Impact of perceived on-field performance on sport celebrity source credibility.

Koo, Gi-Yong ; Ruihley, Brody J. ; Dittmore, Stephen W. 等

Impact of Perceived On-Field Performance on Sport Celebrity Source Credibility

In the past decade, sport celebrities have been collectively earning millions of dollars, annually, from their endorsement contracts. In 2010, U.S. companies paid nearly $17.2 billion to leagues, teams, athletes, coaches, and sports personalities to endorse their goods and services, while worldwide spending on sponsorships continued to grow 5.2% to $46.3 billion (IEG, 2011). According to the company's 2010 annual report, Nike was projected to spend about $712 million dollars for endorsements using celebrity athletes. This figure calculated base endorsement compensation and minimum royalty fees paid to athletes and teams, excluding the cost of the products supplied to the endorsers (Nike, Inc., 2010). When athletes sign heavily financed endorsement deals, they receive compensation for endorsing a certain product or organization. The assumption is that the company will be able to reap major rewards from this financial commitment to an athlete via increased sales and use of the athlete's image.

The underlying principle of paying millions of dollars to celebrity athlete endorsers is that the source of the message will add credibility to an advertisement (Yoon, Kim, & Kim, 1998). The advertisers then consider source credibility as a significant basis for selecting a celebrity spokesperson. Ohanian (1990) provides a model of source credibility utilizing dimensions of perceived attractiveness, trustworthiness, and expertise as the primary characteristics in defining an endorser's source credibility. The more credible and attractive a spokesperson is, the more persuasive he or she will be as an endorser in order to generate favorable attitudes toward an endorsed brand or product (Miciak & Shanklin, 1994). Thus, a well-constructed endorsement can do much to enhance the attitude toward the brand and the purchase intentions; conversely, a poorly planned endorsement can have no effect or even an adverse effect.

Research indicates that as negative information is circulating about an athlete, a negative impact can be directed toward the endorsed brand or organization. With "a strong associative link between the celebrity and the brand, negative information about the celebrity will lower brand evaluations" (Till & Shimp, 1998, p. 72). In particular, off-the-field issues, outside of the physical realm of sport, have been considered as the source of negative information about an athlete (e.g., a scandal, an issue with the law, a crime, a fight, or some other negative situation outside of their athletic career). The negative behavior of sport celebrities receives more attention, is better encoded, and is more simply evoked than positive information (Money, Shimp, & Sakano, 2006). This phenomenon tends to result in a negative perception about an endorsed brand or product. With the recent cases of Tiger Woods and Michael Phelps, there is evidence supporting this circumstance. For instance, Gatorade immediately dropped its endorsement of Tiger Woods when his sex scandal became known to the general public, while a sponsor, Kellogg, using athletes' wholesome image to sell cereal products severed its ties from Michael Phelps because of his indictment for illegal drug use (Macur, 2009).

Many athletes sign tremendously financed endorsement deals based on their performance ability and potential, even though they have not proven their ability to compete at the professional level. A teenage golf star, Michelle Wie, was signing major endorsement contracts (e.g., Sony, Nike) without having played her first professional match. At the age of 17, Wie was a target for sports marketers because of her blossoming talent, personality, and desire to compete with male golfers (Story, 2005). Marketers felt that Wie could do things for women's golf comparable to what Tiger Woods has done for men's golf. However, for sport marketers, as in any endorsement situation, this could be a financial risk because off-the-field issues aside, there might be some possibility of reducing the ability of the athlete caused by an unexpected injury or simply poor performance during game play. Although previous research has examined how off-the-field issues related to the celebrity athlete (e.g., scandal) affect the endorsing brand, the majority of studies have not scrutinized the impact of on-field issues (e.g., performance) on an athletic endorser's source credibility and the potential impact on the endorsed brand (i.e., Louie, Kulik, & Jacobson, 2001; Money, Shimp, & Sakano, 2006; Till & Shimp, 1998).

Therefore, the purpose of this study is two-fold: (1) to examine the influence athletic performance has on the elements of source credibility, and (2) to investigate its impact on the causal relationships among consumers' brand attitude, attitude toward the advertisement, and purchase intentions. As the study has been conducted with an exploratory nature, the study could provide a better understanding of the impact that an athletic endorser's performance has on his or her overall source credibility. Thus, this study could serve as groundwork for sport marketers to make strategic decisions as to how to leverage their relationships with sport celebrities. It is necessary to recognize what impact sport celebrity endorsers have in advertising and on endorsed brands based on their current performance, given the issue of whether or not the sport celebrity's on-field success affects individual dimensions of source credibility.

Source Credibility

According to Ohanian (1990), source credibility is "a term commonly used to imply a communicator's positive characteristics that affect the receiver's acceptance of a message" (p. 41). Ohanian (1990 & 1991) identified and defined three dimensions of a credible source while the landmark study by Hovland, Janis, and Kelley (1953) employed two of the three dimensions in their initial work. Expertise was the first dimension, and it was defined as "the extent to which a communicator is perceived to be a source of valid assertions" (Hovland, Janis, & Kelley, 1953, p. 21). Essentially, it reflects the amount of knowledge the source (e.g., the endorser) has about the particular topic/product he or she is endorsing. In the field of sports, expertise is often determined by the athletic performance of the source. For example, when a professional golfer endorses a titanium driver, through his detailed description of the features and the benefits of the driver, consumers may feel that he is an expert not only because he knows about this driver, but also because he wins a major tournament using this endorsed titanium driver. Accordingly, athletic performance will drastically alter a consumer's view of his or her expertise. The perception of an athlete as an expert may significantly increase when the athlete wins consistently at his or her sport.

The second dimension of source credibility is trustworthiness. Trustworthiness "refers to the consumer's confidence in the source for providing information in an objective and honest manner" (Ohanian, 1991, p. 47). When an endorser is perceived to be highly trustworthy, a message delivered by him or her is more effective in changing attitude than those, from endorsers who have low source trustworthiness (Ohanian, 1990; Stevens, Lathrop, & Bradish, 2003). McGinnies and Ward (1980) indicated that source trustworthiness was correlated with source expertise. They found that an endorser who had both expertise and trustworthiness was most influential on the level of attitude change. A large number of sport fans regard Tiger Woods as being a talented golfer and an expert in the sport of golf; however, because of his indiscretions with regards to his marriage, they may not trust him. Tiger Woods' trustworthiness could have been sustained to a higher degree, had he continued to experience success in the PGA. Trustworthiness is an important construct in persuasion and attitude change (Ohanian, 1990). Thus, with an elevated importance on trustworthiness, a need exists to examine whether perceived athletic performance has an impact on an endorser's source credibility.

The last dimension of source credibility is attractiveness. Physical attractiveness is the perceived familiarity, likability, and similarity of the source to the receiver (Yoon, Kim, & Kim, 1998). Joseph (1982) indicated the attractiveness of endorsers resulted in a positive image and a positive evaluation of products with which they were associated. This phenomenon was consistent with the study conducted by Ohanian (1991) indicating that "physically attractive communicators are more successful in changing beliefs than are unattractive communicators" (p. 47). The current study has employed attractiveness as an important aspect to explain source credibility and has considered the relationships of three components (e.g., expertise, trustworthiness, and attractiveness) defining the latent construct, source credibility (Ferle & Choi, 2005).

For the last few decades, while certain dimensions of source credibility were widely agreed upon, a considerable amount of source credibility research found that a highly credible source effectively influences attitude changes and purchase intentions. For instance, Sternthal, Phillips, and Dholakia (1978) indicated that consumer attitudes and behavioral intentions were more influenced by a highly credible person than a less credible person. Atkin and Block (1983) also specified that celebrities were more effective in forming positive responses derived from consumers than non-celebrities. Findings were consistent with the study conducted by Ohanian (1991) indicating that a highly credible person was likely to make more positive attitude changes toward the endorsed product than a person who was considered to be a less credible source.

Priester and Petty (2003) also indicated that information delivered by trustworthy endorsers was likely to be taken at face value which, in turn, could create more favorable attitudes and a higher probability of the consumer's purchasing the endorsed product and service. Lafferty and Goldsmith (1999) presented that both corporate credibility and endorser credibility had a strong impact on attitude toward the advertisement, the brand attitude, and purchase intentions while Lafferty, Goldsmith, and Newell's (2002) dual credibility model added the weight of evidence that credible endorsers lead to positive attitude toward the advertisement, the brand attitude, and purchase intentions. Therefore, the widely dispersed recognition and popularity of an athletic endorser is expected to have a greater impact on attitudes toward services (Priester & Petty, 2003), the advertisement (Lafferty, Goldsmith, & Newell, 2002), the brand (Lafferty & Goldsmith, 1999), and purchase intentions (Lafferty & Goldsmith, 1999; Ohanian, 1991).

Theoretical Understanding of Source Credibility

The primary motivation for the development of cognitive structure theories was to look at the way people think. Wright's (1973) cognitive structure model (CSM), derived from Lavidge and Steiner's (1961) original hierarchy of effects model, has been used comprehensively in celebrity endorsement and spokesperson advertising research. This model aims to elucidate the way people process information. A concept that is closely related to the CSM is the fundamental role of the belief (cognitive) component (Olson, Toy, & Dover, 1982). The belief component is also known to produce a series of "primary thoughts" often mediating a message of acceptance as well as consecutively affecting consumer attitude and behavioral intentions (Fishbein & Ajzen, 1975; Wright, 1973). Wright (1973) introduced four different types of primary thoughts containing counterarguments, support arguments, source derogations, and curiosity statements. For instance, the source derogating response may be used in circumstances where an athletic endorser is easily viewed as having poor or unstable performance. In this context, the athlete endorser may spontaneously derogate the sponsoring brand, product, or advertising in general. In addition, Wright (1973) designated that the impact of the source derogation process on message acceptance in advertising may be as devastating as those of counter-arguing effects. Accordingly, understanding the primary thoughts (i.e., source derogations, curiosity statements, etc.) is as significant as measuring the relationships among the cognitive structure components (i.e., beliefs, attitudes, and purchase intentions) if one is to fully evaluate and understand the consumer's information process (Smith & Swinyard, 1988).

With regard to the CSM, Fishbein and Ajzen (1975) and Wright (1973) argued that a person evaluates incoming information with his or her existing knowledge, called schema. Also, a schema is defined as "an active organization of past experiences, which must always be supposed to be operating in any well-adapted organic response" (Fiske, 1982, p. 60). Since the primary motivation for the development of the CSM was to offer an alternative explanation for human information processing, the present study employed the CSM and its related concepts to explain how people process an athlete endorser's on-field performance and, in turn, to illustrate how their belief components might lead to favorable attitudes toward the endorsing brand.

This systematic information process is also consistent with the theoretical position derived from Fishbein and Ajzen (1975, p. 222) indicating that "a person's attitude is a function of his salient beliefs at a given point in time." For instance, an athlete endorser's on-field performance information can influence beliefs, opinions, attitudes, and/or behavior through an internalized process that occurs when people perceive a source influence in terms of their personal attitude and value structures (Kelman, 1961). In other words, the CSM posits a hierarchy of consumer responses of: belief--attitude--intentions--behavior.

In the field of sports, athletes acquire most of their credibility through their on-field performances taking place in real time. Athletes are regularly in the headlines, talked about on the radio, displayed on the Internet, and viewed on television. For many athletes, this exposure is what helps define them as a celebrity. In particular, athletes differ from other celebrities in the fashion in which their performance affects credibility because the athletic competition is not staged but a volatile and live production. Oftentimes, other celebrities build their credibility through pre-existing conditions such as rehearsed and written speeches, digitally recorded music, and scripted shows or movies. This is why it is worthwhile to examine how unstable/stable performance affects the credibility of an athlete endorser.

Consequently, in the current study, performance has been examined as to its influence on source credibility, especially the source trustworthiness and expertise. Essentially, the better an athlete performs, the stronger the consumers' perception of that athlete's trustworthiness and expertise becomes. It is speculated that poor athletic performances by an athlete endorser will signal a negative shift in consumers' perceptions of this individual's level of expertise, thereby detrimentally affecting his or her overall source credibility. In addition, successful athletic performances signified by winning or consistent successful finishes fabricating overall source credibility would foster the endorser's effectiveness to positively influence consumers' brand attitude, attitude toward the advertisement, and purchase intentions. This study seeks to add to existing research by examining the influence athletic performance has on source credibility as well as its impact on consumers' brand attitude, attitude toward the advertisement, and purchase intentions. The following research questions (RQs) were scrutinized.

RQ1: Does athletic performance cause a significant difference in each element of source credibility?

RQ2: Does athletic performance have a significant influence on source credibility?

RQ3: What are the causal relationships among consumers' brand attitude, attitude toward the advertisement, and purchase intentions associated with source credibility?

Method

Research Design

A between-group experimental design was chosen to assess differences in the source credibility based on an athletic endorser's on-field performance. This design involved two treatments, and each participant was randomly assigned to only one treatment: either a good or bad performance scenario about a sport athlete (e.g., endorser). Prior to the experiment, a pretest was conducted to create the profile of a fictitious endorser, stimuli (e.g., good and bad on-field performance scenarios), and a fictitious product for the advertisement. Students involved in the pretest did not participate in the main experiment.

Creation of a credible sport athlete. In the line of source credibility and advertising research, the use of a fictitious person as an endorser minimizes prior exposure to and perceptions about him or her because "with the use of familiar endorsers, such as well-known celebrities, there can be a significant amount of variation in subjects' knowledge and attitude toward that familiar individual" (Till & Busler, 2000, p. 4). In addition, it was a necessary condition that a selected fictitious endorser should be a credible source in terms of attractiveness, trustworthiness, and expertise to solely examine overall source credibility of the endorser. Thus, based on these premises the pretests were intended to formulate a fictitious endorser as follows.

First, the endorser should be attractive to the respondents in order to manipulate source attractiveness. Ten headshot pictures of women found to be attractive were chosen by the researchers. The pictures were laid out on a single piece of paper and fifty undergraduate students were asked to indicate their level of agreement on this person's being an attractive female golfer. The respondents indicated their view on a 5-point Likert type scale ranging from "strongly disagree" to "strongly agree." The picture with the highest mean (M = 3.92, SD =.80) was chosen as the picture of the fictitious endorser for the experiment. After choosing the fictitious endorser, the next step was to select a name for the fictitious endorser. The researchers also created ten fictional names (first and last names) and then all the names were laid out on a single piece of paper underneath the picture of the person. Fifty undergraduate students were asked to indicate their level of agreement on this name being an appropriate fit for the picture above. The students indicated their view on a 5-point Likert type scale ranging from "strongly disagree" to "strongly agree." The name with the highest mean (M = 3.50, SD = 1.24) was chosen as the name of the fictitious endorser. From the results of the above pretests, LPGA golfer Morgan Mitchell was created as the sport celebrity for the experiment, but participants were not aware of that fact.

Second, in order to manipulate trustworthiness and expertise, a biographical sketch of the created endorser, Morgan Mitchell, was developed. Since Giffin (1967) considered "favorable disposition" and "perceived supportive climate" as favorable consequences of trust, the following statement was used to manipulate trustworthiness of Morgan Mitchell.

"Mitchell is the co-founder of an organization designed to raise awareness and donations for breast cancer research. An annual golf tournament is held to raise money for this cause."

Also, as expertise was referred to as "qualification" (Berlo, Lemert, & Mertz, 1969) in this study, LPGA membership and years as a LPGA professional golfer were used to manipulate expertise of the created endorser. As a result, the fictitious endorser used for the experiment was considered to possess all elements of source credibility addressed by Ohanian (1990, 1991).

Stimuli. Two different fictional scenarios were developed to manipulate an athletic endorser's on-field performance, since this study was designed to examine the differences in source credibility based on an athletic endorser's recent performance. Prior research has shown the use of fictional scenarios to be effective. For example, Till and Shimp (1998) presented negative information about a fictional cyclist to measure brand evaluations after negative information had been released about an athlete. Priester and Petty (2003) used a fictional scenario to present contrasting information about a fictitious endorser in order to manipulate endorser trustworthiness. As a result, using a fictional scenario for the fictitious endorser is based on similar reasoning that differences will arise with two different scenario types (negative and positive).

A creation of brand. The use of a fictitious brand name with a predetermined product category could control preexisting cognitive and affective reactions caused by prior exposure and experience (Till & Shimp, 1998). This study used a fictitious brand for a product category selected through a two-step pretest as follows.

First, the product category screening test was used to select a popular product category and was administered to fifty undergraduate students in order to learn about students' interest in product categories. In the product category screening test, the respondents (N = 50) were instructed to write down the top three product categories based on their interest and involvement. The product category with the highest frequency was cellular phones (16%). Therefore, the endorsed product used for the experiment was a cellular phone.

Second, after selecting the product categories, the next step was given to find a name for the product category. The researchers chose a picture of a cellular phone, which appeared to be a generic flip phone without any brand name on it, and created 10 fictional brand names. All the names were laid out on a single piece of paper underneath the picture of the cellular phone, and fifty undergraduate students were asked to indicate their level of agreement for the names being used as the brand name for the picture of the cellular phone. They responded on a 5-point Likert-type scale ranging from "strongly disagree" to "strongly agree." The brand name with the highest mean score, Axon Max, was chosen for the experiment (M = 3.24, SD = 1.26).

Participants

The experiment was administered to 208 undergraduate students enrolled in sport management and communication courses at a large, public university in the southeastern region of the United States. Although the use of a convenience sample of student participants is certainly limited with regard to the issue of external validity, a student was deemed acceptable for the study because students are representative of the target market for the product category and are familiar with the product (Ferber, 1977). Of the total 208 participants in this study, the majority (N = 140, 67.3%) were males with 32.7% females (N =68), and participants had a mean age of 21.87 (SD = 3.34). Sample sizes of 200 or more have been considered acceptable for use of Structure Equation Model (SEM) (Garver & Mentzer, 1999; Holeter, 1983).

Data Collection Procedure

The experiment given to participants was to respond to a series of questions (i.e., a total of 18 questions) during and after viewing a packet of information. For example, the respondents first viewed a brief biography of the fictitious endorser, accompanied by a picture, and then viewed a fictional news article outlying a positive or negative performance scenario. Source credibility and perceived on-field performance of the endorser were then evaluated using 7-point semantic differential scales. An advertisement featuring a picture of the cellular phone with its product name and an endorsement of the product from the fictitious endorser also continued to be viewed in the next page. After viewing the advertisement, the respondents were then asked to evaluate brand attitude, attitude toward the advertisement, and purchase intentions of the endorsed product via answering 7-point semantic differential scales and 7-point Likert scales.

Measures

All latent constructs included in the study were measured using multi-item scales. The initial reliability of each latent construct ranged from .702 for source credibility to .942 for brand attitude. Therefore, all scales were found to be internally consistent as reliability coefficients exceeded the .70 threshold suggested by Nunnally (1978).

First, source credibility was examined by using three 7-point semantic differential scales anchored by unattractive/attractive, not an expert/expert, and untrustworthy/trustworthy derived from the 9-item scale in Ferle and Choi (2005). In particular, Ferle and Choi examined a high-order measurement model and found that three components of source credibility developed by Ohanian (1990) could converge into one latent construct. Consequently, the current study followed the recommendations of Ferle and Choi and considered source credibility as an overall construct, featuring the three common characteristics (e.g., expertise, trustworthiness, and attractiveness) which previous research discussed.

Second, a 4-item measure conducted using 7-point semantic differential scales was developed to evaluate the fictitious endorser's on-field performance. For example, statements used to measure on-field performance after viewing the fictional news articles were "Morgan Mitchell's performance has been: (1) unreliable/reliable; (2) bad/good; (3) inconsistent/consistent; (4) undependable/dependable."

Third, brand attitude was examined by using three 7-point semantic differential scales anchored by unfavorable/favorable, bad/good and negative/positive used by McDaniel and Kinney (1996). Fourth, in order to measure attitude toward the advertisement, the 4-item scale developed by Lee (2000) was employed. These items were measured using 7-point Likert scales ranging from 1 as "strongly disagree to 7 as "strongly agree." For example, statements were: (1) I like the advertisement that I saw; (2) the advertisement that I saw is attractive to me; (3) the advertisement that I saw is appealing to me; (4) the advertisement that I saw is interesting to me.

Finally, a 3-item measure conducted using 7-point semantic differential scales was developed to evaluate purchase intentions of the fictitious brand/product. For example, statements used to measure purchase intentions were "The next time I consider purchasing a cellular phone, I will consider Axon Max: (1) impossible/possible; (2) unlikely/likely; (3) improbable/probable."

Data Analysis

The analysis of data from the experiment was performed using the SPSS 19.0 and EQS 6.1 programs. Independent sample t-tests and multivariate analysis of variance (MANOVA) were used to examine differences in perceived on-field performance and overall source credibility, as well as differences in each element of source credibility. For this analysis, two different performance scenarios (e.g., experiments) were considered as the independent variable, while perceived performance and source credibility were considered as the dependent variables. In addition, a 2-stage modeling strategy, examining the measurement model and the structural equation model (SEM), recommended by Anderson and Gerbing (1988) was employed to evaluate the model fit as well as the direct and indirect relationships among the hypothesized latent constructs.

Results

Manipulation Checks

The manipulation checks specified whether the experimental manipulations worked or not, as the current study employed a between-group experimental design to examine differences in the source credibility based on an athletic endorser's performance. Levene's test was used to assess the assumption of variance homogeneity in the groups exposed to two different scenarios and resulted in failure to reject decisions, indicating that the variances were equal over the groups: F(1, 206) = .019, p = .890.

Findings from an independent sample t-test (t = 17.93, p < .000) revealed that participants exposed to the positive performance scenario (PPS) had a higher perceived value (m = 1.78) of an athletic endorser's on-field performance as compared to the value of students (m = -1.21) exposed to the negative performance scenario (NPS). These findings supported the further use of manipulations to examine differences in source credibility.

Differences in Source Credibility

In order to examine differences in the source credibility of the endorser, participants were assigned to either PPS or NPS groups. Levene's test revealed that the assumption of variance homogeneity in the two scenario groups resulted in failure to reject decisions, indicating that the variances were equal over the groups: F(1, 206) = .136, p = .713. Findings from the t-test indicated that students exposed to the PPS reported a higher value ( m = 1.22) on overall source credibility as compared to the value reported by students (m = .61) exposed to the NPS.

MANOVA was also used to examine differences in respective elements of source credibility. The test of the assumption of homogeneity of covariance matrices in the two scenario groups resulted in a reject decision: Box's M = 19.59, F(6, 307460.8) = 3.21, p = .004, indicating a likely violation of the assumption. However, a follow-up analysis with Levene's test for respective elements of source credibility resulted in failure to reject decisions for all elements, indicating that the variances were equal across the two scenario groups: expertise, F(1, 206) = 3.09, p = .080 ; attractiveness, F(1, 206) = 1.00, p = .318; trustworthiness, F(1, 206) = .081, p = .776.

Results of the MANOVA indicated that the equality of the means over the PPS and NPS groups as to the elements of source credibility was rejected at the .05 level: Wilk's [LAMBDA] = .824, F (3, 204) = 3.33, p < .000. Univariate F-tests provided additional support, indicating the differences in expertise and trustworthiness were statistically significant between positive and negative scenario groups: F(1, 206) = 31.92, p < .000; F(1, 206) = 15.56, p < .000, while a non-statistical difference in attractiveness was found between those two groups: F(1, 206) = .163, p = .687.

[FIGURE 1 OMITTED]

Tests of the Hypothesized Relationships

A 2-stage structural equation modeling strategy recommended by Anderson and Gerbing (1988) was employed to examine the effects of perceived on-field performance on source credibility and to investigate the relationship between source credibility and other related properties such as brand attitude, attitude toward the advertisement, and purchase intentions. In particular, tests of all measurement and structural models were based on the covariance matrix and used maximum likelihood estimation as implemented in EQS 6.1 (Bentler & Hu, 2005).

First, the values of selected fit indices indicated a favorable model fit for the initial measurement model while the results of chi-square estimated that the hypothesis of exact fit was rejected ([chi square] (109) = 214.33, p < .000). For example, the Standardized Root Mean Squared Residual (SRMR), one of the absolute fit indices, was .05; the Root Mean Square Error of Approximation (RMSEA), one of the parsimonious fit indices, was .06; the Comparative Fit Index (CFI), one of the incremental fit indices, was .964. Although the results of the LM test recommended model modifications, providing a slightly better fit to the data, the recommended modifications were not theoretically justifiable. As a result, the initial measurement model was deemed acceptable for the further use of the final measurement model as part of an SEM hypothesizing causal paths among latent constructs.

Second, the SEM provided a good fit to the data, [chi square] (114) = 223.256, p < .001, SRMR = .063, RMSEA = .068, CFI = .962, and all estimated parameters were statistically significant. The results of the LM test did not recommend any model modifications which would provide a slightly better fit to the data. Therefore, no further consideration was given to the inclusion of additional paths. Decomposition of effects derived from the SEM indicated that perceived on-field performance had a significant influence on source credibility (t = 7.70, p < .01), which explained approximately 44% of the variance in source credibility. In particular, this path had the strongest relationship, where a one standard deviation increase in perceived on-field performance led to a .662 standard deviation increase in source credibility, holding all else constant. Also, brand attitude was predicted by source credibility (t = 5.27, p < .01) while brand attitude had a significant influence on attitude toward the advertisement (t = 6.23, p < .01). The model explained approximately 20 % of the variance in brand attitude and 24% of the variance in attitude toward the advertisement.

Finally, brand attitude (t = 2.19, p < .01) and attitude toward the advertisement (t = 3.31, p < .01) were found to be significant determinants of purchase intentions, respectively. Approximately 15% of the variance for purchase intentions was explained by brand attitude and attitude toward the advertisement. The standardized parameter estimates for measurement and structure components were presented in Figure 1.

Discussion and Conclusions

Many advertisers and marketers have spent a substantial amount of money for athlete endorsement, as various benefits emerge from using athlete endorsers in product or brand advertising (Ferle & Choi, 2005). However, little is known about the impact of an athlete's on-field performance on source credibility. Accordingly, the primary purpose of this study was to gain a better understanding of the impact of athletic performance on source credibility that facilitates the relationship between the associated attitudes (e.g., brand, advertisements) and purchase intentions. Results of the study have provided several marketing implications for advertisers to consider and researchers to pursue.

First, the results of the present study identified differences in the elements of source credibility based on an athlete endorser's on-field performance. In particular, the perceived on-field performance was found to have a significant influence on source trustworthiness and expertise, while a non-statistical difference was found in source attractiveness. These findings imply that overall source credibility of an athlete endorser could be affected by his or her on-field athletic performance. Sport marketers or advertisers make strategic decisions as to how to maintain their relationships with sport celebrities. This study emphasizes the importance of recognizing the impact of current athlete performance in advertising and on endorsed brands.

Moreover, the decomposition of the effects derived from the SEM revealed the positive relationship between perceived on-field performance and source credibility. This suggests that when a consumer perceives positive information about an athlete endorser's on-field performance, he or she is likely to consider the athlete endorser as a more credible source. However, if that individual is exposed to negative on-field performance, his or her perception of the endorser's credibility is likely to decrease. Findings are consistent with the notion derived from Wright's (1973) CSM and Fishbein's attitude theory (Fishbein & Ajzen, 1975). They believed that one's attitude is generated from his or her salient beliefs. In this context, source credibility formed by on-field performance (e.g., cognition) could function as salient beliefs which affect subsequent attitude change.

In relation to the formation of beliefs, another considerable implication is that on-field performance might be closely associated with off-field issues. For example, Tiger Woods' highly publicized off-field issues appear to have negatively impacted his on-field performance. This in turn has affected his source credibility. However, what if Tiger Woods had performed better in tournaments following his difficulties off of the course? His diminished credibility might have recovered more rapidly with a good/satisfactory performance on the course, resulting in renewed endorsement contracts. Findings from the current study indicated that an athlete's on-field performance improves a consumer's belief in their endorsement-related expertise and trustworthiness.

Second, the results from the decomposition of the effects indicated that source credibility had a positive influence on brand attitude, while the effect of source credibility on attitude toward the advertisement was mediated by brand attitude. Findings were consistent with the results from previous research (e.g., Anderson, 1976; Collins & Loftus, 1975; Rumelhart, Hinton, & McClelland, 1986) demonstrating that consumers' beliefs toward an endorser are expected to transfer to the endorsed brand based upon their cognitive association. In other words, as a person's evaluation of an endorser has an opportunity to link to an endorsed brand, poor on-field performance can deteriorate evaluation of the endorser, which will directly influence a consumer's attitude toward the associated brand (Till & Shimp, 1998). Our findings clearly indicate that source credibility has a positive effect on brand attitude.

Till and Shimp (1998) also argued that the association between beliefs and attitudes in the context of advertisements is relatively stronger when new or unfamiliar brands have the association with a celebrity endorser who is essentially the primary cause for evaluation. Since the current study was designed to manipulate the effects of a fictional endorser rather than those of an advertisement, source credibility was directly related to brand attitude while attitude toward the advertisement was indirectly influenced by source credibility via brand attitude. However, considering that there are also direct effects of attitude toward the advertisement on brand attitude, the relationships between brand attitude and attitude toward the advertisement may have more variability than can be explained by a single hierarchy like the previous studies (Ferle & Choi, 2005; Gardner 1985; Mitchell & Olson, 1981; Shimp, 1981). Therefore, it seems reasonable to expect that the relationship between brand attitude and attitude toward the advertisement may be diverse in accordance with the levels of endorser equity and brand equity in advertisements, as well as the effects of advertisements.

Finally, brand attitude and attitude toward the advertisement were found to be significant determinants of purchase intentions. Findings were consistent with the theoretical perspectives suggested by Deogun and Beatty (1998) and Mitchell and Olson (1981) indicating that the consumers' affective reactions (i.e., attitudes) lead to purchase intentions. In the field of sport marketing and communication, consumers' affective reactions to endorsement activities have important implications in terms of measuring the effectiveness of a particular communication message. This is because the relationship between beliefs and consumers' affective reactions in the cognitive information process presents diagnostic information about the effectiveness of a message strategy (Mitchell & Olson, 1981). Thus, findings in support of a consistent pattern among beliefs, attitudes, and purchase intentions verify the notion that purchase intentions are a function of the consumers' affective reactions to endorsement activities that could have an effect on actual sales and consumption of endorsed-brand products.

In relation to a mediating effect of the consumers' affective reactions, the current study revealed that brand attitude has a direct, positive influence as well as an indirect influence on purchase intentions mediated by attitude toward the advertisement. In addition, a direct causal relationship was found between attitude toward the advertisement and purchase intentions. In particular, the direct causal relationship from brand attitude to purchase intentions was consistent with the authoritative direction of the extended Fishbein's attitude theory, while Mehta and Purvis (1997) explicitly revealed the direct causal link between attitude toward the advertisement and purchase intentions in their Adverting Response Model, which was also found to support our findings.

However, a number of studies have proposed that an affective construct representing consumers' favorable feelings toward the advertisement (i.e., attitude toward the advertisement) has a mediating effect on brand attitude and purchase intention (Ferle & Choi, 2005; Mitchell & Olson, 1981; Shimp, 1981), whereas the current study showed a mediating role of brand attitude in a path between attitude toward the advertisement and purchase intentions. The discrepancy might occur since the primary focus of the manipulation employed in this study was intended to examine differences in the source credibility based on an athlete endorser's performance rather than evaluating the effectiveness of an advertisement. In addition, an athlete endorser's credibility devolved by his or her athletic performance seems to have a direct association with brand attitude, which is then linked to attitude toward the advertisement and purchase intentions. Thus, it would be interesting to compare the effect of source credibility constituted by a particular type of stimuli (e.g., performance, scandal, campaign, etc.) on the causal sequence of consumers' affective and purchase reactions. It seems reasonable to hypothesize that the causal sequence could be different depending on the way athlete endorsers reduce and also attempt to restore their credibility.

In summary, athletes are considered more credible when they maintain their athletic performance. Consumer perception of athlete credibility is one of the critical factors mediating the relationships among perceived on-field performance, attitude toward the advertisement/brand, and purchase intentions. Since the issue of whether or not an athlete celebrity's on-field success affects the dimensions of source credibility is scarcely addressed in athlete endorsement literature, the present study will provide valuable information for sport marketers to consider when implementing athlete endorsement strategies. In particular, the current study allows sport marketers as well as advertisers to make knowledgeable decisions about who will endorse their product, and for how long they will endorse it. It is always necessary for sport marketers to realize the practical implications, due to their substantial financial commitment, regarding athlete endorsements and the athletes they choose to endorse their brand.

Limitation and Future Studies

This study provides important insights regarding how source credibility founded on perceived on-field performance mediates the relationship between consumers' affective reactions and purchase intentions. The results of the study should be interpreted within certain constraints.

The current study was confined to a convenient sample of undergraduate students at a major public university in the southeastern region of the United States. Therefore, the findings cannot be generalized to any specific populations or geographical areas. Future studies would gain external validity by employing probability samples of consumers.

Another constraint that should be considered is that the experiments utilized manipulated levels of an athlete endorser's on-field performance rather than measuring naturally occurring perceptions of his or her athletic performance. For instance, the study employed a between-group experimental design, and participants were asked to view one of the experiments including the fictitious endorser's on-field performance. These manipulations were artificially extreme in order to fabricate large differences in source credibility. Utilizing this, we can examine the effects of on-field performance on source credibility in a consumer's cognitive structure. However, this procedure could result in a lack of realism (Goldsmith, Lafferty, & Newell, 2000). Future studies could incorporate real athlete endorsers to gain the generalizability of the findings.

Ohanian (1991) found that male and female endorsers were perceived as being significantly different regarding their source credibility in the eyes of respondents, while a respondent's age and gender did not affect the evaluations of the endorser's source credibility. These findings were consistent with the study conducted by Boyd and Shank (2004). They indicated that although a respondent's gender had no effect on the three dimensions of source credibility, there was a significant interaction effect between a respondent's gender and an endorser's gender on the dimensions of source credibility. However, the current study only employed a female golfer as the athletic endorser to examine differences on her source credibility when exposed to negative or positive performance. Therefore, future research might employ a similar experimental design including a male athletic endorser to examine the effect of endorser's gender on the endorsement information process, as role differentiation between male and female athletes perceived by individuals might lead to variations of endorser effectiveness (Ferle & Choi, 2005).

One should also consider that featuring other types of sports and products would extend the scope of the findings. An athlete endorser may have a different association with a sponsoring brand due to the characteristics/preferences of that athlete's fans. For example, a professional golfer's endorsement of a luxury golf club brand will have more value than that of a NASCAR driver. Also, sport-related products endorsed by an athlete may reveal a stronger relationship between brand attitude and purchase intentions than those of non-sport related products. A concept related closely to this phenomenon is the match-up hypothesis. A successful match-up is one in which "... the highly relevant characteristics of the spokesperson are consistent with the highly relevant attributes of the brand" (Misra & Beatty, 1990, p. 161). Consumer reactions to endorsees (e.g., a brand, product, or service) are expected to be more positive if the endorser has a good match-up with them (Misra & Beatty, 1990). As sport marketers and advertisers are likely to choose an endorser according to their promotional and advertising emphasis, future studies would extend a similar research design to other sports settings as well as to different product categories, accompanied by focusing on the issue of the match-up hypothesis.

The finding that an athlete's on-field performance improves a consumer's belief in the athlete's endorsement-related expertise and trustworthiness might function as a partial solution for off-field issues in order to maintain the endorser's credibility. Future studies should consider whether athletes with positive on-field performance are able to maintain endorser credibility in the face of off-field issues better than those athletes who have low levels of on-field performance.

Finally, the current study was confined to "non-natural" exposure of the advertisement (Mitchell & Olson, 1981). The advertisement exposure used in this study was obligatory and timed, while frequent exposures to an advertisement might naturally affect changes in beliefs and attitude toward the advertisement. Future studies will be required to have rigorous controls in advertising stimuli (e.g., advertisement exposure intervals, the format of advertisement, etc.) in order to obtain natural advertisement effects. In addition, they should be performed to understand the time frame associated with negative performance and its effect on source credibility. For instance, if an athlete endorser has one year of bad performance, does that in turn result in negative consumer perceptions of the sponsoring brand? How much negative performance does it take to impact consumer-based brand equity?

The current study clearly demonstrates how an athlete endorser's on-field performance changes his or her overall credibility mediating brand attitude, attitude toward the advertisement, and purchase intentions. By examining the role of athletic performance in endorsement activities, sport marketers and advertising practitioners may be able to develop more effective communication strategies, as our findings justify the consideration of athlete endorser's on-field performance to evaluate source credibility as well as advertising effects.

References

Anderson, J. (1976). Language, memory and thought. Hillsdale, NJ: Erlbaum Associates.

Anderson, J. C., & Gerbing, D. W. (1988). Structural equation modeling in practice: A review and recommended two-step approach. Psychological Bulletin, 103(3), 411-423.

Atkin, C., & Block, M. (1983). Effectiveness of celebrity endorsers. Journal of Advertising Research, 23(1), 57-62.

Bentler, P. M., & Hu, E. J. C. (2005). EQS 6.1 for Windows user's guide. Multivariate Software, Inc., Encino, CA.

Berlo, D. K., Lemert, J. B., & Mertz, R. J. (1969). Dimensions for evaluating the acceptability of message sources. Public Opinion Quarterly, 33(4), 563-576.

Boyd, T., & Shank, M. (2004). Athletes as product endorsers: The effect of gender and product relatedness. Sport Marketing Quarterly, 13(2), 82-93.

Collins, A. M., & Loftus, E. F. (1975). A spreading activation theory of semantic processing. Psychological Review, 82, 407-428.

Deogun, N., & Beatty, S.G. (1998, March 19). Coke's marketing chief, Sergio Zyman, to quit. Wall Street Journal.

Ferber, R. (1977). Research by convenience. Journal of Consumer Research, 4, 57-58.

Ferle, C. L., & Choi, S. M. (2005). The importance of perceived endorser credibility in South Korean advertising. Journal of Current Issues and Research in Advertising, 27(2), 67-81.

Fishbein, M., & Ajzen, I. (1975), Belief, attitude. intention, and behavior. Reading, MA: Addison-Wesley.

Fiske, S. T. (1982). Schema-triggered affect: Applications to social perception. In Clark, M. S., & Fiske, S. T. (Eds.), Affect and Cognition: The Seventeenth Annual Carnegie Symposium on Cognition (pp. 55-78). Hillsdale, NJ: Lawrence Erlbaum Associates.

Gardner, M. P. (1985). Does attitude toward the ad affect brand attitudes under a brand evaluation set? Journal of Marketing Research, 22(2), 192-198.

Garver, M. S., & Mentzer, J. T. (1999). Logistics research methods: Employing structural equation modeling to test for construct validity. Journal of Business Logistics, 20(1), 33-57.

Goldsmith, R. E., Lafferty, B. A., & Newell, S. J. (2000). The impact of corporate credibility and celebrity credibility on consumer reaction to advertisements and brands. Journal of Advertising, 29, 43-54.

Giffin, K. (1967). The contribution of studies of source credibility to a theory of interpersonal trust in the communication process. Psychological Bulletin, 68(2), 104-120.

Hoelter, D. R. (1983). The analysis of covariance structures: Goodness-of-fit indices. Sociological Methods and Research, 11, 325-344.

Hovland, C. I., Janis, I. K., & Kelley, H. H. (1953) Communication and persuasion, New Haven, CT: Yale University Press.

IEG (2011). Sponsorship Spending: 2010 Proves Better Than Expected; Bigger Gains Set For 2011. Retrieved from http://sponsorship.com

Joseph, W. B. (1982). The credibility of physically attractive communicators: A review. Journal of Advertising, 11, 15-24.

Kelman, H. C. (1961). Processes of opinion change. Public Opinion Quarterly, 25, 57-78.

Lafferty, B. A., & Goldsmith, R. E. (1999). Corporate credibility's role in consumers' attitudes and purchase intentions when a high versus a low credibility endorser is used in the ad. Journal of Business Research, 44(2), 109-116.

Lafferty, B. A., Goldsmith, R. E. & Newell, S. J. (2002). The dual credibility model: The influence of corporate and endorser credibility on attitudes and purchase intentions. Journal of Marketing Theory and Practice, 10(3), 1-13.

Lavidge, R. C., & Steiner, G. A. (1961). A model for predictive measurements of advertising effectiveness. Journal of Marketing, 25, 59-62.

Lee, Y. H. (2000). Manipulating ad message involvement through information expectancy: Effects on attitude evaluation and confidence. Journal of Advertising, 29(2), 29-43.

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.

Macur, J. (2009, February 6). Phelps disciplined over marijuana pipe incident. New York Times. Retrieved from http://www.nytimes.com/2009/02/06/sports/othersports/06phelps.html

McDaniel, S. R., & Kinney, L. (1996). Ambush marketing revisited: An experimental study of perceived sponsorship effects on brand awareness, attitude toward the brand and purchase intention. Journal of Promotion Management, 3, 141-67.

McGinnies, E., & Ward, C. D. (1980). Better liked than right: Trustworthiness and expertise as factors in credibility. Personality and Social Psychology Bulletin, 6, 467-472.

Mehta, A., & Purvis, S. C. (1997). Evaluating advertising effectiveness through Advertising Response Modeling (ARM). In W. D. Wells (Ed), Measuring advertising effectiveness (pp. 325-334). Mahwah, NJ: Lawrence Erlbaum Associates.

Miciak, A. R., & Shanklin, W. L. (1994). Choosing celebrity endorsers. Marketing Management, 3(3), 50-59.

Misra, S., & Beatty, S. E. (1990). Celebrity spokesperson and brand congruence: An assessment of recall and affect. Journal of Business Research, 21 , 159-173.

Mitchell, A. A. & Olson, J. C. (1981). Are product attribute beliefs the only mediators of advertising effects on brand attitudes? Journal of Marketing Research, 18(3), 318-322.

Money, R. P., Shimp, T. A., & Sakano, T. (2006). Celebrity endorsements in Japan and the United States: Is negative information all that harmful? Journal of Advertising Research, 26(1), 113-123.

Nike, Inc. (2010). Annual Report on Form 10-K. Retrieved from http://www.nd.edu/~scorwin/fin70610/documents/Nike_10K_2009.pdf

Nunnally, J. C. (1978). Psychometric theory (2nd ed.). New York, NY: McGraw-Hill.

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 consumers' intention to purchase. Journal of Advertising Research, 31, 46-53.

Olson, J. C., Toy, D. R., & Dover. P. (1982). Do cognitive responses mediate the effects of advertising content on cognitive structure? Journal of Consumer Research, 9(3), 245-262.

Priester, J. R., & Petty, R. E. (2003). The influence of spokesperson trustworthiness on message elaboration, attitude strength, and advertising effectiveness. Journal of Consumer Psychology, 13, 408-421.

Rumelhart, D. E., Hinton, G. E., & McClelland, J. L. (1986). A framework for parallel distributed processing. Cambridge, MA: MIT Press.

Shimp, T. A., (1981). Attitude toward the ad as a mediator of consumer brand choice. Journal of Advertising, 10(2), 9-15.

Smith, R., & Swinyard, W. (1988). Cognitive response to advertising & trial: Belief strength, belief confidence, and product curiosity. Journal of Advertising, 17, 3-14.

Sternthal, B., Phillips, L. W., & Dholakia, R. R. (1978). The persuasive effect of source credibility: A situational analysis. Public Opinion Quarterly, 285-314.

Stevens, J., Lathrop, A., & Bradish, C. (2003). Who is your hero? Implications for athlete endorsement strategies. Sport Marketing Quarterly, 12(2), 103-110.

Story, L. (2005, July 18). 'Michelle mania' a marketer's dream. New York Times News Service. Retrieved from: http://www.findarticles.com/p/articles/mi_qn4188/is_20050718/ai_n14781038

Till, B. D., & Busler, M. (2000). The match-up hypothesis: Physical attraction, expertise, and the role of fit on brand attitude, purchase intent, and brand beliefs. Journal of Advertising, 29(3), 1-13.

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

Yoon, K., Kim, C., & Kim, M. (1998). A cross-cultural comparison of the effects of source credibility on attitudes and behavioral intentions. Mass Communication & Society, 1(3/4), 153-173.

Wright, P. L. (1973). Cognitive processes mediating acceptance of advertising. Journal of Marketing Research, 4, 53-62.

Gi-Yong Koo, MS, is a doctoral candidate at the University of Arkansas. His research interests include sport marketing and statistics, especially fan behavior, corporate sponsorship, and service marketing.

Brody J. Ruihley, PhD, is an assistant professor of sport administration at the University of Cincinnati. His research interests lie in the area of sport communication, specifically sport and media, fantasy sport, and public relations in sport.

Stephen W. Dittmore, PhD, is an assistant professor of recreation and sport management at the University of Arkansas. His research interests include public relations in sport, media rights in sport, and the U.S. Olympic movement.