Sponsorship and Corporate Image: An Empirical Investigation," Journal of Advertising, 23(4), 47-58.

Keller, K. L. (1993), "Conceptualizing, Measuring and Managing Customer Based Brand Equity," Journal of Marketing, 57, 1-22.

Machleit, K. A. and R. D. Wilson (1988), "Emotional Feelings And Attitude Toward The Advertisement: The roles of brand familiarity and repetition," Journal of Advertising, 17(3), 27-35.

Mael, Fred and Blake E. Ashforth (1992), "Alumni and Their Alma Mater: A Partial Test of the Reformulated Model of Organizational Identification," Journal of Organizational Behavior, 13, 103-123.

Madrigal, Robert (2000), "The Influence of Social Alliances with Sports Teams on Intentions to Purchase Corporate Sponsors' Products," Journal of Advertising, 29(4), 14-24.

Mathur, Lynette Knowles, Ike Mathur and Nanda Rangan (1997), "The Wealth Effects Associated with a Celebrity Endorser: The Michael Jordan Phenomenon," Journal of Advertising Research, 37(May/June), 67-73.

McCracken, Grant (1989), "Who Is The Celebrity Endorser?  Cultural Foundations of the Endorsement Process," Journal of Consumer Research, 16(December), 310-321.

McDaniel, Stephen R. (1999), "An Investigation of Match-Up Effects in Sport Sponsorship Advertising: The Implications of Consumer Advertising Schemas," Psychology & Marketing, 16(2), 163-184.

McGuire, William J. (1984), "Search for the Self: Going Beyond self-esteem and the Reactive Self," in Personality and the Prediction of Behavior, R.A. Zucker, J. Aronoff, and A. I. Rabin, eds.  New York: Academic Press, 73-120.

McGuire, William J. (1985), "Attitudes and Attitude Change," in Gardner Lindzey and Elliot Aronson (eds.), Handbook of Social Psychology, 2, 233-346, New York: Random Hourse,

Meenaghan, Tony (1991), "The Role of Sponsorship in the Marketing Communications Mix," International Journal of Advertising, 10(1), 35-47.

Misra, Shekhar and Sharon E. Beatty (1990), "Celebrity Spokesperson and Brand Congruence: An Assessment of Recall and Affect," Journal of Business Research, 21, 159-173.

Ohanian, Roobina (1990), "Construction and Validation of a Scale to Measure Celebrity Endorsers' Perceived Expertise, Trustworthiness, and Attractiveness," Journal of Advertising, 19(3), 39-52.

Ohanian, Roobina (1991), "The Impact of Celebrity Spokesperson's Perceived Image on Consumer's Intention to Purchase," Journal of Advertising Research, 31(February/March), 46-54.

Oliver, Richard L. (1997), Satisfaction: A Behavioral Perspective on the Consumer, New York: Irwin/McGraw-Hill.

Oliver, Richard L. (1999), "Whence Consumer Loyalty?" Journal of Marketing, 63 (Special Issue), 33-44.

Petty, Richard E., John T. Cacioppo and David Schumann (1983), "Central and Peripheral Routes to Advertising Effectiveness: The Moderating Role of Involvement," Journal of Consumer Research, 10(September), 135-146.

Petty, Richard E. and John T. Cacioppo (1980), "Effects of Issue Involvement on Attitudes in an Advertising Context," in Proceedings of the Division 23 Program, eds. Gerald G. Gorn and Marvin E. Goldberg, Montreal, Canada: American Psychological Association, 75-79.

Petty, Richard E. and John T. Cacioppo (1986), "The Elaboration Likelihood Model of Persuasion," in Leonard Berkowitz (eds.), Advances in Experimental Social Psychology, 19, 124-192, London: Academic Press, Inc.

Sandage, C. H. (1993), Roads to Be Taken, Center for the Study of Free Enterprise and Entrepreneurship, 213-267, Graceland College, Iowa.

Shimp, T. A. (1997), Advertising, Promotion, and Supplemental Aspects of Integrated Marketing Communications (4th ed.), Orlando, FL: Dryden.

Stipp, Horst, and Nicholas P. Schiavone (1996), "Modeling the Impact of Olympic Sponsorship on Corporate Image," Journal of Advertising Research, 36(July/August), 22-27.

Tajfel, Henri (1982), "Introduction," in Henri Tajfel (eds.), Social Identity and Intergroup Relations, New York: Academic Press.

Till, Brian D. and Terence A. Shimp (1998), "Endorsers in Advertising: The Case of Negative Celebrity Information," Journal of Advertising, 27(1), 67-82.

Turner, John C. (1984), "Social Identification and Psychological Group Formation," in Henri Tajfel (eds.), The Social Dimension: European Developments in Social Psychology, 2, 518-538, Cambridge, England: Cambridge University,



























APENDICS A
Stimulus Material

Anglo-American celebrity / sponsorship condition
Anglo-American celebrity / non- sponsorship condition



Asian celebrity / sponsorship condition
Asian celebrity / no sponsorship condition



Anglo-American non-celebrity / sponsorship condition
Anglo-American non-celebrity / no sponsorship condition



Asian non-celebrity / sponsorship condition
Asian non-celebrity / no sponsorship condition