Sport Management Review 13 (2010) 142–157

Contents lists available at ScienceDirect

Sport Management Review journal homepage: www.elsevier.com/locate/smr

Dimensions of general market demand associated with professional team sports: Development of a scale Kevin K. Byon a,*, James J. Zhang b,1, Daniel P. Connaughton c,2 a

Sport Management Program, School of Human Performance & Recreation, The University of Southern Mississippi, 118 College Drive #5142, Hattiesburg, MS 39406-0001, United States b Department of Tourism, Recreation and Sport Management, College of Health and Human Performance, University of Florida, Post Office Box 118208, Gainesville, FL 32611-8208, United States c Department of Recreation, Tourism & Sport Management, College of Health and Human Performance, University of Florida, Post Office Box 118208, Gainesville, FL 32611, United States

A R T I C L E I N F O

A B S T R A C T

Article history: Received 3 October 2008 Received in revised form 30 July 2009 Accepted 31 July 2009

The purpose of this study was to develop the Scale of Market Demand to assess general market demand factors affecting the consumption of professional team sports, which was completed through the following five steps: (a) formulation of a theoretical framework, (b) development of a preliminary scale, (c) exploratory factor analysis (EFA), (d) confirmatory factor analysis (CFA), and (e) examination of predictive validity through conducting a structural equation modeling (SEM) analysis. Following a community intercept method, professional sport consumers (N = 453) in four southeastern metropolitan areas responded to the scale. Data were randomly split into two halves: one for EFA and the other for CFA. In the EFA with alpha extraction and promax rotation, six factors with 31 items emerged: opposing team, home team, game promotion, economic consideration, sport epitome, and schedule convenience. In the CFA with maximum likelihood estimation, five factors with 17 most pertinent items were retained, without the sport epitome factor. This five-factor model displayed good fit to the data, discriminant validity, and high reliability. The SEM revealed that home team, opposing team, and game promotion were predictive of game re-attendance behaviour. ß 2009 Sport Management Association of Australia and New Zealand. Published by Elsevier Ltd. All rights reserved.

Keywords: Market demand Sport consumption Exploratory factor analysis Confirmatory factor analysis Scale development Structural equation modeling

1. Introduction According to Shank (2005), game attendance is the most traditional and important form of sport consumption behaviour. Chelladurai (1999) classified the sport industry into three segments: (a) sport economic activities, (b) spectator sports, and (c) participant sports. Of the three segments, spectator sports have been the fastest growing in terms of annual business transactions (Street & Smith’s Sport Group, 2007). Other researchers have also recognised the continuous growth of spectator sports in North America and pointed out that spectator sports have become increasingly popular leisure behaviour for Americans (Ross & James, 2006; Trail, Anderson, & Fink, 2005). The rapid and vast growth of professional sport teams in

* Corresponding author. Tel.: +1 601 266 6085; fax: +1 601 266 4445. E-mail addresses: [email protected] (K.K. Byon), [email protected]fl.edu (J.J. Zhang), [email protected]fl.edu (D.P. Connaughton). 1 Tel.: +1 352 392 4042x1274; fax: +1 352 392 7588. 2 Tel.: +1 352 392 4042x1296; fax: +1 352 392 7588. 1441-3523/$ – see front matter ß 2009 Sport Management Association of Australia and New Zealand. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.smr.2009.07.005

K.K. Byon et al. / Sport Management Review 13 (2010) 142–157

143

North America correlates highly with the increased interest in spectator sports among American people. Masteralexis, Barr, and Hums (2008) indicated that as of 2007, a total of 149 franchise teams belong to the five major professional sport leagues: Major League Baseball (MLB), National Football League (NFL), National Basketball Association (NBA), National Hockey League (NHL), and Major League Soccer (MLS). This figure does not include teams in less-prominent professional sport leagues such as the Women’s National Basketball Association (WNBA), National Lacrosse League (NLL), and many other major and minor leagues. The augmentation of spectator sports has also been confirmed through attendance and media viewership rates. The same phenomenon is also true in NCAA Division I men’s basketball and football, which are considered to be the two main revenue producers for collegiate athletic departments (Fulks, 2003). In 2005, nearly 1.3 million people watched March Madness men’s college basketball games online (Rein, Kotler, & Shields, 2006). The increasing popularity of spectator sports has led to the establishments of new leagues (e.g., Women’s Professional Soccer, National Lacrosse League), teams (e.g., Gold Coast Titans, San Jose Earthquakes, Oklahoma City Thunder), and multimedia outlets (e.g., NFL Sunday Ticket via Direct TV, ESPN 360), which have not only provided more spectating options for sport consumers but have also created greater competition among various leagues and teams for consumers. With such a crowded sport marketplace, sport consumers have many options in which to spend their leisure time and discretionary dollars. As a result, professional sport organizations face stiff competition in an effort to gain market share. Mullin, Hardy, and Sutton (2007) stated that ‘‘competition for sport dollar is growing at the pace of a full-court press’’ (p. 7) to describe the intensity of the competitive sport marketplace. According to Zhang, Smith, Pease, and Jambor (1997), market competition also comes from many other forms of leisure and entertainment options. As market competition becomes even more intense in professional sports, it is important for both researchers and practitioners to understand game consumption related variables in order to improve the quality of product offering and enhance competitiveness of the sport product (i.e., sport games). Also, marketers of professional sport teams need to develop strategic marketing plans that are based on an in-depth understanding of consumers. It is critical for them to identify those contributing variables that affect the consumption of spectators. Following an extensive literature review on factors influencing game attendance variables, Schofield (1983) proposed four market demand categories including demographic variables, economic variables, game attractiveness, and residual preference. Greenstein and Marcum (1981) and Jones (1984) focused their studies on game production functions and found that team performance variables, such as winning/losing record and presence of star player(s) were related to game attendance. Synthesizing key game demand variables and production functions, Zhang, Pease, Hui, and Thomas (1995) proposed the systematic concept of market demand, which was defined as the spectators’ expectations towards the main attributes of the core product (i.e., the game itself). Braunstein, Zhang, Trail, and Gibson (2005) further explained that market demand was a set of essential constructs associated with the game that a sport team could offer to its existing and prospective consumers. Essentially, it is a cluster of pull factors associated with the game that a professional sport team can offer to its prospective and returning spectators (Braunstein et al., 2005; Hansen & Gauthier, 1989; Schofield, 1983). In a sense, market demand variables are comprised of the attributes of core game quality variables that are directly related to athlete/team performance, game schedule, and/or ticket affordability. Although market demand has consistently been found to be an important concept that influences sport consumption behaviour (Zhang, Lam, & Connaughton, 2003; Zhang et al., 1995), previous studies were usually conducted in the context of a specific sport (Braunstein et al., 2005; Zhang et al., 2003; Zhang et al., 1995), lacking generalizability and application to a broader market environment of professional sports. To fill this void, the unique characteristics associated with general professional team sports were incorporated into the scale development in this study to inspire and enhance scientific inquiry into this topic area. It was anticipated that the developed Scale of Market Demand (SMD) would be frequently adopted by researchers and marketers to examine the target markets of professional team sports and their demands for the core attributes of game product. 1.1. Conceptual framework A theoretical justification for the market demand studies can be partially attributed to the cognitive-oriented attitude models, including the Adequacy Importance Model by Mazis, Ahtola, and Klippel (1975), the Theory of Reasoned Action proposed by Fishbein and Ajzen (1975), and the Theory of Planned Behaviour (Ajzen, 1991). The Adequacy Importance Model is a multi-attribute model that predicts consumer behaviour and was developed to understand consumers’ behaviour from a cognitive structure perspective. In this model, attitude was operationalized as an importance-weighted evaluation toward the attributes and dimensions of the products or services. Based on three studies, Mazis et al. (1975) found that this model explained consumer’s behavioural intentions well. Mashiach (1980) supported the use of a multi-attribute model to predict sport consumers’ attendance behaviour by arguing that behavioural intentions formed by sport consumers were not a function of a single attribute, such as winning; instead, they were influenced by multiple attributes. In terms of predicting attitude in purchase behaviour, Oliver (1980) contended that attitude affects the purchase intentions of products and services. Consumers tend to form their attitude based on prior expectations regarding the performance of the product or service, which often results in ensuing behavioural intentions. The importance of attitude as a predictor of behavioural intentions was well explained by the Theory of Reasoned Action (Fishbein & Ajzen, 1975) and its updated version, termed the Theory of Planned Behaviour (Ajzen, 1991). These theories postulate that human behaviour is a direct consequence of behavioural intentions, which are functions of attitude, subjective norms, and perceived behavioural control. Several researchers have found that the attitude construct better explained behavioural intentions than subjective

144

K.K. Byon et al. / Sport Management Review 13 (2010) 142–157

norms and perceived behavioural controls did (Stutzman & Green, 1982; Warshaw, Calantone, & Joyce, 1986). A strong attitude towards a certain object or phenomenon could act as a powerful heuristic that positively influences consumer behaviour (Fazio, Powell, & Williams, 1989). Likewise, when a sport consumer holds a positive attitude towards the attributes of game product such as home team/athlete performance, and/or game schedule, the positive attitude tends to be transformed into attendance and re-attendance behaviour (Zhang et al., 2003). Based on previous studies, it can be suggested that a multi-attribute model of market demand which reflects the unique nature of sport games can be a driving force of sport consumption behaviour, including game attendance or re-attendance. Following the various attitude models (i.e., Adequacy Importance Model, Theory of Reasoned Action, and Theory of Planned Behaviour) as primary theoretical frameworks and also based on the empirical findings of previous market demand studies (e.g., Braunstein et al., 2005; Zhang et al., 2004; Zhang et al., 2003; Zhang et al., 1995), a six-factor model of general market demand was proposed in the current study. The six factors were: home team, opposing team, game promotion, economic consideration, sport epitome, and schedule convenience. Based on a review of related literature, the documentation and rationale for selecting each of the factors are presented below. 1.1.1. Home team The first proposed dimension, home team, is defined as the perceived quality of the home team that is represented by such attributes as home team performance, presence of superstar(s), athletic quality of home team players, win/loss record, home team reputation, and/or home team league standing. Previous studies found that this factor was most influential on the attendance of NBA (Zhang et al., 1995) and minor league hockey (Zhang et al., 1997) games. Zhang, Wall, and Smith (2000) found that win/loss record was positively related to NBA season ticket holder’s game attendance. Bird (1982) in his football game attendance study found that league standing had a direct relationship with game attendance. 1.1.2. Opposing team Opposing team refers to the perceived quality of the opposing team and includes variables such as opposing team performance, athletic quality of opposing team, overall athletic quality of opposing team players, opposing team history and tradition, opposing team league standing, opposing team as a rivalry, and/or opposing team superstar(s). Madrigal (1995) found that competitive quality of opponent was related to affective reactions (Enjoyment and Basking-In-Reflected-Glory), both of which had a direct relationship with spectator satisfaction. In the context of the NHL, Jones (1984) found that the presence of star players on the home team, visiting team, or both was what motivated sport consumers to attend hockey games. Athletic quality of opposing team, opposing team history, league standing, and presence of superstar(s) on the opposing team were consistently found to be contributing variables to game attendance (Greenwell, Fink, & Pastore, 2002; Zhang et al., 1995, 1997). 1.1.3. Sport epitome Sport epitome is defined as the overall perception toward the general features of professional team sports. A wide range of sport epitome attributes have been identified as influencing variables, including but not limited to, closeness of competition, popularity of professional team sport, duration of game, high level of skills, best players in a sport, and speed of game. At times, this factor has been referred as ‘Love of a Sport’ (Braunstein et al., 2005; Ferreira & Armstrong, 2004; Zhang et al., 2003). Braunstein et al. found that this factor was an important factor representing Spectator Decision Making InventorySpring Training (SDMI-ST). Similar findings were discovered in Zhang et al.’s (2003) study on general market demand associated with professional sport consumption as this factor was found to be positively related to game attendance and media consumption. When studying game attributes that influenced college students’ game attendance, Ferreira and Armstrong (2004) found such variables as the duration of event, popularity of sport, high level of skill displayed, and speed of game were salient in influencing game attendance. Additionally, a number of previous studies on spectator motivation have recognised various sport epitome elements that were related to sport consumption. For example, closeness of competition was regularly identified as a variable for the drama dimension by Funk, Mahony, Nakazawa, and Hirakawa (2001), Trail and James (2001), and Wann (1995). Funk et al. and Trail and James also included high level of skills as a part of the physical skills of the athletes factor in their studies. 1.1.4. Economic consideration Economic consideration is defined as an individual’s concern of economic issues, including such variables as ticket price, ticket affordability, good seats, and/or ticket discounts. Previous studies have shown statistical significance but conflicting results regarding the impact of economic consideration on game attendance. In their longitudinal study on MLB game attendance, Baade and Tiehen (1990) found that economic consideration was negatively related to game attendance. Similar findings were found in Bird’s (1982) football game study and another study on general market demand of professional sports conducted by Hansen and Gauthier (1989). On the other hand, Zhang et al. (1995) found that ticket discounts, good seats, and group ticket cost were positively associated with the attendance of NBA games. 1.1.5. Game promotion Game promotion is defined as the specific mixture of marketing tools used by a sport organization for persuading consumers to attend the events or consume the product (Kotler & Armstrong, 1996). Fullerton and Merz (2008) argued that one of the frequently used marketing tools by professional sport marketers is sales promotion because oftentimes it helps to

K.K. Byon et al. / Sport Management Review 13 (2010) 142–157

145

gain the attention of consumers, increase their interest, and eventually sell sport products. For instance, Pavelchak, Antil, and Munch (1988) found that people who recalled Super Bowl advertisements tended to possess a favorable attitude toward the companies and brands that were featured in commercials during the Super Bowl. Direct marketing has been found to be an effective promotional strategy that forms a positive relationship with the company (Fullerton & Merz, 2008; Kotler & Armstrong, 1996). Recently, Seo and Green (2008) recognised that the Internet has become a crucial promotional tool used to influence sport consumption of professional sport consumers. The game promotion factor can be represented by such attributes as advertising, direct mail and notification, publicity, and web information. This factor should be distinguished from in-game entertainment amenities that can be manipulated by team marketers on a game-by-game basis (Zhang et al., 2005). Previous studies have indicated that game promotion activities (e.g., advertising, direct mail/notification) were positively related to various professional sports’ event attendance (Zhang et al., 2003; Zhang et al., 1995). 1.1.6. Schedule convenience The last dimension in the proposed model, schedule convenience, is defined as the assigned time and day in which a sport game takes place. Zhang et al. (1995) found that schedule convenience was related to game attendance. Hill, Madura, and Zuber (1982) found that schedule convenience was positively related to game attendance at weekend and season ending games, but not afternoon games, in the MLB. Zhang (1998) examined minor league hockey spectators’ preferred time for game attendance and found that spectators preferred evening times (7:00 pm) for weekday and Saturday games, and late afternoon times (4:00 pm) for Sunday games. This factor appears to be less predictive of game attendance when compared to other market demand factors in the proposed model. However, in various studies on sport market demand, the schedule convenience factor usually emerged as an influencing dimension of consumer behaviour (Braunstein et al., 2005; Zhang et al., 2003; Zhang et al., 1995). Numerous studies on sport market demand have been conducted to examine its ability to predict game attendance and/ or re-attendance (e.g., Hansen & Gauthier, 1989; Zhang et al., 2004; Zhang et al., 2003; Zhang et al., 1995). Market demand factors have been consistently identified as influencing factors for spectator attendance of professional team sport events (Braunstein et al., 2005; Greenstein & Marcum, 1981; Hansen & Gauthier, 1989; Jones, 1984; Zhang et al., 2003; Zhang et al., 1995). Involving a sample of spectators of NBA regular season games, Zhang et al. (1995) found that four factors (home team, opposing team, game promotion, and schedule convenience) were related to NBA game attendance, with 6% of the variance explained. Similarly, Zhang et al. (2004) found that game attractiveness, economic consideration, and marketing promotion factors were positively related to past, current, and future consumption of NFL games, with 14% of the variance explained. In Zhang et al.’s (2003), study on general market demand of professional sports, Game Attractiveness and economic consideration factors were found to be predictive of game consumption level, with 22% of the variance explained. Although tremendous research efforts have been made, two major limitations have been identified in previous studies. First, previous studies mainly focused on specific professional sports such as the NBA (Zhang et al., 1995), MLB (Braunstein et al., 2005), and NFL (Zhang et al., 2004), suggesting a possible lack of a greater generalizability of the measurement instruments and the research findings to consumers beyond a specific sport. Second, Zhang, Lam et al.’s study was limited by including only three general factors in the study (i.e., game attractiveness, marketing promotion, and economic consideration) and ignored the presence of other possible factors. For instance, Braunstein et al. (2005) identified love of professional sport as an important dimension of market demand that was related to MLB Spring Training games. Schedule convenience is another factor that was consistently found to be an important factor of market demand. From an analytical perspective, Bagozzi (1980) argued that one reason for model misspecification in marketing research could be due to omitting important variables from the model. Further examination of the concept and factor structure of general market demand appears necessary in order to formulate general promotion guidelines for professional team sport researchers and practitioners. Therefore, the purpose of this study was to examine the dimensions of general market demand associated with professional team sports and develop a scale that reflected the identified dimensions. The objective of this study was accomplished through a comprehensive measurement process, including formulation of a preliminary scale, a test of content validity, a pilot study, test administration, and various statistical analyses of the scale’s psychometric properties. 2. Method 2.1. Participants For the purpose of including professional team sport spectators from diverse backgrounds in terms of sociodemographics, sport preference, and sport consumption levels, the current study employed a community intercept sampling method to recruit research participants (Brenner, 1996). A total of 453 respondents participated in the survey study. Research participants were those who resided in southeastern United States metropolitan cities (Atlanta, Jacksonville, Tampa, and Miami) or within close proximity of those cities, where one or more professional team sports were franchised at the time this survey was conducted. Participation in the survey was voluntary and a participant had to be 18 years of age or older. To qualify for participation in the study, an individual must have also attended at least one professional team sport event within the previous 12 months. These sampling conditions enabled the research participants to be familiar with the game products of professional team sport events (Petrick, 2002). The final responses resulted from data collection at seven sport bars, three shopping malls, one grocery store, one community park, and one college campus. As shown in Table 1, the characteristics of respondents were consistent with

146

K.K. Byon et al. / Sport Management Review 13 (2010) 142–157

Table 1 Frequency distributions for the sociodemographic variables (N = 453). Variables

Category

Frequency (%) (N = 453)

Cumulative %

Gender

Male Female

274 (60.5) 179 (39.5)

60.5 100.0

Age

18–22 23–30 31–40 41–50 51–65

41 175 151 58 28

(9.0) (38.6) (33.3) (12.8) (6.2)

9.0 47.7 81.0 93.8 100.0

Number of people in household

1 2 3–4 5–6 7–8 9 or more

90 112 179 62 7 3

(19.9) (24.7) (39.5) (13.7) (1.5) (0.7)

19.9 44.6 84.1 97.8 99.3 100.0

Household income

Below $20,000 $20,000–39,999 $40,000–59,999 $60,000–79,999 $80,000–99,999 $100,000–149,999 $150,000–199,999 Above $200,000

23 79 128 81 56 42 27 17

(5.1) (17.4) (28.3) (17.9) (12.4) (9.3) (6.0) (3.8)

5.1 22.5 50.8 68.7 81.0 90.3 96.2 100.0

Marital status

Single Married Divorced

241 (53.2) 195 (43.0) 17 (3.8)

53.2 96.2 100.0

Education

In school now High school graduate In college now College graduate Advanced degree

1 47 45 265 95

(0.2) (10.4) (9.9) (58.5) (21.0)

0.2 10.6 20.5 79.0 100.0

Ethnicity

Caucasian African American Hispanic Asian/Pacific islander American Indian Interracial Other

259 61 87 40 2 2 2

(57.2) (13.5) (19.2) (8.8) (0.4) (0.4) (0.4)

57.2 70.6 89.8 98.7 99.1 99.6 100.0

Occupation

Management Technical Professional Sales Clerical Education Skilled worker Non-skilled worker Other

79 28 128 60 12 111 30 3 2

(17.4) (6.2) (28.3) (13.2) (2.6) (24.5) (6.6) (0.7) (0.4)

17.4 23.6 51.9 65.1 67.8 92.3 98.9 99.6 100.0

Attended game

AFL MLB NBA NFL NHL SOCCER

15 99 117 203 18 1

(3.3) (21.9) (25.8) (44.8) (4.0) (0.2)

3.3 25.2 51.0 95.8 99.8 100.0

those general backgrounds of professional sport consumers as described by the Simmons Market Research Bureau (2007). In terms of the most recent game attended, 44.8% of the respondents indicated that they attended a NFL game, followed by a NBA game (25.8%), a MLB game (21.9%), a NHL game (4%), an AFL game (3.3%), and a MLS game (0.2%). 2.2. Scale development Major concepts in the Spectator Decision-Making Inventory (Zhang, Lam, Bennett, Connaughton, et al., 2003; Zhang et al., 2003; Zhang et al., 1995) were adopted, which originally consisted of four dimensions (home team, opposing team, game promotion, and schedule convenience). Two additional factors, economic consideration and sport epitome, were added to

K.K. Byon et al. / Sport Management Review 13 (2010) 142–157

147

the scale based on research indications made by numerous investigators such as Braunstein et al. (2005), Hansen and Gauthier (1989), and Schofield (1983). A majority of the items were derived from adoptions and modifications of various existing scales. A very small number of items (approximately 10%) were generated from a review of other published literature. In particular, all adoptions and modifications took into consideration the following three aspects: (a) unique product and service features of professional team sports, (b) general nature of this study with an attempt to include all professional team sports, and (c) relevance and reliability of related factors and items that were reported in previous studies. Consequently, a total of 46 items were generated for the six SMD factors: home team (10 items), opposing team (10 items), sport epitome (10 items), economic consideration (6 items), game promotion (4 items), and schedule convenience (6 items). These items were phrased into statements in a Likert-type scale with 5 ranking points, ranging from 1 = ‘Not at All’ to 5 = ‘Very Much.’ The SMD section was preceded by the following statement: ‘Please rate each of the following variables that might have influenced your decision making to attend the most recent professional team sport event within the past 12 months.’ Besides the SMD items, the questionnaire also included three game consumption and nine sociodemographic variables. Inclusion of the game consumption variables was for the purpose of examining the predictability of SMD factors. Three items of re-attendance intentions were adopted from So¨derlund (2006) consumer behaviour scale that was derived through proper measurement procedures. These items were all under the conative consumption factor. In this study, they were slightly modified to be reflective of the professional sport game setting. The items were preceded with the following statement: ‘With respect to the professional team sport event that you most recently attended, please rate the following statements that assess your intentions for future attendance.’ Each item was phrased into a Likert-type scale with 5 ranking points, ranging from 1 = ‘Extremely Unlikely’ to 5 = ‘Extremely Likely.’ Additionally, inclusion of sociodemographic variables for the purpose of sample description included gender, age, ethnicity, household size, household income, marital status, education occupation, and professional sport game most recently attended. 2.3. Procedures The preliminary questionnaire that included SMD, game re-attendance, and sociodemographic items was submitted to a panel of 10 experts for content validity testing. The panel included five university professors and five practitioners. Each panel member was requested to examine the overall content of the items under each of the stipulated factors. Following the feedback of the panel members, the preliminary instrument was modified in the areas of item adequacy, factor relevance, and word clarity. With the modified version of the questionnaire, a pilot study was conducted to a small sample (n = 32) of sport consumers who attended one or more professional team sport events within the previous 12 months. The purpose of the pilot study was to further examine the content validity from the perspective of the targeted population. At this stage, suggested changes and improvements were all minor and primarily related to word clarifications. A survey packet which included the revised questionnaire, a cover letter explaining the purpose of the study and requesting cooperation from a participant, and the Informed Consent form was prepared. Approval from the Institutional Review Board for the Protection of Human Participants was obtained prior to data collection. The researchers first contacted the general managers of the targeted community locations to obtain permission to conduct the study. The survey was administered only with the permission from the general manager of a location. To ensure a good representation of professional team sport consumers with different backgrounds in the sample, data collections were conducted on both weekdays and weekend. Four trained research assistants provided on-site support with the data collection process. Completing a questionnaire, on average, took approximately 15 minutes. In terms of the sample size required for factor analyses (i.e., EFA and CFA) employed in this study, Gorsuch (1983) suggested at least 5 respondents are desirable for each measured variable. Considering that the current study utilized both EFA and CFA for market demand factors that had a total of 46 observed variables, it was appropriate for the researchers to target a minimum number of 460 participants. Consequently, a total of 470 questionnaires were collected. Of those, 17 questionnaires were discarded due to having non-sporadic missing values, following the suggestions made by Zhang, Pease, and Hui (1996). Therefore, a total of 453 questionnaires were included in subsequent data analyses necessary for scale development. Missing values were rarely spotted within the remaining sample of 453 respondents. Among those occasional missing data points, there were no Not-Missing-At-Random (NMAR) data (Rubin, 1987; Schafer & Graham, 2002). Only a few Missing-At-Random (MAR) were detected. For those MAR data, mean substitution was applied. 2.4. Data analyses The total sample of 453 was randomly split into two halves. The first data set (n = 231) was used for conducting EFA and the second set (n = 222) was employed for conducting CFA. Procedures in the SPSS version 15.0 (SPSS, 2006) were carried out to calculate descriptive statistics for sociodemographic, SMD, and re-attendance variables, execute the EFA, and calculate reliability coefficients. Because the factors and items in the SMD were compiled from numerous existing scales and indications of previous studies, an EFA was deemed necessary as the initial analytical step to examine the factor structure of the measure. The primary purpose of the EFA was to identify a unique and reliable simple factor structure that was of the potential to be generalized to a universe of variables from a sample of variables, so as to reduce any redundant data.

148

K.K. Byon et al. / Sport Management Review 13 (2010) 142–157

Following an EFA, internal consistency reliability was examined by calculating the Cronbach’s alpha coefficients for the identified factors (Cronbach, 1951). In the EFA, alpha factoring extraction was applied, which was a preferred extraction approach when the intention was to enhance the reliability and generalizability of factors (Kaiser & Caffrey, 1965). The extracted factors were rotated by employing the promax rotation approach to identify distinct factors (Hendrikson & White, 1964). The promax rotation was developed by combining the advantages of varimax (orthogonal) and oblique rotation techniques (Fabrigar, Wegener, MacCallum, & Strahan, 1999). The following criteria were used to determine the factors and their items: (a) a factor had an eigenvalue equal to or greater than 1.0 (Kaiser, 1974), (b) an item had a factor loading equal to or greater than .40 (Nunnally & Bernstein, 1994), (c) a factor had at least 3 items (Hair, Black, Babin, Anderson, & Tatham, 2006), and (d) an identified factor and retained items must be interpretable in the theoretical context. The scree plot test was also utilized to help make a determination on the number of extracted factors (Cattell, 1966). Procedures in the AMOS version 7.0 (Arbuckle, 2006) were executed for conducting the CFA with maximum likelihood (ML) estimation for the retained SMD factors that were resolved from the EFA. According to Bollen (1989) and Hair et al. (2006), conducting a CFA needs to take the following five steps: (a) model specification, (b) model identification, (c) model estimation, (d) testing model fit, and (e) model respecification. Following the suggestions of Hair et al. (2006), several goodness of fit measures were adopted, which included chi-square statistic (x2), normed chisquare (x2/df), root mean square error of approximation (RMSEA), standardized root mean residual (SRMR), comparative fit index (CFI), and expected cross-validation index (ECVI) (Bollen, 1989; Hu & Bentler, 1999). Bollen suggested that cutoff values of less than 3.0 for the normed chi-square are considered reasonable fit. Brown and Cudeck (1992) indicated that any RMSEA values less than .05 show a close fit. More recently, Hu and Bentler (1999) suggested that RMSEA value of .06 also indicates a close fit. Any values of RMSEA between .06 and .08 indicate acceptable fit. Values of RMSEA between .08 and .10 show mediocre fit. A value greater than .10 indicates unacceptable fit (Hu and Bentler). SRMR indicates how large residuals are; therefore, smaller values of SRMR show good fit. Any values from 0.10 to 0.05 are considered indicative of an acceptable fit, while values below 0.05 indicate an excellent fit (Hu & Bentler, 1998). The CFI assesses ‘‘the relative improvement in fit of the researcher’s model compared with a baseline model (i.e., null model)’’ (Kline, 2005, p. 140). A rule of thumb for CFI is that any value larger than .90 indicates an acceptable fit and any value greater than .95 shows a close fit. Lastly, the ECVI measures the fit across samples and has no set criteria. Generally, ECVI is used to compare models, with a smaller value indicating better model fit. Three tests were employed to measure the reliability of the scales: Cronbach’s coefficient alpha (a) values, construct reliability (CR), and average variance extracted (AVE). Nunnally and Bernstein (1994) suggested .70 be used as a cut-off value for the Cronbach’s alpha in the newly developed scale. However, they further recommended .80 be considered for the scale refinement study. Since the purpose of the current study was to develop the SMD, the recommended .70 cutoff value was adopted to determine internal consistency of Cronbach’s alpha (Nunnally and Bernstein). Fornell and Larcker (1981) defined CR as an internal consistency measure that accounts for the measurement errors of all indicators. We adopted .70 as a threshold of CR as suggested by Fornell and Larcker. AVE is an alternative way to examine reliability of the latent construct. AVE is defined as an amount of variance that is accounted for by the construct, relative to the amount of variance due to measurement errors of all indicators (Fornell & Larcker, 1981). The benchmark value for AVE was equal to or greater than .50 suggested by Bagozzi and Yi (1988). Since AMOS program did not provide CR and AVE values, CR and AVE values were calculated using the procedures outlined by Fornell and Larcker. Indicator loadings and significant z-values were evaluated with an item loading equal to or greater than .707 (i.e., R2 value .50) (Anderson & Gerbing, 1988). Additionally, discriminant validity was examined to measure how distinct the constructs were between one another. To establish discriminant validity, two methods were employed. According to Kline (2005), discriminant validity can be established when an interfactor correlation is below .85. Fornell and Larcker suggested a more robust way of measuring discriminant validity in which a squared correlation between two constructs should be lower than the AVE value for any one of the two constructs. Finally, a structural equation modeling (SEM) test was conducted following the procedures in the AMOS version 7.0 to examine the relationship between SMD factors and re-attendance intention variables. The same fit index criteria as in the CFA were adopted to examine the goodness of fit level in the structural model. Path coefficients were calculated and used to determine the direct effect of market demand factors on re-attendance intentions (Kline, 2005). 3. Results 3.1. Descriptive statistics Descriptive statistics for the SMD variables are presented in Table 2. Of the 46 items, 38 had a mean score greater than 3.0 (i.e., midpoint on the five-point Likert scale), indicating that overall market demand variables were considered important when making a decision to attend a professional team sport event. Seven items had a mean score that was lower than the midpoint. Of the variables in the scale, the ‘love professional team sport(s)’ item had the highest mean score (M = 4.22; SD = 0.95) and the ‘web information’ item had the lowest mean score (M = 2.42; SD = 1.20). Additionally, skewness and kurtosis for the items were examined. For the skewness cut-off value, an absolute value of 3.0 would be considered extreme. For the kurtosis threshold value, an absolute score greater than 3.0 would be considered extreme (Chou & Bentler, 1995). In this study, all skewness and kurtosis values for the SMD items were well within the acceptable threshold.

K.K. Byon et al. / Sport Management Review 13 (2010) 142–157

149

Table 2 Descriptive statistics for the Scale of Market Demand variables (N = 453). Variable

M

SD

1. Home team win/loss record (HT1) 2. Home team star player(s) (HT2) 3. Home team record breaking performance (HT3) 4. Overall quality of home team players (HT4) 5. Home team reputation (HT5) 6. Home team league standing (HT6) 7. Home team history and tradition (HT7) 8. Home team exciting play (HT8) 9. Support the home team (HT9) 10. High level of skills (HT10) 11. Opposing team’s overall performance (OT1) 12. Opposing team star player(s) (OT2) 13. Opposing team history and tradition (OT3) 14. Opposing team reputation (OT4) 15. Overall quality of opposing team players (OT5) 16. Opposing team league standing (OT6) 17. Quality of opposing team (OT7) 18. Opposing team as a rivalry (OT8) 19. Opposing team exciting play (OT9) 20. Player charisma of opposing team (OT10) 21. Played that sport(s) (SE1) 22. Closeness of competition (SE2) 23. Popularity of professional team sport (SE3) 24. Duration of the game (SE4) 25. High level of skills (SE5) 26. Best players in a sport (SE6) 27. Speed of game (SE7) 28. Athleticism of professional team sport (SE8) 29. High level of competitiveness (SE9) 30. Love professional team sport(s) (SE10) 31. Personal ticket price (EC1) 32. Ticket affordability (EC2) 33. Good seats (EC3) 34. Group ticket cost (EC4) 35. Ticket discount (EC5) 36. Sales promotions (EC6) 37. Advertising (GP1) 38. Direct mail and notification (GP2) 39. Publicity (GP3) 40. Web information (GP4) 41. Game time of the day (SC1) 42. Convenient game schedule (SC2) 43. Weather condition (SC3) 44. Day of the week (SC4) 45. Travel distance (SC5) 46. Location of venue (SC6)

3.59 3.93 3.27 4.02 4.16 3.83 4.00 4.01 4.16 3.77 3.23 3.23 3.42 3.25 3.35 3.24 3.47 3.69 3.12 3.20 2.89 3.39 3.82 2.87 3.88 3.79 3.21 3.64 3.94 4.22 3.17 3.32 3.67 2.67 3.17 2.92 2.87 2.50 3.22 2.42 3.66 3.75 3.35 3.66 3.15 3.59

1.22 1.19 1.38 .93 1.03 1.14 1.12 .979 1.02 1.24 1.15 1.28 1.08 1.10 1.15 1.16 1.06 1.14 1.22 1.22 1.47 1.09 1.21 1.29 1.03 1.23 1.30 1.09 1.05 .95 1.27 1.21 1.15 1.30 1.37 1.30 1.20 1.31 1.18 1.20 1.05 .93 1.32 1.02 1.19 1.25

Skewness

Kurtosis

.69 1.0 .28 .95 1.23 .89 .96 1.08 1.20 .93 .44 .47 .46 .33 .52 .47 .63 .72 .35 .46 .03 .54 .84 .05 .90 .80 .38 .52 .90 1.14 .18 .38 .61 .10 .27 .15 .002 .24 .31 .38 .98 .82 .43 .76 .11 .66

.49 .05 1.13 .95 .91 .14 .02 1.03 .85 .07 .59 .85 .28 .47 .46 .53 .19 .12 .83 .76 1.33 .30 .18 1.15 .37 .42 .93 .40 .25 .63 .94 .61 .42 1.20 1.10 1.17 .91 1.23 .78 .89 .72 .86 .89 .26 .73 .49

Note: HT = home team; OT = opposing team; SE = sport epitome; EC = economic consideration; GP = game promotion; SC = schedule convenience.

3.2. Exploratory factor analyses Using the first data set, an EFA was conducted for the purpose of identifying a simple structure among the SMD items (Stevens, 1996). The Kaiser–Meyer–Olkin (KMO) measure of sampling adequacy value was .845, which exceeded the cut-off value of .70, indicating that the degree of common variance was meritorious (Kaiser, 1974). As a result, the sample was adequate for a factor analysis. Bartlett’s Test of Sphericity (BTS) was 4521.27 (p < .001), indicating that the hypothesis of the variance and covariance matrix of the variables as an identity matrix was significantly rejected; hence, a factor analysis was deemed appropriate. In the EFA, six factors emerged with 31 items meeting the retention criteria, explaining a total of 57.69% variance among the variables. The scree plot test also suggested that a six-factor model was most interpretable. The results of the rotated pattern matrix from promax rotation are presented in Table 3. Based on the predetermined criterion of an item loading equal to or greater than .40, nine items were eliminated (i.e., high level of performance, home team star player(s), support the home team, high level of skills, weather condition, closeness of competition, opposing team as a rivalry, high level of competitiveness, and good seats). Six other items were removed due to having only one or two items loaded on the respective factors (i.e., home team record breaking performance, athleticism of professional team sport, best players in a sport, location of venue, love professional team sport(s), and popularity of professional team sport). Consequently, the six factors were labeled as opposing team (9 items), home team (6 items), game promotion (5 items), economic consideration (4 items), sport epitome (4 items), and schedule convenience (3 items). Alpha coefficients for the factors were .93, .85, .86, .83,

150

K.K. Byon et al. / Sport Management Review 13 (2010) 142–157

Table 3 Factor pattern matrix for the Scale of Market Demand variables: alpha factoring with promax rotation using first half data (n = 231). F1

F2

F3

F4

F5

F6

Opposing team (9 items) Quality of opposing team Overall quality of opposing team players Opposing team exciting play Opposing team star player(s) Opposing team reputation Player charisma of opposing team Opposing team league standing Opposing team’s overall performance Opposing team history and tradition

.86 .83 .82 .82 .81 .77 .73 .72 .64

.09 .10 .07 .17 .07 .14 .14 .07 .24

.11 .01 .09 .08 .19 .24 .02 .03 .16

.01 .00 .14 .10 .00 .10 .14 .16 .36

.00 .04 .11 .06 .12 .03 .11 .08 .22

.12 .12 .15 .02 .02 .05 .09 .04 .00

Home team (6 items) Home team win/loss record Home team league standing Home team reputation Home team history and tradition Overall quality of home team players Home team exciting play

.15 .01 .12 .08 .09 .08

.77 .74 .72 .70 .63 .57

.10 .06 .04 .07 .06 .04

.05 .07 .00 .01 .05 .13

.09 .02 .00 .18 .03 .02

.03 .02 .14 .03 .05 .08

Game promotion (5 items) Advertising Sales Promotions Direct mail and notification Publicity Web information

.01 .14 .04 .20 .09

.04 .01 .00 .02 .12

.90 .87 .82 .49 .47

.13 .09 .13 .06 .11

.10 .02 .01 .10 .06

.03 .08 .01 .23 .16

Economic consideration (4 items) Ticket affordability Travel distance Ticket discount Personal ticket price

.04 .01 .09 .13

.01 .07 .07 .03

.08 .09 .08 .08

.95 .71 .63 .59

.12 .07 .31 .04

.06 .04 .03 .18

Sport epitome (4 items) Group ticket cost Speed of game Duration of the game Played that sport(s)

.18 .09 .11 .12

.11 .02 .08 .12

.04 .03 .15 .06

.09 .30 .21 .10

.65 .57 .51 .51

.14 .18 .04 .02

Schedule convenience (3 items) Day of the week Convenient game schedule Game time of the day

.01 .11 .02

.01 .17 .01

.06 .04 .13

.08 .06 .05

.01 .06 .07

.66 .62 .62

Note: F1 = opposing team; F2 = home team; F3 = game promotion; F4 = economic consideration; F5 = sport epitome; F6 = schedule convenience.

.70, and .75, respectively, indicating that they were all internally consistent and reliable. The resolved factor structure was overall consistent with the conceptual model for developing the SMD scale in this study. Only items that were retained in the EFA (31 items) were included in the subsequent CFA. 3.3. Confirmatory factor analyses The second data set for the SMD variables, which contained 31 items under six factors, was submitted to a CFA with ML estimation (Hair et al., 2006). Goodness of fit indexes revealed that the six-factor and 31-item measurement model did not fit the data well. The chi-square statistic was significant (x2 = 1340.89, p < .001), indicating that the hypothesized model and the observed model had a statistically significant difference. Because chi-square value is known to be sensitive to sample size (Kline, 2005), alternative fit indices were further examined, including the normed chi-square, RMSEA, SRMR, CFI, and ECVI. The normed chi-square (x2/df = 3.20) was above the suggested cut-off value (i.e., <3.0), showing a poor fit. The RMSEA value indicated that the six-factor model showed a poor fit (RMSEA = .10, 90% CI = .094–.106). Although the value of SRMR (.077) was within the range of acceptable fit (.10), the CFI value of .78 was substantially lower than the recommended cut-off ratio (>.90), indicating an overall lack of fit to the data. The model fit tests overall suggested a need for model respecification. According to Tabachnick and Fidell (2001), model respecification would be needed if the proposed model did not fit the data well. In the current analysis, indicator loadings ranged from .398 (group ticket cost) to .903 (advertising). Of 31 items, nine items were below .707, indicating that these indicators’ error variance was greater than their shared variance. Therefore, the nine items were removed (i.e., opposing team history and tradition, home team exciting play, web information, travel distance, played that sport(s), speed of game, group ticket cost, duration of the game, and home team history and tradition). Furthermore, modification indexes suggested problems associated with five other items (i.e., opposing team star player(s), overall athletic quality of home team players, opposing team league standing, publicity, and player charisma of opposing

K.K. Byon et al. / Sport Management Review 13 (2010) 142–157

151

Table 4 Model fit comparison between the six-factor model and five-factor model of market demand using second half data (n = 222). Model

x2

df

x2/df

RMSEA

RMSEA CI

SRMR

CFI

ECVI

Six-factor model (31 items) Five-factor model (17 items)

1340.89 278.31

419 109

3.20 2.55

.10 .084

.094–.106 .072–.096

.077 .054

.78 .92

6.76 1.66

CI = confidence interval.

team). When examining the indicator loading matrix, it was apparent that these items had highly double loading coefficients that were in the range of .50–.60, indicating that the items loaded on more than one predetermined factor. After careful consideration of statistical justifications, the decision to remove these items was made. As a result of the model respecification, a five-factor model with 17 items was conceptualized: home team (3 items), opposing team (5 items), game promotion (3 items), economic consideration (3 items), and schedule convenience (3 items). This five-factor model with 17 items was further submitted to a CFA. Overall, goodness of fit revealed that the five-factor model fit the data reasonably well. Chi-square statistic was significant (x2 = 278.31, p < .001). The normed chi-square (x2/ df = 2.55) was lower than the suggested cut-off value (i.e., <3.0). The RMSEA value indicated that the five-factor model had an acceptable fit (RMSEA = .084, 90% CI = .072–.096). The SRMR (.054) was of a good value (.10). CFI was .92, which was considered acceptable. ECVI was 1.66, which indicated a much better fit than that of the six-factor model. Overall goodness of fit of the five-factor model improved drastically, indicating an increased level of acceptability when compared with the sixfactor model (Table 4). The reliability of the factors and their respective items was evaluated by Cronbach’s alpha, CR, and AVE (Table 5). Cronbach’s alpha values for the five-factor model ranged from .80 (schedule convenience) to .91 (opposing team). The CR values for the five constructs ranged from .76 (economic consideration) to .82 (opposing team; game promotion). Furthermore, all AVE values were above the suggested standard, ranging from .52 (economic consideration) to .64 (opposing team). Based on the overall information of reliability, the determined factors were deemed reliable. All of the indicator loadings were statistically significant with z scores ranging from 8.99 to 16.79 (p < .05). In addition, all indicator loadings were greater than the suggested standard of .707 (Anderson & Gerbing, 1988) except for two items on home team and schedule convenience (‘home team reputation’ with a value of .60 and ‘day of the week’ with a value of .67). A decision was made to retain these items due to their theoretical relevance to the home team or schedule convenience factors and the values were only slightly lower than the .707 threshold. Overall, the five-factor SMD supported a parsimonious structure (Table 5). No interfactor correlations were above .85, ranging from .19 (between game promotion and economic consideration) to .51 (between economic consideration and schedule convenience), indicating discriminant validity. The Fornell and Larcker’s test found that all squared correlations in the scale were less than AVE value for each respective construct, indicating excellent discriminant validity, as illustrated in Table 5 for AVE values and Fig. 1 for interfactor correlations. As a result of the Table 5 Indicator loadings, critical ratios, Cronbach’s alpha, construct reliability, average variance extracted for the market demand using second half data (n = 222). Variables

Indicator loadings

Opposing team (5 items) Opposing team’s overall performance Opposing team reputation Overall quality of opposing team players Quality of opposing team Opposing team exciting play

.90 .74 .84 .83 .76

Home team (3 items) Home team win/loss record Home team reputation Home team league standing

.94 .60 .77

Game promotion (3 items) Advertising Direct mail and notification Sales promotions

.92 .80 .82

Economic consideration (3 items) Personal ticket price Ticket affordability Ticket discount

.71 .91 .77

Schedule convenience (3 items) Game time of the day Convenient game schedule Day of the week

.81 .79 .67

Critical ratios

Cronbach’s alpha

Construct reliability

Average variance extracted

.91

.82

.64

.81

.80

.58

.88

.82

.61

.83

.76

.52

.80

.80

.57

13.52 16.79 16.43 13.91

8.99 11.56

14.84 15.31

10.92 10.41

10.91 9.54

K.K. Byon et al. / Sport Management Review 13 (2010) 142–157

152

Fig. 1. First-order confirmatory factor analysis for the Scale of Market Demand.

two factor analyses, a preliminary construct validity test, (but pending further validation), and reliability tests, it was found that the five-factor model with 17 items was the most appropriate model measuring general market demand associated with professional team sports. 3.4. Structural equation modeling To examine the predictability of the SMD factors to re-attendance intentions, a SEM analysis was conducted. The SMD items under the five factors were predicting variables and the items under the re-attendance intentions factor were the criterion variables. The re-attendance intentions factor was found to have good factor loading and internal consistency. The factor loadings of the items were .97, 74, and .84, respectively. Cronbach’s alpha value for the factor was .84. Therefore, it was appropriate to use the re-attendance intentions factor for the SEM analysis. Prior to estimating path coefficients for the hypothesized structural model, goodness of fit indexes for the overall structural model was first evaluated. The overall model fit was good (x2 = 332.32, p < .001, x2/df = 2.14, CFI = .93, RMSEA = .072, 90% CI = .061–.085, SRMR = .052). Having a satisfied model fit, it was appropriate to proceed with a SEM analysis. It was hypothesized that the SMD factors would be predictive of re-attendance intentions. The SEM analysis revealed that the SMD predicted a total of 16.6% of the variance in spectators’ re-attendance intentions. More specifically, the standardized direct effect of home team (b = .245, p < .001) and opposing team (b = .210, p < .001) showed a positive influence on re-attendance intentions, respectively. These indicated that when market demand towards home team increased by one standard deviation unit, re-attendance intentions also increased by .245 standard deviations; similarly, when market demand of opposing team increased by one standard deviation unit, re-attendance intentions increased also by .210 standard

K.K. Byon et al. / Sport Management Review 13 (2010) 142–157

153

deviations. However, the SEM analysis revealed that the standardized direct effect of game promotion had an inverse relationship with re-attendance intentions (b = .245, p < .001), indicating that when market demand of game promotion increased by one standard deviation unit, re-attendance intentions decreased by .245 standard deviations. The economic consideration and schedule convenience factors were not found to be related to re-attendance intentions. 4. Discussion As market competition in professional team sports becomes more intense, it is imperative for both researchers and practitioners to identify those variables that directly and indirectly influence game consumption (Hansen & Gauthier, 1989; Zhang et al., 1995). In-depth understanding of what factors influence spectators’ decision making to attend games, and how they refer the game products and services to others, is crucial for professional teams to better understand spectator consumption behaviour. Findings of previous studies revealed that market demand variables were important predictors of sport spectator consumption behaviour (Murray & Howat, 2002; Zhang, 1998; Zhang et al., 2004, 1995). Although previous researchers recognised the importance of market demand variables when marketing professional sport games, most of the studies were conducted in the context of a specific professional sport (Greenwell et al., 2002; Zhang, Lam, Bennett, et al., 2003; Zhang et al., 1995). In these studies, scales developed under specific settings might not be directly applicable to other settings due to the lack of consideration of marketing characteristics that are generalizable to various professional team sport events. Thus, it was necessary to conduct this study that incorporated the attributes of core game products and market environment related to professional team sports (Mullin et al., 2007; Zhang et al., 2003). The current study was designed to develop the SMD measuring general market demand of professional team sports. In this study, psychometric testing procedures were meticulously conducted. For instance, systematic procedures were undertaken to formulate the preliminary scale, which included a review of literature, test of content validity by a panel of experts, and a pilot study by a group of consumers representing the targeted population. Previous research typically studied professional sport consumers at a limited number of sport events located in one geographic location. It was the intention of this study to include consumers of more diverse backgrounds in terms of geographic locations, sport types, consumption levels, and sociodemographics in an effort to improve the external validity of the developed scale. In this study, both EFA and CFA were conducted to ensure factor validity and generalizability of the resolved factor structure. For the SMD variables, six factors with 31 items were retained in the EFA: (a) opposing team, (b) home team, (c) game promotion, (d) economic consideration, (e) sport epitome, and (f) schedule convenience. The derived factors from the EFA were consistent with the theoretical dimensions suggested by previous researchers (Greenstein & Marcum, 1981; Hansen & Gauthier, 1989; Schofield, 1983; Zhang et al., 1995). However, the six-factor model did not fit the data well in the initial CFA. After careful consideration of modification indices, the scale was revised to a five-factor model with a total of 17 items: opposing team (5 items), home team (3 items), game promotion (3 items), economic consideration (3 items), and schedule convenience (3 items). This respecified model exhibited much improved fit indexes. As a result of the respecification, the sport epitome factor was eliminated, mainly due to low indicator loadings. In previous studies, this factor was sometimes labeled as ‘love of the sport,’ and it was found to be a contributing variable to game attendance of intercollegiate competitions (Ferreira & Armstrong, 2004) and professional games (Braunstein et al., 2005; Zhang, Lam, et al., 2003; Zhang et al., 2003). In Braunstein et al.’s (2005) study, the researchers found that Love of Baseball was an important factor related to MLB spring training game market demand; however, the factor in fact displayed poor psychometric properties. The factor was eventually retained by Braunstein et al. based on the consideration that Love of Baseball covered essential characteristics of MLB games, such as closeness of competition, duration of game, high level of skills, best players in a sport, and speed of game. Although the researchers of that study were reluctant to eliminate this factor, they did suggest the need for further studies examining this factor. The researchers of the current study conducted rigorous procedures in item purification, test of content validity, and a pilot study; yet, similar findings occurred with this factor. There was a wide array of aspects that were included in the sport epitome factor. Including a player perspective in this factor could have also led to correlations with the player aspects that were already in the previous dimensions. As Braunstein et al. suggested, additional examination of this factor is necessary in future studies. A key issue in future studies is how to keep the sport epitome factor theoretically separated from home team and opposing team factors because the factor analyses in the current study revealed that sport epitome items were double loaded on these two factors. Although the number of factors was reduced to five, the resolved constructs of the SMD were essentially consistent with previously suggested factors (Braunstein et al., 2005; Schofield, 1983; Zhang et al., 2004; Zhang et al., 2003; Zhang, Lam, et al., 2003; Zhang et al., 1995). Nevertheless, findings of this study were likely more reliable and generalizable when considering the following aspects: (a) a more representative sample was involved and (b) more appropriate items were derived from advanced statistical analyses (i.e., CFA and SEM), In previous studies, data were usually collected on-site in arenas or stadiums (e.g., Zhang, Lam, Bennett, et al., 2003; Zhang et al., 1995), where only spectators of one sporting event participated in the study and they might have been under temporal influence due to occurrence of winning or losing. The

154

K.K. Byon et al. / Sport Management Review 13 (2010) 142–157

respondents of the current study were current sport consumers who indicated that they had attended one or more professional team sport events within the previous 12 months. Descriptive statistics indicated that a total of six premier professional team sport leagues were attended by the respondents, which may help improve the external validity of the findings and usefulness of the scale. Additionally, Zhang et al. (2003) suggested that besides EFA and CFA, other types of construct validity including convergent and discriminant validity tests be utilized to improve the factor structure. These suggestions were materialized in the current study, however, convergent validity of the current scale needs to be further validated in future studies. The current study retained at least three items per factor through the CFA. One of the limitations found in Zhang et al. (2003) study was that opposing team and schedule convenience factors were measured by only two items. When using CFA and SEM analyses, having the optimal number of items per factor is an important consideration for measurement precision (Bollen, 1989). The possible reason that two items of opposing team and schedule convenience factors were consistently used in previous studies may be due to the use of EFA as the primary item selection method, which is data-driven. Although our results were also data-driven due to reliance on modification indices when the scale was modified, the current study improved in that five items and three items related to opposing team and schedule convenience, respectively, were retained in the factor analyses (i.e., EFA and CFA). The items of the opposing team represented overall performance, athletic quality of the opposing team, athletic quality of players, opposing team’s exciting play, and team reputation. Schedule convenience was represented by such attributes as game time of the day, convenient game schedule, and day of the week. In the current study, one of the items (i.e., home team reputation) in the home team factor did not meet the Anderson and Gerbing (1988) conservative criterion, but the item was retained due to its high level of theoretical relevance. Nonetheless, more research to validate the items related to the opposing team, schedule convenience, and home team factors appears necessary in future studies. Overall, the five-factor SMD measurement model displayed good psychometric properties, particularly in the areas of construct reliability and discriminant validity as verified by findings of EFA and CFA. Meanwhile, the research findings on content validity were still preliminary; yet, they showed promising results. The panel members for the test of content validity included academicians, practitioners, and sport consumers, who examined item relevance, representativeness, and clarity of the SMD items. Even so, a substantial number of initial items in the preliminary scale had to be excluded from the final version of the scale, suggesting that discrepancies existed between the expert panel and information obtained from the sport consumers. Future studies may consider applying a comparatively more rigorous approach of testing content validity in the social science domain as outlined by Aiken (1996) and Dunn, Bouffard, and Rogers (1999). In this contemporary approach, item examinations by the panel members are focused on the constructs. The three aspects of content validity (i.e., relevance, representativeness, and clarity) are simultaneously evaluated through various statistical procedures, including content validity coefficient, rater homogeneity coefficient, and rater reliability coefficient. Through this process, discrepant raters can be determined, more stringent items are retained, and weak or double loaded items are identified. Improved content validity would lead to improved convergent validity. Of the research findings, the high indicator loadings, high AVE values, and low interfactor correlations provided strong evidence for achieving a good factor structure of the SMD scale; nonetheless, no effort was made in this study to examine the convergent validity of the SMD factors. Convergent validity can usually be examined by relating to the scale and its factors of a previously validated measure that was developed in same or similar measurement environment and purpose. Although some similar factors were found in previous studies that measured market demand for professional sports (Zhang et al., 2003; Zhang et al., 1995), they were developed in different marketing contexts. Thus, due to the absence of previously validated scales, no attempt was made in this study to adopt this approach. However, future studies should make an effort to accomplish this needed measurement aspect. Ideally, the multitrait and multi-method approach should be adopted to substantiate and confirm the construct validity of the scale (Campell & Fiske, 1959). By implementing this advanced procedure, both convergent and discriminant validity can be examined simultaneously and a better understanding can be obtained not only on what is being measured, but also on how an attribute is being measured. A number of previous studies have provided evidence of criterion validity (e.g., sport consumption) for the market demand factors and indicated that market demand factors would explain 6–22% of game consumption variance (Zhang et al., 2003; Zhang et al., 1995). In the current study, examination of criterion validity focused on one aspect of criterion validity, the predictability of the SMD factors to re-attendance intentions. Findings that the home team, opposing team, and game promotion factors were predictive of re-attendance intentions revealed that these factors were of basic capacity for analysing an individual’s conative consumption behaviours, and these findings were consistent with those of previous studies (e.g., Braunstein et al., 2005; Hansen & Gauthier, 1989; Zhang et al., 2003). It is necessary to note that cumulatively, the three factors explained less than 17% of variance in re-attendance intentions, indicating that 83% was explained by concepts and factors not included in the SMD. In particular, none of the factors explained more than 8% of variance by itself. In addition, two other SMD factors (economic consideration and schedule convenience) explained no variance at all. Thus, the overall predictive validity was deemed partial. The negative relationship between game promotion and re-attendance intentions was interesting and somewhat different from the findings of previous studies. This finding may have in fact led to the following two speculations on the relevance level of game promotion to the market demand of professional team sports: (a) those with high demand for game promotion activities may have lower intention to attend sport events when compared to those with low demand for game promotion activities and (b) satisfying market demand for game promotion activities would likely be effective in making the game attractive to consumers with comparatively low consumption levels. Certainly,

K.K. Byon et al. / Sport Management Review 13 (2010) 142–157

155

these speculations need further examination in future studies. The current study did not make an attempt to examine the predictability of the SMD factors to other past, present, and future consumption variables. This study also failed to examine the relationships between the SMD factors and those same or similar factors identified in previous studies (e.g., Braunstein et al., 2005; Zhang et al., 2003; Zhang et al., 1995) by adopting the previously validated-test approach of criterion validity. Conducting all these tests in future studies would certainly be helpful in understanding all aspects of the criterion validity of the SMD scale. This study was also limited due to its sample size. Due to the necessity of conducting the EFA, only one half of the sample (n = 222) was used for the CFA. Although Wetson and Gore (2006) suggested that a minimum sample size of 200 was adequate for a CFA, the small sample size might have negatively influenced model fit for the market demand construct in the current study (Cheung & Rensvold, 2002). According to MacCallum, Roznowski, and Necowitz (1992), any model respecification should have an additional independent sample for cross-validation for the respecified model in order to avoid capitalization on chance. In the current study, the respecified model was not validated by an independent sample. Therefore, a careful interpretation of the results is required due to the following three reasons: (a) the preliminary model was revised to improve the overall model fit, (b) generalizability of the revised model to the population of interest for the current study remains unknown, and (c) the revised model was data-driven and considered as plausible, pending further validation using an independent sample in future studies. In the current study, no effort was made to examine if differences existed between die-hard and fair-weather fans in terms of the factor structures of the market demand construct. According to Wann and Branscombe (1990), spectators can at least be categorized as die-hard and fair-weather fans based on their consumption levels and socio-motivations. Die-hard fans generally are of higher team identification, involvement, and consumption levels than fair-weather fans; and are likely to support a team when the team does not perform well (Heere & James, 2007; Trail, Fink, & Anderson, 2003). Perhaps, diehard fans pay more attention to core product attributes (e.g., win/loss, level of performance, and/or the presence of star players); whereas fair-weather fans pay more attention to peripheral attributes (e.g., venue, promotion, and/or entertainment). These speculations need further examination by assessing invariance issues with respect to the factor structure of the SMD. When doing this, a number of sociodemographic and psychological variables, such as gender, team identification, and consumer involvement level, may be incorporated into the study. Additionally, data in this study were collected through a community intervention approach; thus, research participants were those who attended professional team sport events in the past. Due to the decay of memory, some of the respondents might not have been able to provide their responses with specificity. Hence, future studies should also examine the invariance issues between on-site and recall settings. With continued improvement, the SMD has great potential for adoption by researchers and practitioners. When considering its reasonable numbers of items (i.e., 17 items), the scale can be easily administered during on-site events or in various community locations. The scale can be used as a marketing tool for professional team sport organizations to better understand their spectator consumption behaviours. It can also be utilized in conjunction with other theoretically related variables, including but not limited to, game support programs, team identification, and word-of-mouth to examine how they function together to enhance sport consumption behaviours. Acknowledgements The authors would like to thank two anonymous reviewers and the editor of SMR (Graham Cuskelly) for their detailed and insightful comments on earlier drafts of this article. References Aiken, L. R. (1996). Rating scales and checklists: Evaluating behavior, personality, and attitude. New York: Wiley. Anderson, D. R., & Gerbing, D. W. (1988). Structural equation modeling practice: A review and recommended two-step approach. Psychological Bulletin, 103, 411– 423. Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50, 179–211. Arbuckle, J. L. (2006). AMOS 7. 0 user’s guide. Chicago, IL: Small Walters Corporation. Baade, R., & Tiehen, L. (1990). An analysis of major league baseball attendance: 1969–1987. Journal of Sport and Social Issues, 14, 14–32. Bagozzi, R. P. (1980). Causal models in marketing. New York: John Wiley. Bagozzi, R. P., & Yi, Y. (1988). On the evaluation of structural equation models. Journal of the Academy of Marketing Science, 16, 74–94. Bird, P. J. (1982). The demand for league football. Applied Economics, 14, 637–649. Bollen, K. A. (1989). Structural equations with latent variables. New York: John Wiley & Sons. Braunstein, J. R., Zhang, J. J., Trail, G. T., & Gibson, H. J. (2005). Dimensions of market demand associated with pre-season training: Development of a scale for major league baseball spring training. Sport Management Review, 8, 271–296. Brenner, S. (1996). 76ers community intercept survey suggests new marketing tactics. Team Marketing Report, 8(June (9)), 11–15. Brown, M. W., & Cudeck, R. (1992). Alternative ways of assessing model fit. Sociological Methods and Research, 21, 230–258. Campell, D. T., & Fiske, D. W. (1959). Convergent and discriminant validation by the multitrait–multimethod matrix. Psychological Bulletin, 56, 81–105. Cattell, R. B. (1966). The scree test for the number of factors. Multivariate Behavioral Research, 1, 245–276. Chelladurai, P. (1999). Human resource management in sport and recreation. Champaign, IL: Human Kinetics. Cheung, G. W., & Rensvold, R. B. (2002). Evaluating goodness-of-fit indexes for testing measurement. Structural Equation Modeling, 9, 233–255. Chou, C. P., & Bentler, P. M. (1995). Estimates and tests in structural equation modelling. In R. H. Hoyle (Ed.), Structural equation modeling: Concepts, issues and applications (pp. 37–55). Thousand Oaks, CA: Sage. Cronbach, L. J. (1951). Coefficient alpha and the internal structure of test. Psychometrika, 16, 297–334.

156

K.K. Byon et al. / Sport Management Review 13 (2010) 142–157

Dunn, J. G., Bouffard, M., & Rogers, W. T. (1999). Assessing item content-relevance in sport psychology scale construction research: Issues and recommendations. Measurement in Physical Education and Exercise Science, 3, 15–36. Fabrigar, L. R., Wegener, D. T., MacCallum, R. C., & Strahan, E. J. (1999). Evaluating the use of exploratory factor analysis in psychological research. Psychological Methods, 4, 272–299. Fazio, R. H., Powell, M. C., & Williams, C. J. (1989). The role of attitude accessibility attitude-to-behavior process. Journal of Consumer Research, 16, 280–288. Ferreira, M., & Armstrong, K. L. (2004). An exploratory examination of attributes influencing students’ decisions to attend college sport events. Sport Marketing Quarterly, 13, 194–208. Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention, and behavior: An introduction to theory and research. Reading, MA: Addison-Wesley. Fornell, C., & Larcker, D. (1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 18, 39– 50. Fulks, D. L. (2003). Revenues and expenses of Divisions I and II intercollegiate athletic programs: Financial trends and relationships—2002–2003. Overland Park, KS: National Collegiate Athletic Association. Funk, D. C., Mahony, D. F., Nakazawa, M., & Hirakawa, S. (2001). Development of the Sport Interest Inventory (SII): Implications for measuring unique consumer motives at team sporting events. International Journal of Sports Marketing & Sponsorship, 3, 291–316. Fullerton, S., & Merz, R. (2008). The four domains of sports marketing: A conceptual framework. Sport Marketing Quarterly, 17, 90–108. Gorsuch, R. L. (1983). Factor analysis (2nd ed.). Hillsdale, NJ: Erlbaum. Greenstein, T., & Marcum, J. (1981). Factors affecting attendance of major league baseball: Team performance. Review of Sport and Leisure, 6(2), 21–34. Greenwell, T. C., Fink, J. S., & Pastore, D. L. (2002). Assessing the influence of the physical sports facility on customer satisfaction within the context of the service experience. Sport Management Review, 5, 129–148. Hair, J. F., Black, W. C., Babin, B. J., Anderson, R. E., & Tatham, R. L. (2006). Multivariate data analysis (6th ed.). Upper Saddle River, NJ: Prentice Hall. Hansen, H., & Gauthier, R. (1989). Factors affecting attendance at professional sport events. Journal of Sport Management, 3, 15–32. Heere, B., & James, J. D. (2007). Stepping outside the lines: Developing a multi-dimensional team identity scale based on social identity theory. Sport Management Review, 10, 65–91. Hendrikson, A. E., & White, P. O. (1964). Promax: A quick method of rotation to oblique simple structure. British Journal of Statistical Psychology, 17, 65–70. Hill, J. R., Madura, J., & Zuber, R. A. (1982). The short run demand for major league baseball. Atlantic Economic Journal, 10(2), 31–35. Hu, L. T., & Bentler, P. M. (1998). Fit indices in covariance structure modeling: Sensitivity to underparameterized model misspecification. Psychological Methods, 3(4), 424–453. Hu, L. T., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6, 1–55. Jones, J. C. H. (1984). Winner, losers, and hosers: Demand and survival in the National Hockey League. Atlantic Economic Journal, 12(3), 54–63. Kaiser, H. F. (1974). An index of factorial simplicity. Psychometrika, 39, 31–36. Kaiser, H. F., & Caffrey, J. (1965). Alpha factor analysis. Psychometrika, 30, 1–14. Kline, R. B. (2005). Principles and practice of structural equation modeling (2nd ed.). New York: Guilford. Kotler, P., & Armstrong, G. (1996). Principles of marketing. Englewood, Cliffs, NJ: Prentice-Hall. MacCallum, R. C., Roznowski, M., & Necowitz, L. B. (1992). Model modifications in covariance structure analysis: The problem of capitalization on chance. Psychological Bulletin, 111, 490–504. Madrigal, R. (1995). Cognitive and affective determinants of fan satisfaction with sporting event attendance. Journal of Leisure Research, 27, 205–227. Mashiach, A. (1980). A study to determine the factors which influence American spectators to go to see the summer Olympics in Montreal, 1976. Journal of Sport Behavior, 3, 16–28. Masteralexis, L. P., Barr, C. A., & Hums, M. A. (2008). Principles and practice of sport management (3rd ed.). Sudbury, MA: Jones and Bartlett. Mazis, M. B., Ahtola, O. T., & Klippel, R. E. (1975). A comparison of four multi-attribute models in the prediction of consumer attitudes. Journal of Consumer Research, 2(2), 38–52. Mullin, B. J., Hardy, S., & Sutton, W. A. (2007). Sport marketing (3rd ed.). Champaign, IL: Human Kinetics. Murray, D., & Howat, G. (2002). The relationships among service quality, value, satisfaction, and future intentions of customers at an Australian sports and leisure center. Sport Management Review, 5, 25–43. Nunnally, J. C., & Bernstein, I. H. (1994). Psychometric theory (3rd ed.). New York: McGraw-Hill. Oliver, R. (1980). A cognitive model of the antecedents and consequences of satisfaction decisions. Journal of Marketing Research, 17, 460–469. Pavelchak, M. A., Antil, J., & Munch, J. M. (1988). The Super Bowl: An investigation into the relationship among program context, emotional experience, and ad recall. Journal of Consumer Research, 35, 360–367. Petrick, J. F. (2002). Development of a multi-dimensional scale for measuring the perceived value of a service. Journal of Leisure Research, 34, 119–134. Rein, I., Kotler, P., & Shields, B. (2006). The elusive fan: Reinventing sports in a crowded marketplace. New York: McGraw-Hill. Ross, S. D., & James, J. D. (2006). Major versus minor league baseball: The relative importance of factors influencing spectator attendance. International Journal of Sport Management, 7, 217–233. Rubin, D. B. (1987). Multiple imputations for nonresponse in surveys. New York: Wiley. Schafer, J. L., & Graham, J. W. (2002). Missing data: Our view of the state of the art. Psychological Methods, 7, 147–177. Schofield, J. (1983). Performance and attendance at professional team sports. Journal of Sport Behavior, 6, 196–206. Seo, W. J., & Green, B. C. (2008). Development of the motivation scale for sport online consumption. Journal of Sport Management, 22, 82–109. Shank, M. D. (2005). Sport Marketing: A strategic perspective (3rd ed.). Upper Saddle River, NJ: Prentice Hall. Simmons Market Research Bureau. (2007). Study of media and markets. New York: Market Research Bureau. So¨derlund, M. (2006). Measuring customer loyalty with multi-item scales: A case for caution. International Journal of Service Industry Management, 17, 76–98. SPSS. (2006). SPSS 15.0: Guide to data analysis. Upper Saddle River, NY: Prentice Hall. Stevens, J. (1996). Applied multivariate statistics for the social sciences (3rd ed.). Mahwah, NJ: Lawrence Erlbaum. Street. Smith’s Sport Group. (2007). About us: The sport industry Retrieved June 1, 2009 from http://www.sportsbusinessjournal.com/index.cfm?fuseaction= page.feature&featureId=43. Stutzman, T. M., & Green, S. B. (1982). Factors affecting energy consumption: Two field tests of the Fishbein–Ajzen model. Journal of Social Psychology, 117, 183– 201. Tabachnick, B. G., & Fidell, L. S. (2001). Using multivariate statistics (4th ed.). New York: Harper Collins. Trail, G. T., & James, J. D. (2001). The motivation scale for sport consumption: Assessment of the scale’s psychometric properties. Journal of Sport Behavior, 24, 108– 127. Trail, G. T., Anderson, D., & Fink, J. (2005). Consumer satisfaction and identity theory: A model of sport spectator conative loyalty. Sport Marketing Quarterly, 14, 98– 111. Trail, G. T., Fink, J., & Anderson, D. (2003). Sport spectator consumption behavior. Sport Marketing Quarterly, 12, 8–17. Wann, D. L. (1995). Preliminary validation of the sport fan motivation scale. Journal of Sport & Social Issues, 19, 377–396. Wann, D. L., & Branscombe, N. R. (1990). Die-hard and fair-weather fans: Effects of identification on BIRGing and CORFing tendencies. Journal of Sport and Social Issues, 14, 103–117. Warshaw, R., Calantone, B., & Joyce, M. (1986). A field application of the Fishbein and Ajzen intention model. Journal of Social Psychology, 126, 135–136. Wetson, R., & Gore, P. A. (2006). A brief guide to structural equation modeling. The Counseling Psychologist, 34, 719–751. Zhang, J. J. (1998). Preference of spectators toward professional hockey game time. International Sports Journal, 2, 57–70. Zhang, J. J., Connaughton, D. P., Ellis, M. H., Braunstein, J. R., Cianfrone, B., & Vaughn, C. (2004). Consumer expectations of market demand variables of an NFL expansion team. Journal of Contemporary Athletics, 1, 15–39.

K.K. Byon et al. / Sport Management Review 13 (2010) 142–157

157

Zhang, J. J., Lam, E. T. C., Bennett, G., & Connaughton, D. P. (2003). Confirmatory factor analysis of the spectator decision making inventory (SDMI). Measurement in Physical Education and Exercise Science, 7, 57–70. Zhang, J. J., Lam, E. T. C., & Connaughton, D. P. (2003). General market demand variables associated with professional sport consumption. International Journal of Sports Marketing and Sponsorship, 5, 33–55. Zhang, J. J., Pease, D. G., & Hui, S. C. (1996). Value dimensions of professional sport as viewed by spectators. Journal of Sport and Social Issues, 21, 78–94. Zhang, J. J., Pease, D. G., Hui, S. C., & Thomas, J. M. (1995). Variables affecting the spectator decision to attend NBA games. Sport Marketing Quarterly, 4, 29–39. Zhang, J. J., Piatt, D. M., Ostroff, D. H., & Wright, J. W. (2005). Importance of in-game entertainment amenities at professional sporting events: A case for NBA season ticket holders. Journal of Contemporary Athletics, 2, 1–24. Zhang, J. J., Smith, D. W., Pease, D. G., & Jambor, E. A. (1997). Negative influence of market competitors on the attendance of professional sport games: The case of a minor league hockey team. Sport Marketing Quarterly, 6(3), 31–40. Zhang, J. J., Wall, K. A., & Smith, D. W. (2000). To go or not? Relationship of selected variables to game attendance of professional basketball season ticket holders. International Journal of Sport Management, 1, 200–226.