International Journal of Sport Communication, 2015, 8,411 -430 http://dx.doi.org/10.1123/IJSC.2015-0043 © 2 0 1 5 H um an Kinetics, Inc.
Human Kinetics STUDENT RESEARCH
The Effect of World Ranking on the Selection of Athlete Endorsers: The Case of the PGA Wonseok (Eric) Jang,Yong Jae Ko, and Hee Youn Kim University of Florida, USA
Seung Hoon Jeong Woosuk University, South Korea The purpose of the current exploratory study was twofold: First, to outline cur rent trends in athlete endorsement in the golf industry, and second, to discuss specific patterns of athlete endorsement in practice by considering an athlete’s world ranking and product type (low vs. high involvement and informational vs. transformational products). The results indicate that firms in 23 different types of industries are currently using professional golfers as athlete endorsers to position their products in their target markets. Specifically, the results of correspondence analysis indicate that highly ranked golfers tend to endorse high-involved, expen sive, and informational products, while both highly ranked and lowly ranked golfers are similarly used as endorsers for low-involved, inexpensive, and transformational products. Implication, limitations, and future research directions are discussed. Keywords: athlete endorsement, expert endorser, golf industry, correspondence analysis H igh-profile athletes receive im m ense attention from m ass m edia. Therefore, in contem porary society, the pow er o f athletes to influence peo p le’s behavior has significantly increased. H igh-profile athletes affect peop le’s lives in several ways such as influencing purchasing behaviors (M artin & Bush, 2000) and exercise behaviors (Lines, 2001); som e athletes even becom e societal role m odels (M utter & Paw low ski, 2014). C om m unication studies in particular have found that the m edia play a key role in such effects. T hat is, the m assive exposure o f athletes in m edia m akes people feel m ore involved and engaged w ith high-profile athletes (Brown, Basil, & B ocarnea, 2003); therefore, people are m ore influenced by those athletes w hen they m ake decisions (M artin & Bush, 2000).
Jang, a doctoral student; Ko; and Kim are with the Dept, of Tourism, Recreation, and Sport Management, University of Florida, Gainesville, FL. Jeong is with the College of Sports Science, Woosuk University, Wanju-gun, South Korea. Address author correspondence to Yong Jae Ko at
[email protected].
411
412
J a n g e ta l.
Due to the enormous public interest in athletes, several firms have used highprofile athletes (instead of other celebrities) as a key marketing communication tool to positively promote their products and firms (Carlson & Donavan, 2008; Carroll, 2009). Marketing-communication research has generally shown that high-profile athletes enhance consumers’ evaluations of endorsed products and brands, thus leading to increased positive attitudes toward advertisements (Koernig & Boyd, 2009), brands (Boyd & Shank, 2004; Koernig & Boyd, 2009; Martin, 1996), and purchase intentions (Bush, Martin, & Bush, 2004). In general, the engagement of a celebrity endorser requires a huge financial outlay (Stone, Joseph, & Jones, 2003). Recent reports, for example, indicate that Nike spent almost $100 million for a 5-year contract with Rory Mcllroy and $10 million for a 1-year contract with Lebron James in 2014 (Crouse, 2014). Since firms are spending millions of dollars to hire high-profile athletes, the key issue in celebrity endorsement in marketing-communication research is providing an effective guideline for how to select the best possible endorsers to maximize effec tiveness. In this respect, a number of theoretical frameworks have been developed to examine endorsers’ effectiveness, such as the source-credibility model (Ohanian, 1990), the source-attractiveness model (McGuire, 1969), the meaning-transfer model (McCracken, 1989), and the match-up hypothesis (Boyd & Shank, 2004; Martin, 1996; McDaniel, 1999). Although these frameworks have used slightly different theoretical and methodological approaches, a key commonality exists— the importance of developing an appropriate match between the endorsed product and the characteristics of the endorser. Consistent with this notion, sport manage ment research has found that high-profile athletes are more effective endorsers for sports-related firms (e.g., Nike, Adidas, UnderArmour) than for others (e.g., BMW, State farm Insurance, HSBC bank), because their expertise in sports cre ates a better match with sports-related products (Boyd & Shank, 2004; Martin, 1996; McDaniel, 1999). Although there is a wealth of literature on the subject of athlete endorsement, a few key limitations still exist. First, most of earlier athlete-endorsement studies were conducted using an experimental setup (e.g., Boyd & Shank, 2004; Fink, Cunningham, & Kensicki, 2004; Hahn & Cummins, 2014; Yoon & Choi, 2005); so existing knowledge provides limited managerial implications in terms of cur rent trends in athlete-endorsement deals in practice. By assessing current trends, firms may be able to use these findings as an effective marketing-communication guideline to select the best possible endorsers. Second, very little is known about how product type influences firms’ hiring decisions of high-profile athletes as endorsers. In fact, prior endorsement literature has suggested that it is important to consider the product type to develop an appropriate match with the characteristics of the endorser (Biswas, Biswas, & Das, 2006; Lord & Putrevu, 2009). This aspect is important because high-profile athletes are not only endorsing sports-related products; they are also endorsing various categories of products such as watches, automobiles, and even financial products. Third, very little is known about how athletes’ world rankings affect firms’ hiring decisions of athletes as their endorsers. Indeed, scholars have pointed out that an athlete’s world ranking plays a crucial role in an endorsement context (e.g., Chung, Derdenger, & Srinivasan, 2013). To assist in filling these gaps, the purpose of the current exploratory study was twofold. First, this study outlines current trends in athlete endorsement in the IJ S C V o l. 8, N o . 4, 2015
Athlete Endorsement
413
golf industry. Second, this study discusses specific patterns of athlete endorsement in practice, by considering an athlete’s world ranking and product type (low vs. high involvement and informational vs. transformational products), to provide an effective marketing-communication guideline to firms.
Theoretical Background Consumer Trends in the Golf Industry More than 20 million people all over the world enjoy golf as a lifelong recreational sport. While the growth of the golf industry was remarkable just a decade ago, the industry has been experiencing a downward trend in tenns of numbers of participants and spectators. According to the National Golf Foundation (2015), the numbers of golfers were approximately 25 and 24.7 million in 2014 and 2013, respectively, down 17% from 30 million in 2010. In terms of golfers’ demographics, among those 25 million golfers in 2014,77.5% were male, their average household income was $930,000, and the highest percentage of core golfers came from the 30- to 39-year (18%) and 40- to 49-year age groups (17.6%; Stachura, 2015). A similar pattern can be found in golf as a major spectator sport. Golf televi sion viewership continues to tumble. In 2011, the PGA Tour’s cumulative audience reached 175.9 million from all TV networks. However, after that golden age, the number dropped to 171.2 million in 2012, 172.6 million in 2013, and 170.7 million in 2014 season (Smith, 2014), although the audience numbers have started to rise once more in 2015. A recent report indicated that the 2015 Masters final-round coverage on NBC had a 26% increase in audience (PGA Tour, 2015), and the audi ence for the final-round coverage of the PGA Tours’ Valspar Championship on NBC increased more than 30% compared with the 2014 season (Golf Channel, 2015).
Trends in Golf Endorsement The golf industry has received enormous attention from a number of major firms and corporations (Forbes, 2015). PGA events have received considerable attention as a key marketing platform from corporate partners. As a result, PGA tournaments remain fully sponsored for 2015 by the most prestigious brands (e.g., MercedesBenz, OMEGA, Samsung). Furthermore, Fortune 500 firms continue to use profes sional golfers as a key marketing-communication tool to promote their firms and product brands (Forbes, 2015; Rovell, 2015). Currently, professional golfers endorse a wide range of products (both sport and general products) including financial, automobile, and even medical services. For example, in 2014, Phil Mickelson had endorsement deals with multiple firms from various industries (e.g., luxury watch, Rolex; financial service, Barclays; educational service, ExxonMobil) and earned more than $48 million (Forbes, 2015). Why do professional golfers receive more endorsements than other profes sional athletes? First, a professional golfer is a human billboard. Firms are able to put their logos on professional golfers’ hats, shirts, and golf bags and even on their caddies. Therefore, media exposure of firms’ logos is dramatically increased. When a firm sponsors a team, however, the sponsor’s logo gets little direct exposure because of the team’s prominent logo. Second, firms’ media exposure is greater IJ S C V o l. 8, No. 4, 2015
414
J a n g e ta l.
when using professional golfers because PGA Tour golf-tournament broadcasts are longer (e.g., 5 hr/day) and more frequent (four times a week) than most other professional sports (e.g., National Basketball Association). Third, golf fans are generally wealthier than other sports fans (CNN, 2015). According to the National Golf Foundation’s report, golfers’ average household income was $93,000, which is almost twice the average household income of Americans ($53,891; Stachura, 2015). Therefore, firms can effectively reach affluent potential consumers by using professional golfers (Kennedy, 2014). Fourth, firms can effectively enhance aware ness in international markets because professional golfers travel around the world to compete, and major golf tournaments are broadcast in numerous other countries (e.g., China and Korea; Golf’s 2020 Vision). In the following section, we use several theoretical frameworks to address how athlete endorsements can benefit firms in promoting their brands.
Criteria for the Selection of an Athlete Endorser There is a wealth of literature showing that high-profile athletes provide several benefits to firms. First, high-profile athletes (e.g., higher ranked athletes) receive massive media exposures (Jensen, 2012). Second, high-profile celebrities (athletes) grasp and enhance consumers’ attention for endorsed products and brands (Atkin & Block, 1983). Third, when high-profile athletes perform substantially well by using the endorsed products, this implies that the products have great quality; therefore, it benefits athletes to perform better (Boyd & Shank, 2004). Finally, high-profile athletes enhance consumers’ evaluations of endorsed products and brands, thus leading to increased positive attitudes toward advertisements (Koernig & Boyd, 2009), brands (Boyd & Shank, 2004; Koernig & Boyd, 2009; Martin, 1996), and purchase intentions (Bush et al., 2004). In academia, a number of theoretical frameworks have been developed to examine endorsers’ effectiveness. In particular, the source-credibility model (SCM; Ohanian, 1990) and the source-attractiveness model (SAM; McGuire, 1969) have focused on the characteristics of celebrity endorsers to measure endorsement effectiveness. The SCM and the SAM suggest that the effectiveness of celebrity endorsements depends on the specific characteristics of the endorsers. The SCM identifies trustworthiness, expertise, and physical attractiveness (Kahle & Homer, 1985; Ohanian, 1990) of a celebrity as characteristics of an effective spokesperson. According to SAM, endorsers embody three dimensions— familiarity, similarity, and liking (McGuire, 1969). Consumers perceive celebrity endorsers as more persuasive and attractive when they are well known, similar to themselves, and well liked (Biswas et al., 2006). The idea is that those specific characteristics of endorsers significantly influence the persuasiveness of the advertising message and subsequently enhance consumers’ attitudes and beliefs about the products (Amos, Holmes, & Strutton, 2008; Lord & Putrevu, 2009).
World Ranking and Athlete Expertise While researchers have identified several important characteristics of endors ers, expertise was found to be the most influential characteristic that determines athletes’ endorsement effectiveness over other characteristics (e.g., attractiveness IJSC Vol. 8, No. 4, 2015
Athlete Endorsement
415
and likeability; Boyd & Shank, 2004; Fink et al., 2004). Fink et al., for instance, found that an athlete’s expertise is a stronger determinant of consumers’ percep tions of an athlete-event fit than is attractiveness. Furthermore, athletes were more effective spokespersons for sports-related products (e.g., energy bars) than actors (Till & Busier, 1998). Similarly, athlete endorsers were more effective in promoting sports-related products than were nonathlete endorsers (Koernig & Boyd, 2009). Expertise accumulates “from an actor’s ability to provide information to others because of his [or her] experience, education, or competence” (Horai, Naccari, & Fatoullah, 1974, p. 601). In some ways, celebrity endorsers might be viewed as experts in their particular fields, due to their higher ability to provide information based on their successful experiences, skills, and knowledge (Kahle & Homer, 1985; Ohanian, 1990; Biswas et al., 2006). Therefore, endorsers increase the credibility of advertising messages when they endorse a product related to their own field (Koo, Ruihley, & Dittmore, 2012) and ultimately influence consumers’ perceptions of endorsed products and brands (Eisend & Langner, 2010). For example, in the sports setting, athletes are considered experts when they endorse sports-related products rather than general (nonsports) products, because of their successful experiences and accomplishments in their own fields. What, then, is the most effective and convenient method for firms to assess athletes’ expertise levels? One simple way is assessing world ranking. Although several valid psychometric scales were developed to measure an athlete’s exper tise (e.g., DeSarbo & Harshman, 1985; Ohanian, 1990; Simpson & Kahler, 1981), there are few pieces of evidence in the literature that show that world rankings can also indicate athlete expertise level. First, consumers usually determine endorsers’ expertise levels based on the degree of their success, skills, and knowledge in their particular fields (Ohanian, 1990; Biswas et al., 2006). Similarly, an athlete’s world ranking is determined by how he or she has performed in tournaments for a given period of time (OWGR, 2015). Second, Koo et al. (2012) found that an athlete’s on-field performance significantly influences their trustworthiness and expertise. Third, higher ranked professional athletes garner more media exposure than lower ranked professional athletes (Jensen, 2012). This aspect is also important in determining the perception of an athlete’s expertise, because endorsers’ familiarity influences their levels of expertise (Buhr, Simpson, & Pryor, 1987). In this respect, the current study used world rankings to assess athletes’ expertise levels. In sum, numerous athlete-endorsement studies have found that athletes’ expertise has a positive impact on consumers’ evaluation of endorsed products and brands (Boyd & Shank, 2004; Koernig & Boyd, 2009; Martin, 1996). Expert endorsers increase the credibility of advertising messages significantly more than nonexpert endorsers (Biswas et al., 2006; Koo et al., 2012). Furthermore, Jensen (2012) found that during a golf tournament (e.g., PGA Tour), higher ranked golf ers were more likely to be exposed in the media than lower ranked golfers. In this respect, it is not surprising that many firms hire higher ranked professional athletes as communication spokespeople to better transfer their successful experiences, accomplishments, and credibility to advertising messages, endorsed products, and brands (Biswas et al., 2006). Accordingly, the first hypothesis was formulated as follows: IJ S C V o l. 8, N o. 4, 2015
416
Jangetal.
HI: Firms are more likely to make endorsement deals with higher ranked professional athletes than with lower ranked professional athletes
World Ranking and Product Category To provide an effective marketing-communication guideline, it is important to consider how firms are hiring high-profile athletes as endorsers based on product type. According to the Rossiter-Percy grid, product type can be classified based on motivation (informational vs. transformational products) and involvement (high- vs. low-involvement products; Rossiter & Percy, 1987; Rossiter, Percy, & Donavan, 1991). The Rossiter-Percy grid suggests that product-purchase motivation acts as a significant agent in activating consumption behavior. Specifically, consumers purchase informational products to fulfill their cognitive beliefs and functional motivations (Rossiter & Percy, 1987). Meanwhile, consumers purchase transfor mational products to enhance their self-images and achieve positive emotional gratification (Rossiter et al., 1991). Prior literature has found that the effectiveness of the endorser increases when the characteristics of the endorser match with the underlying motivation of product consumption (Stafford, Stafford, & Day, 2002; Till & Busier, 1998). In this view, informational products more effectively communicate with consumers via the characteristics of endorser expertise compared with other endorser characteristics (i.e., attractiveness), because expert endorsers’ knowledge, skills, and successful experiences help resolve consumption-related problems for informational products (Lord & Putrevu, 2009). For example, Stafford et al. (2002) found that expert endorsers were more effective in creating trust and belief in advertising messages for informational services (e.g., banks). Likewise, consumers’ perceived risk of purchasing informational products (e.g., microwave ovens, aspirin) was diminished through expert endorsers’ recommendations; therefore, expert endorsers increased consumers’ purchase intentions for informational products more than for transfor mational products (Lord & Putrevu, 2009). Meanwhile, transformational products are more effectively communicated by physically attractive or socially likeable endorsers (Kamins, 1990; Lord & Putrevu, 2009). For example, Kamins found that attractive celebrity endorsers were more effective in endorsing products used to enhance consumers’ attractiveness. According to the Rossiter-Percy grid, involvement is closely related with “a risk perceived by the typical target audience” (Rossiter et al., 1991, p. 14). In gen eral, consumer product consumption is heavily determined by level of involvement toward the product (Biswas et al., 2006; Homer & Kahle, 1990; Lord & Putrevu, 2009). Specifically, consumers perceive a higher level of risk when they make a purchase decision for high-involvement products (e.g., expensive, high-technology) than low-involvement products (Kelman, 1961). In this respect, expert endorsers are more effective in representing high-involvement products than low-involvement products because their credibility may have a stronger effect in reducing perceived financial and performance-related risks (Biswas et al., 2006; Petty & Cacioppo, 1983). Building on previous literature, the current study proposed that firms are more likely to hire higher ranked professional athletes as their endorsers for infor mational and high-involvement products than lower ranked athletes. Therefore, we developed the following: IJSCVol. 8, No. 4, 2015
Athlete Endorsement
417
H2: Higher ranked professional athletes are used more as endorsers for infor mational products than transformational products.
H3: Higher ranked professional athletes are used more as endorsers for highinvolvement products (i.e., expensive and highly technology-oriented products) than low-involvement products (i.e., inexpensive and less technology-oriented products).
RQ1: Is there any relationship between the athletes’ world rankings and types of endorsed products?
Method Sample Professional golfers were chosen for this study. There are several rationales behind this decision. First, professional golfers endorse a wider range of products than other professional athletes. Perhaps the most popular examples are Tiger Woods and Phil Mickelson. In 2014, Woods and Mickelson earned roughly $55 million and $48 million and ranked sixth and eighth, respectively, among athlete endors ers (Forbes, 2015). Second, we could clearly identify actual endorsement deals via professional golfers’ official Web sites. Thus, the sample pool was created based on the list of professional golfers who competed in PGA of America and European Golf Association tours in 2011. Included in the sample pool were 557 professional golfers.
Coding Procedures To investigate the specification pattern of professional golfers’ endorsement deals and their world rankings, it was necessary to identify endorsement deals for prolessional golfers. Only professional golfers who provide their endorsement deals on official Web sites in 2011 were included in the sample pool. As a result, 288 (51.7%) of professional golfers out of 557 were included in the final sample. On the list, we found 1,748 endorsement deals in various industries (e.g., financial, hotel, luxury brands). The world ranking as of Week 35 of the 2011 season (ending August 28, 2011) was used to group professional golfers into five categories based on their world rankings. World ranking is determined by the International Federation of PGA Tours based on golfers’ most recent 2-year performance in the PGA, European PGA, Japan PGA, South Africa PGA, OneAsia, and Australia Tours. By using a median-split technique, we categorized professional golfers into higher ranked professional golfers (Group 1 [Gl] and Group 2 [G2]), middleranked professional golfers (Group 3 [G3]), and lower ranked professional golf ers (Group 4 [G4] and Group 5 G5]). Specifically, Gl consisted of 58 (world ranking 1-76), G2 consisted of 58 (world ranking 77-185), G3 consisted of 58 (world ranking 186-370), G4 consisted of 58 (world ranking 373-631), and G5 consisted of 56 (world ranking 632-652) professional golfers. Although each
IJ S C V o l.8 , No. 4, 2015
418
J a n g e ta l.
group has different ranges of world rankings, it was an essential procedure to include similar numbers of professional golfers for each group. Following sug gestions from previous literature, the middle point (G3) was excluded to clearly identify two groups (e.g., Bernichon, Cook, & Brown, 2003; Preacher, Rucker, MacCallum, & Nicewander, 2005). For product classification, we used two dimensions of the Rossiter-Percy grid (Rossiter et al., 1991) developed based on product-purchase motivation (information vs. transformation products) and involvement (high- vs. lowinvolvement products) approaches. By following Davis’s (1997) guidelines, two independent coders who were unaware of the purposes of the study were recruited and trained for data analysis. Before coding the actual data set, they were first trained with a few examples not included in the main data analysis. During the training process, inconsistencies in their coding were identified and subsequently they received additional training until they reached a sufficient level of familiarity. In the main data analysis, the coders were asked to classify 1,748 endorse ment contracts to different product segments. The average intercoder reliability was .87, which is in the acceptable range (Perreault & Leigh, 1989). Furthermore, a focus-group interview (n = 5) was conducted to classify products (e.g., luxury and financial) into each dimension that were not originally placed in the RossiterPercy grid; in addition, nonprofit organizations, charities, and department stores were excluded from the data analysis, because those industries can be placed in all categories (examples are displayed in Table 1).
Data Analysis Chi-square (HI, H2, and H3) and correspondence analysis (RQ1) were employed for the data analysis. Specifically, cross-tabulation was computed to test the chisquare value for HI, H2, and H3. In addition, a correspondence analysis was employed to examine the specific relationship between professional golfers’ expertise level and types of endorsed products (RQ). Correspondence analysis is an interdependent and exploratory data-reduction method designed to analyze multiple categorical data in a joint-space map (Greenacre, 1984; Hoffman & Franke, 1986). The results of the analysis can be displayed graphically in a low dimensional map where similar objects and attributes are plotted close together and different objects and attributes (described by rows and columns as points) are plotted relatively far away from one another (Ferreira, Hall, & Bennett, 2008). To determine the nature of the relationships (or correspondence) between the objects, this graphical depiction of both rows and columns in the same map is an especially useful tool to discover information not revealed by multiple pairwise comparisons (Bendixen, 1996; Hair, Anderson, Tatham, & Black, 1998; Weller & Romney, 1990). However, Weller and Romney argued that it is important for chi-square values to be significant if correspondence analysis is to be employed. Otherwise, the analysis cannot capture significant associations. The chi-square statistic (x2= 151.46, p < .000) was significant, and a singular value for Dimension 1 also met the .20 criterion to establish a significant relationship with meaningful dimensions (Hair et al., 1998).
IJ S C V o l.8 , No. 4, 2015
Athlete Endorsement
419
Table 1 Selected Companies and Industry Segments That Use Golfer Endorsements Type of industry
Exam ples
Financial
O ld M utual, N evada State Bank
C onsulting
K PM G , D eloitte
E nergy
Schoco, A ir E nergi
H igh-tech product
Sharp, E picor
C onstruction
Bennett C onstruction, K N AUF
L ogistic
U PS, D H L
S port-related product
Sport practice equipm ent, sport drink
G olf-related product
G o lf practice equipm ent, go lf grip
O rganization or charity
C om er Stone, O neida India N ation
D epartm ent store
V ery.co.uk, Belk
M edical
G reenw ay M edical, O ptical E xpress
G o lf apparel
C utter & Bucks, A shw orth
L uxury brand
O m ega, R olex, H ugo Boss
A utom obile
Kia, T oyota
C ar repair
C ar S ervice A tlas, Boch N e C oppes
H otel or resort
H yatt, Palm R esort
R estaurant
C afe de C olum bia, Tim H ortons
A lcoholic beverage
Johnnie W alker, G ray G oose Vodka
A irline
U nited A irlines, Korean A irlines
V ideo gam e
EA Sport
M agazine
G o lf D igest, The Scotsm an
N ew spaper
Focus, all over press
Sunglasses
T ransitions, Sundog
Results Effects of Athlete World Ranking on Endorsement Deals The primary purpose of HI was to examine the relationship between athletes’ world rankings and the number of athlete-endorsement contracts in practice. Table 2 dis plays and compares the frequency of endorsement contracts for each ranking group of professional golfers. As shown in Table 2, higher ranked professional golfers (G1 and G2; 48.5%, n = 848) had significantly more endorsement contracts than lower ranked professional golfers did (G4 and G5; 20.3%, n = 548). Specifically, G1 had the highest proportion of endorsement contracts at 27.9% (n = 488), regardless of type of industry. The second-highest group was G2, accounting for 20.6% (n =
IJSCVol. 8, No. 4,2 015
420
Jang et al.
Table 2
Frequency of Golf Endorsement Frequency
Percent
G1
488
27.9
G2
360
20.6
G3
352
20.1
G4
308
17.6
G5
240
13.7
Total
1,748
100
Rank
Note. G1 = World Ranking 1-76; G2 = 77-185; G3 = 186-370; G4 = 373-631; G5 = 652 and lower.
360), followed by G3, with 20.1% (n = 352). The bottom two ranking groups (G4 and G5) had 17.6% (n = 308) and 13.7% (n = 240), respectively. Overall, there was a significant difference in the number of endorsement contracts across profes sional golfers’ expertise levels, %-(22) = 51.45,/? < .001. Thus, HI was supported.
Informational Versus Transformation Products The purpose of H2 was to examine the interaction effects of professional golfers’ world rankings on two types of endorsed products (informational vs. transforma tional). For informational products, higher ranked professional golfers (G 1 n = 88, 38.3%, and G2 n = 70, 30.4%) were more frequently used as product endorsers than lower ranked professional golfers (G4 n = 40, 17.4%, and G5 n = 32, 13.9%). In particular, higher ranked professional golfers more often endorsed the financial, energy, airline, and medical industries (G1 and G2) than did lower ranked profes sional golfers. In contrast, for transformational products, both higher ranked (G 1 n = 396, 34.4%, and G2 n = 286, 24.9%) and lower ranked professional golfers (G4 n = 263, 22.9%, and G5 n == 205, 17.8%) were similarly used as product endors ers. Overall, a significant difference was noted for these variables, x2(3) = 8.43, p < .05. Thus, H2 was supported.
High-Versus Low-Involvement Product The purpose of H3 was to investigate the interaction effects of professional golfers’ expertise levels on types of endorsed products, in terms of product involvement level. For high-involvement products, the results showed that higher ranked (G1 n - 466, 35.0%, and G2 n = 345, 25.9%) professional golfers were more often used as endorsers than lower ranked professional golfers (G4 n = 290, 21.8% and G5 n = 231, 17.3%). More specifically, among high-involvement products, airline and luxury-product industries mainly employed higher ranked professional golfers (G 1 and G2) as their endorsers, whereas financial, hotel, and apparel industries used professional golfers as their endorsers, regardless of their expertise level. On the other hand, for low-involvement products, the results indicate there were few endorsement contracts for either higher (G1 n = 18, 37.5%, and G2 n = 11, 22.9%) or lower ranked professional golfers (G4 n = 13, 27.1%, and G5 1JSC Vol. 8, No. 4, 2015
Athlete Endorsement
421
n = 6, 12.5%). Overall, however, there was no significant difference between professional golfers’ expertise levels and product-involvement levels, %2(3) = 1.47,/? > .05. Thus, H2b was not supported.
Correspondence Analysis Between Professional Golfers’ Expertise and Type of Industry The results of correspondence analyses provide very interesting results regarding the link between professional golfers’ expertise level (column variables) and vari ous types of endorsed industries (row variables) in a two-dimensional figure. The proximity between a pair of points for column and row variables were employed to interpret the strength of the fundamental relationship between them. According to the results of correspondence analyses, a total of four dimen sions perfectly fit the data matrix. The first two dimensions for the eigenvalues accounted for cumulative proportions of inertia equal to 79.4%. Therefore, two-dimensional approaches were deemed appropriate for the data. The third dimension only added an additional 14.4% of the variance. As suggested by Hair et al. (1998), the singular values (eigenvalues) of the first and second dimensions were .312 and .20, which were sufficiently significant to employ correspondence analysis that met the .20 criterion to achieve meaningful dimensions. In addition, the overall chi-square indicated that the relationship between professional golfer group and type of industry was significant (%2= 151.459, df= 88,/? < .001) to use correspondence analysis, as suggested by Weller and Romney (1990). In sum, for this study it was recommended to keep a low-dimensionality (two dimensions) solution to facilitate interpretation. Based on this result, a cross-tabulation matrix was constructed by relating the expertise levels of professional golfers (column variables) to target industries (row variables; see Table 3). The data suggest that golf-related industries (e.g., golf apparel, golf equipment) devotedly used professional golfers as endorsers regardless of expertise level. In addition, the financial and hotel industries frequently used professional golfers as their endorsers. Notably, the luxury-brand industries have the second-highest endorsement contracts with the highest ranking group (Gl) among general products. Overall results are summarized in Table 3. Figure 1 shows the joint-space map display of the professional golfers and types of endorsed industry in a two-dimensional map. The proximity between a pair of points for column and row variables was used to explain the strength of the fundamental relationship between them. In terms of professional golfers’ expertise levels, the financial-services industry has a strong association with professional golfers’ of all expertise level and thus is plotted in middle of the map. Furthermore, as Figure 1 indicates, the highest expertise group (G 1) has a strong association with luxury brands, automobiles, sports-related products, video games, and the medical, construction, energy, and airline industries and formed a cluster in the right side of the map. However, it might be difficult to generalize a strong association for Gl with video games because EA Sports was the only company that used professional golfers as endorsers. In this particular case, most of the higher ranking professional golfers were recruited (i.e., Rickie Fowler and Rory Mclloy). More specifically, as anticipated, the results indicate that the sports- and golfrelated industries seemed better positioned toward G l. The energy, construction, IJ S C V o l.8 , No. 4, 2015
422
Jang et al.
Table 3 Crosstab Results for Endorsers’ World Rankings and Industry Segments Involvem ent
Motivation
Product Category
G1
G2
G3
G4
G5
High
Informational
Financial
30
29
27
21
14
Consulting
8
5
10
7
5
Energy (oil, gas)
8
12
4
0
5
High-tech product
9
6
9
6
4
Transformational
Construction
3
4
1
1
2
Airline
20
4
2
1
1
Logistic
3
2
4
2
0
Automobile
2
1
1
0
0
Medical
4
3
0
1
0
Sport-related product
11
4
3
5
2
Golf-related product
259
189
179
157
123
Apparel (golf)
52
55
56
57
50
Luxury brand
34
11
16
11
6 19 1
23
20
25
21
Informational
Car repair
1
4
1
1
Transformational
Alcoholic beverage
4
1
3
1
1
Video game
6
1
0
0
0
Magazine
1
0
2
4
0
Newspaper
0
0
4
0
1
Hotel or resort Low
Other
Sunglass
2
3
0
2
0
Restaurant
4
2
5
5
3
Organization or charity
2
2
0
3
1
Department store
2
2
0
2
2
Note, x 2 = 143.529, df = 88, p < .001. G1 = 58 professional golfers with World Ranking 1-76; G2 = 58 golfers with World Ranking 77-185; G3 = 58 golfers with World Ranking 186-370; G4 = 58 golfers with World Ranking 373-631; G5 = 56 professional golfers with Word Ranking 632-652.
sunglasses, and car-repair industries were more strongly associated with G2, located in the upper half of the map. In addition, department stores, organizations, and charities formed a cluster positioned at the left of G2. Notably, the hightech-manufacturing, logistics, consulting, and restaurant industries developed a strong linkage with G3, while the golf-apparel and resort (or hotel) industries were displayed between G4 and G5. In the bottom-left quadrant of the map, the newspaper- and magazine-publishing companies formed another cluster close to G3. Finally, alcoholic beverages were located between G3 and G l. The specific relationship between professional golfer groups and types of endorsed industry is displayed in Table 3 and Figure 1. IJ S C V o l.8 , No. 4, 2015
2
Ctt
ca
CD —
•
ca
“ •>
■
^
G5
-----------,_________ U O
B I2 A
AB23
la E o is
JJi
a
■
o wr —
- r - _________ 3
m « CO
◄
CO
0
°
5
LA t- i
&
o
S
5
oo
r-1
s
IJ S C V o l. 8, N o. 4, 2015 a i § H
ct3
re
ra
a _o
G
i hJ
©
◄
rgCQ