A++ Work (quality, no plaigarism, on time)

Official Journal of NASSM www.JSM-Journal.com

ARTICLE

130

Journal of Sport Management, 2013, 27, 130-145 © 2013 Human Kinetics, Inc.

Shapiro and Ridinger are with the Dept. of Human Movement Sciences – Sport Management Program, Old Dominion Uni- versity, Norfolk, VA. Trail is with the Dept. of MSAL, Seattle University, Seattle, WA.

An Analysis of Multiple Spectator Consumption Behaviors, Identification, and Future Behavioral Intentions

Within the Context of a New College Football Program

Stephen L. Shapiro and Lynn L. Ridinger Old Dominion University

Galen T. Trail Seattle University

The growth of college sport over the last several years, combined with increased competition for the sport consumer dollar, has created a need to understand spectator consumption behavior. In addition, the impact of a new football program can generate interest that influences future spectator spending decisions. Using identity theory as a framework, the current study examined the differential effects of past sport consumer behaviors on various future sport consumer intentions within the context of a new college football program. Consumption intentions included attendance, sponsor support, and merchandise purchases. Furthermore, this investigation helped to determine how much variance past behaviors would explain in behavioral intentions after controlling for nine points of attachment. Data were collected from spectators of a Football Championship Subdivision (FCS) football program located in the Mid-Atlantic region. The findings suggest past behavior predicted future intentions; however, the amount of variance explained varied dramatically depending on specific past behaviors and points of attachment. These results can help sport marketers develop strategies to capitalize on the interest generated through new athletic programs.

College athletic departments have continued to increase generated revenues over recent years. According to Fulks (2011), National Collegiate Athletic Association (NCAA) Division I Football Bowl Subdivision (FBS) schools, which is the highest level of college football competition in the Unites States, saw a 9.5% increase from 2009 to 2010 in median generated revenue. NCAA Division I Football Championship Subdivision (FCS) schools, which are one level below FBS schools in regard to football competition, experienced even larger median revenue increases (14%) over the same time period. How- ever, total expenses have increased at approximately the same rate. Only 22 college athletic programs reported a profit in 2010 (Fulks, 2011).

In the current financial landscape of college sport, revenue growth is essential to cover the increase in costs. The primary areas of college athletic revenue, which include ticket sales, charitable contributions, sponsorship, broadcasting rights, and merchandise purchases (Fulks, 2011) are primarily spectator driven. Fans purchase tick- ets and merchandise, make annual contributions, support

program sponsors, and consume games through mediated channels (i.e., television, team/league websites, social media). Past fan consumption behavior through various means helps determine how likely fans are to engage in future sport consumption (Trail, Anderson, & Fink, 2000; Trail, Fink & Anderson, 2003). This is further supported by role identity and identity theory which suggests that identity with certain activities influences behavior related to those activities. According to Callero (1985), identi- ties by their very nature, imply action. The relationship between previous behaviors, identification, and future behavioral intentions become particularly important in a college athletics environment where spectator behav- iors drive revenue production. Therefore, it is important to understand spectator behavior and specifically, the factors that may have an influence on future spectator consumption intentions.

There is a wealth of literature examining the influ- ence of identification on sport consumer behavior. Previous research has focused on the development of identification measures (Robinson & Trail, 2005; Wann & Branscombe, 1993), the influence of team identification on attendance (Laverie & Arnett, 2000; Trail, Anderson, & Lee, 2006; Wakefield, 1995), and the influence of team identification on various future consumption behaviors (James & Trail, 2008; Trail et al., 2000, 2003, 2006; Trail, Anderson, & Fink, 2005). Team identification has been

An Analysis of Multiple Consumption Behaviors 131

shown to significantly influence consumption intentions related to attending games and purchasing merchandise (Trail et al., 2003, 2005, 2006), two vital revenue sources in college athletics.

However, the literature examining the differential effects of past consumption behaviors and identification on future consumption intentions is limited. Only Trail et al. (2006) focused solely on these relationships. The authors argued that the influence of previous behavior and team identification on future behavioral intentions is imperative because consuming an event and establishing a connection with a program helps move consumers up the fan commit- ment escalator; which Mullin, Hardy, and Sutton (2007) suggest increases overall consumption activity.

In addition, examining only one form of previous consumption behavior (attendance), one facet of identi- fication (team), and limited future intentions (attendance and merchandise purchases) is only a piece of the puzzle. Previous research supports the use of multiple facets of identification and the importance of other revenue sources (fundraising, sponsorship, broadcasting rights) in addition to attendance and merchandise (Fulks, 2011; Robinson & Trail, 2005; Trail, Robinson, Dick, & Gil- lentine, 2003). Research examining multiple categories of previous consumption behavior, identification, and future intentions combined, is nonexistent.

It is especially important to examine multiple facets of identification in the context of a new team. New sport teams do not have a history of achievement or well established traditions through which to attract fans. Thus, the facets that influence the formation of identification with a new team may be different from those that affect identification with an existing team (Lock, Taylor, & Darcy, 2011). Lock et al. (2011) found that the forma- tion of new team identification was driven primarily by identification with the sport, rather than with the specific team. These authors encouraged the inclusion of multiple points of attachment in future research on identification with new sport teams.

As leagues expand and new teams emerge (Tierney, 2009), a better understanding of the identity fans develop with a new team and its impact on consumptive behav- iors may help maximize marketing opportunities. To be viable in a competitive sports environment, a new team must attract, develop, and maintain a relationship with a substantial number of sport consumers (James, Kolbe, & Trail, 2002). In college football, as schools look to develop stronger connections with students, alumni, and the community, the growth of new programs has been sizeable. This includes 42 new college football programs in the 1980s, 22 in the 1990s, and 28 in the 2000s (Tier- ney, 2009). Furthermore, 25 additional college football programs are slated to begin by 2014 (McGuire, 2011). Research on consumer attitudes and behavior within the context of a new program is scant. Conceptual and theoretical development of sport consumer identification has focused on established teams, largely ignoring how identification might vary for a new team or league (Lock et al., 2011).

Therefore, the purpose of this study was to examine the relationship between previous fan/spectator behavior, identification, and future behavioral intentions, while incorporating multiple facets of each of these variables within the context of a new college football program.

Review of Literature Early research in sport consumer behavior focused on the factors that influence attendance (Demnert, 1974; Hansen & Gauthier, 1989; Noll, 1974; Whitney, 1988) or the development of economic models to predict atten- dance (Baade & Tiehen, 1990; Greenstein & Marcum, 1981). However, these studies failed to examine con- sumer behavior factors that influence future consumption intentions. Individual factors such as consumer attitudes, feelings, and emotions influence how sport fans think and consume sport-related products and services (Mullin et al., 2007). Therefore, as an extension to this early work, later studies explored relationships between spectator identification and consumption behaviors (Laverie & Arnett, 2000; Madrigal, 1995; Wakefield, 1995; Wann & Branscombe, 1993). Wann and Branscombe devel- oped an instrument to measure team identification and examined the impact of identification on fan behavior. The results provided evidence that fans with high levels of team identification appear to be more involved with their team and more willing to invest time and resources into being a fan. Madrigal (1995) extended this research through an examination of the relationship between team identification and fan satisfaction. Team identification was found to have a dominant influence on fan satisfac- tion. However, actual consumption was not measured in this model. Wakefield (1995) also examined team identification, but focused specifically on repatronage intentions as an outcome variable. The author found a positive relationship between team identification and future intentions providing some of the first empirical evidence regarding the influence of identification on future consumption intentions.

The previous studies demonstrate the significant role team identification plays in fan behavior. Fan iden- tity can be further explained through identity theory (Stryker, 1968, 1980; Stryker & Burke, 2000). Accord- ing to Stryker (1980), identity theory is focused on the concept that individuals develop identities through social experiences and relationships. Multiple aspects of identi- fication are internalized through these social exchanges. It has been hypothesized that the higher the salience of these identities, “the greater the probability of behavioral choices in accord with the expectation attached to that identity” (Stryker & Burke, 2000, p. 186). This is further supported by Callero (1985), who stated that the most discernable consequence of identity salience relates to actual behavior. The relationship between identification and behavior is apparent in both theoretical and practi- cal terms. Stronger identity salience leads to increased actions. The relationship between social experiences, identification, and behavior has been supported in areas

132 Shapiro, Ridinger, and Trail

such as student involvement in university organizations (Serpe & Stryker, 1987) and commitment to religious activities (Stryker & Serpe, 1982).

Within the context of sport, Laverie and Arnett (2000) examined spectator identification and behavior based on identity theory. It was suggested that role identities are formed through past sport-related experiences. High levels of identity salience influence current attitudes and future behavioral intentions. The authors found support for the relationship between team identification and attendance. However, the outcome variable used was past attendance. No other behaviors or behavioral intentions were consid- ered. Past behaviors have been shown to influence future behavioral intentions in a variety of contexts (Ouellette & Wood, 1998). However, to fully understand the role identification plays on consumption behavior through an identity theory framework, various past behaviors and various future intentions should be examined.

Trail et al. (2005) extended the work of Laverie and Arnett (2000) through the development of multiple models examining relationships between team identifi- cation, disconfirmation/confirmation of expectancies, mood, self-esteem, and future behavioral intentions (i.e., attendance, merchandise purchasing, overall team support). These models were created from previous theoretical (Trail et al., 2000) and empirical (Trail et al., 2003) studies that combined multiple determinants of spectator consumption in an effort to further under- stand fan behavior. The findings provided evidence of a direct relationship between team identification and future behavioral intentions along with an indirect rela- tionship between these two variables, influenced by fan self-esteem.

In addition, team identification may influence spe- cific consumption behaviors differently. Trail et al. (2003) found that the relationships between team identification and two types of consumption intentions (attendance and merchandise purchasing) were different, as indicated by the difference in factor loadings on the second-order latent variable (future behavior; Trail et al., 2003). These findings are further supported in the literature (James & Trail, 2008).

Trail et al. (2006) developed a model based on identity theory that focused exclusively on past atten- dance, team identification, future intentions, and actual attendance. This study was the first sport consump- tion examination that included previous consumption behavior, identification, and future behavior. The authors proposed that past attendance would predict preseason team identification, intentions to attend games, and actual game attendance. Findings showed that number of games attended explained approxi- mately 21% of the variance in team identification and past attendance and team identification combined explained 48% of the variance in future intentions. Although Trail et al. (2006) did not specifically test for mediation in their model; there certainly is the poten- tial for team identification to mediate the relationship between past attendance and future attendance. These

results offer further support regarding the influence of identification on future consumption behavior. In addition, this study highlights the impact that previous behavior can have on team identification.

In summary, the previous literature provides empirical support for two distinct relationships, (1) team identification and various spectator consumption behaviors and (2) past behaviors and future intentions. Additional research is needed to further understand these relationships. First, various consumption behaviors, both previous and future, in addition to attendance should be considered (i.e., merchandise purchases, mediated consumption, sponsor purchases). As mentioned previ- ously, merchandise purchases, sponsorship agreements, charitable contributions, and broadcasting contracts generate significant revenue in college athletics. In 2010, these areas accounted for approximately 28.4% of gener- ated revenue for FBS schools and approximately 33.7% of generated revenue for FCS schools (Fulks, 2011). Furthermore, attendance is only part of the complete fan experience. Many fans are not able to attend live games due to cost, location, or other obligations. Still, these fans can build identification through many of the alternative consumption methods previously mentioned.

Second, other forms of identification in addition to team identification must be considered. Previous research has shown support for multiple points of attachment or facets of identification (e.g., player identification, sport identification, coach identification; Robinson & Trail, 2005; Trail et al., 2003; Woo, Trail, Kwon, & Anderson, 2009). Previous consumption experiences may differen- tially influence various points of attachment and these points of attachment may differentially affect aspects of future intentions. Examining only team identification limits the opportunity to reach other segments of the fan market, which identify with alternative facets of the organization.

However, the literature examining the connection between past behaviors, multiple facets of identifica- tion, and various future intentions is underdeveloped. Trail et al. (2006) provided empirical evidence that past behavior and identification combined provided a more thorough explanation of the variance in future behav- ioral intentions. However, this was the only study that provided evidence of this relationship, and both previous and future consumption behaviors were measured only through attendance.

This becomes even more important when dealing with a new program where team identification may not yet have been established. The impact of a new football program is a unique consumption experience which may have an effect on various levels of identification and future consumption behavior. There has been a substan- tial growth of new sport teams in general and college football programs in particular in recent times. A better understanding of past behaviors and fan identification effects on future intentions for sport consumption can help cultivate a fan base for these new teams which is vital to their existence.

An Analysis of Multiple Consumption Behaviors 133

However, only a few studies examined identification in a new sport environment. James et al. (2002) found the reasons for purchasing season tickets for a new Major League Baseball team differed based on psychological connection to the team. Lock, Darcy, and Taylor (2009) examined member identification with a new club soccer team in Australia and concluded that age and income were related to identity strength. Lock et al. (2011) used a mixed-method approach to understand key themes driving the formation of new team identification for fans of Sydney FC, a soccer team in the newly developed Australian A-League. Their findings suggest that to attract fan support, a new team should leverage existing social identities such as identification with the sport or with the city where the new team plays. None of these studies looked at the impact of identification on future intentions.

The current study had two objectives. The primary objective was to test the differential effects of various past behaviors on multiple behavioral intentions. A second- ary goal within this objective was to examine how much variance past behaviors would explain in behavioral intentions after controlling for points of attachment. This methodology provides an evaluation of the total amount of variance explained by points of attachment, and helps to determine whether the points of attachment entirely subsume (mediate) the variance explained by past behaviors. The results will allow marketers and managers to understand whether it is necessary to take into account both past behaviors and points of attachment when trying to ascertain the determinants of future sport consumer behaviors. However, this type of analysis assumes that points of attachment potentially mediate the relationship between past behavior and behavioral intentions. Based on identity theory and the previous literature on identification and spectator consumption behavior noted above, this is a valid assumption, but it should be tested in the current data. Thus, the second objective was to test for mediation.

Method Research Setting The context for this study was a large public university in the Mid-Atlantic region with an enrollment of approxi- mately 23,000 students. It is the largest among several colleges and universities in a metropolitan community with a population of 1.5 million citizens, but has been considered a commuter school for many years. Foot- ball was essential to the university’s goal of shedding its commuter image and developing a greater sense of community on campus (Sander, 2010). After confirming student, alumni, and community interest in 2005, the Board of Visitors unanimously approved a plan to begin playing football at the FCS level and the inaugural home opener occurred on September 5, 2009.

Participants Data were collected from a random sample of fans (season ticket holders and students) who attended at least

one home football game during the inaugural season. Interest in the new team resulted in 73% of the seats being sold as season tickets. Another 20% of the seats were reserved for students, 5% were complimentary tickets for the athletic department, and 2% were provided to the visiting team, half of those on consignment. The only tickets available for purchase on game day were those unsold by the visiting team. Thus, the two largest groups, season ticket holders and students, were targeted for this study. The sample was selected from a list of season ticket holders and student ticket holders during the inaugural season. A total of 3,000 season ticket holders were randomly selected from a list of 14,450. In addition, 2,616 students were randomly selected from a database which included all students who registered for tickets and attended at least one game during the season. Online surveys were sent to a total of 5,616 fans and 1,092 usable surveys were returned for a response rate of 19.4%.

Instrumentation

The questionnaire used for the current study consisted of four sections with a total of 55 items. The first section had 12 items related to demographics to profile the typi- cal respondent. The second section had items measuring various forms of past consumption behavior including attendance, television viewership, radio listenership, print media consumption, merchandise purchases, member status and length of membership in the annual donor club, tailgating, and other mediated consumption (e.g., web content, Facebook, Twitter; see Table 1 for the list of the items/scales used in this research and how they were measured). These measures were adapted from earlier investigations examining previous behavior (Trail et al., 2003, 2005). Some of the questions were open-ended to collect continuous numeric data (e.g., How many home games did you attend this past season?). Other questions (7; e.g., I listened to the weekly football coach’s show) were measured on a 7-point Likert-type scale with end points ranging from Strongly Disagree (1) to Strongly Agree. Means and standard deviations along with reliabil- ity measures are listed in Table 1. The internal consistency was satisfactory for all multi-item past behaviors (alpha values ranging from .85 to .86). The correlations among the past behavior items/scales indicated that they could not be reduced into higher order factors so they were used as 11 distinct independent variables.

The third section of the survey included 27 items measuring identification. The Points of Attachment Index (PAI), a scale developed to measure facets of identifica- tion with a sport program (Robinson & Trail, 2005), was used to measure various aspects of attachment to the new football program. The PAI consists of nine categories of attachment (player, team, coach, university, sport, community, athletic department, general sport fan, level of sport), which have shown past reliability and validity related evidence, with alpha scores ranging from .70 to .88 (Robinson & Trail, 2005; Robinson, Trail, & Kwon, 2004; Woo et al., 2009) and Average Variance Extracted

134 Shapiro, Ridinger, and Trail

Table 1 Means (M) and Standard Deviation (SD) Values for the Past Behaviors and Behavioral Intentions

Item/Scale M (SD)

Past Behaviors

Television Consumption (mean score of two items: I watched sports broadcasts on the local TV news for information about the team; I watched TV for news about the team—α= .86)

4.89 (1.84)

Print Media Consumption (I read about the (TEAM NAME) football team in the daily sport pages.) 5.53 (1.79)

Radio Consumption (mean score of four items: I listened to the weekly (TEAM NAME) football coach’s show; I got my information about (TEAM NAME) football from radio stations; I listened to the pregame shows on the radio; I listened to the postgame shows on the radio— α = .85)

3.31 (1.66)

Website (I read about the (TEAM NAME) football team on the (TEAM NAME) website.) 5.26 (1.80)

Past Attendance (How many home (TEAM NAME) football games did you attend this season? 0–7.) 6.10 (1.81)

Tailgating (7-point Likert-type scale, Very negative influence on my attendance (-3) to Very positive influence on my attendance (+3))

5.66 (1.39)

Facebook (I got information about (TEAM NAME) football from Facebook.) 2.51 (1.88)

Twitter (I got information about (TEAM NAME) football through Twitter.) 1.53 (1.28)

Annual Donor Club (6-point scale: How long have you been a member of the (Donor) club? (1) Less than a year, (2) 1–2 years, (3) 3–5 years, (4) 6–10 years, (5) 11–20 years, (6) More than 20 years)

2.58 (1.58)

Web Broadcast (I was aware that I could watch (TEAM NAME) home games online at odusports.com.) 3.76 (2.43)

Merchandise Purchase (fill-in-the-blank item: Please estimate the total dollar amount (if any) that you spent during this current season on (TEAM NAME) football team merchandise and paraphernalia for yourself and others.)

$136.61 (194.87)

Behavioral Intentions Mean (SD)

Support Sponsors of Football team (mean score of three items: When I’m planning to purchase a product, I would be more likely to choose a particular brand if that company sponsors (TEAM NAME) athletics; I will support companies that sponsor (TEAM NAME) athletics when I have a choice between two products; When a company sponsors (TEAM NAME) athletics, I am more likely to purchase their products/services when I have that option (α = .96).

4.86 (1.46)

Purchase Football team Merchandise (Please estimate the total dollar amount (if any) that you intend to spend next year on (TEAM NAME) football team merchandise and paraphernalia for yourself and others.)

$129.03 (192.23)

Attend Football Games (What is the number of (TEAM NAME) football home game(s) that you intend to attend next season?)

6.93 (1.79)

Attend Men’s Basketball Games (I am likely to attend (TEAM NAME) men’s basketball games.) 5.51 (1.58)

Attend Women’s Basketball Games (I am likely to attend (TEAM NAME) women’s basketball games.) 4.15 (1.82)

Note: All items were measured on a 7-point Likert-type scale ranging from Strongly Disagree (1) to Strongly Agree (7) unless otherwise noted.

(AVE) values ranging from .48 to .73 (Robinson & Trail, 2005; Robinson et al., 2004; Woo et al., 2009). The PAI items were measured on a 7-point Likert-type scale and indicated good internal consistency (alpha values rang- ing from .82–.93) and construct reliability (AVE values ranging from .615–.809; Table 2).

Finally, the fourth section of the survey consisted of items measuring future intentions. These items were measured by asking participants how likely they were to attend future football games, attend men’s basketball games, attend women’s basketball games, consume spon- sor (of the football team) products (3-item scale), and purchase football team merchandise in the future. Future intentions were adapted from previous examinations of

identification and future behavior (James & Trail, 2008; Trail et al., 2003, 2005). These items were measured individually using a 7-item Likert-type scale and were retained as single items (except for the sponsored prod- ucts scale which showed satisfactory internal consistency, a = .96), with each used as the dependent variable in the different regression analyses (see Table 1 for means and standard deviations).

Procedure

Questionnaires were administered through an online format. Surveys were sent out one week after the final home game during the inaugural season. Each potential

An Analysis of Multiple Consumption Behaviors 135

participant received an introductory e-mail explain- ing the purpose of the study along with a link to the web-based survey. A follow up e-mail was sent to all nonrespondents ten days later in an effort to increase response rate. In addition, incentives were offered to

respondents who completed the survey. Respondents had the option to enter a drawing to win one of several prizes. The information collected for the drawings was kept separate from survey responses to maximize ano- nymity and confidentiality.

Table 2 Factor Loadings (β), Confidence Intervals (CI), Standard Errors (SE), and Average Variance Explained (AVE) Values for the Points of Attachment Index (PAI)

Factor and Item β CI SE α AVE Identification with the players .91 .773

I am a fan of the individual players on the team .788 .767–.809 .013

I am a big fan of specific players .923 .910–.936 .008

I consider myself a fan of certain players .920 .907–.933 .008

Identification with the team .90 .752

Being a fan of (university) football team is very important to me .819 .800–.838 .012

I am a committed fan of (university) football team .887 .873–.902 .009

I consider myself to be a “real” fan of the (university) football team .893 .879–.907 .009

Identification with the coach .87 .705

I am a big fan of (head coach) .788 .765–.810 .014

I would experience a loss if (head coach) was no longer the coach .850 .831–.868 .011

Being a fan of (head coach) is very important to me .878 .861–.895 .010

Identification with the university .85 .665

I feel connected to numerous aspects of the university .865 .847–.883 .011

I feel that I am part of the university community .833 .813–.854 .012

I support the university as a whole .744 .718–.770 .016

Identification with sport .84 .686

First and foremost I consider myself a football fan. .562 .527–.598 .022

Football is my favorite sport. .958 .944–.972 .008

Of all sports, I prefer football .908 .893–.923 .009

Identification with the community .93 .809

I feel connected to numerous aspects of the community .874 .859–.888 .009

I feel that I am a part of the community .940 .930–.950 .006

I support the community as a whole .883 .869–.897 .008

Identification with the Athletic Department .82 .615

I connect with numerous aspects of (university) athletics .644 .633–.695 .019

I am a fan of all (university) teams .830 .810–.851 .012

Being a fan of all (university) teams is very important to me .861 .843–.880 .011

Identification as a general sport fan .85 .661

I am a sport fan in general .853 .833–.872 .012

I am a fan of lots of different sports .716 .688–.744 .017

Being a sport fan is very important to me .862 .843–880 .011

Identification with level of sport .87 .681

I am a fan of college football regardless of who is playing. .787 .764–.810 .014

I consider myself a fan of college football, and not just one specific team .781 .758–.804 .014

I am a big fan of college football .902 .885–.918 .010

136 Shapiro, Ridinger, and Trail

Results Demographics Of the 1,091 respondents, 32.3% were students, 7.7% were faculty/staff, 38.6% were alumni, 12.6% indicated no formal affiliation and 8.8% indicated other. The mean age of the respondents was 40.5 years with a range from 18 to 81. Approximately 65% were male, and over 60% had a bachelor’s degree or higher. Slightly more than 86% were Caucasian and 59% were married or partnered, with the largest percentage (26%) indicating an average household income in the $100,000-$150,000 range. Approximately 50% of the respondents went to either six or seven of the homes games and 42.6% of the 1,091 indicated that they were annual donors.

Mediated Models

To determine whether points of attachment were potential mediators of the past behavior to behavioral intentions relationship, five separate mediated models were tested. Model testing provides empirical evidence of a mediated relationship, which has been hypothesized in previous identification and sport consumption literature. With 11 IVs (behaviors), 9 mediators (points of attachment), and 5 DVs (behavioral intentions), it was not feasible to test

all 495 potential mediated models. In addition, because the IVs were not highly correlated (Table 3), nor were the DVs (Table 3), it would be improper to create a higher order latent variable to represent each set. However, the points of attachment were correlated to a greater extent, and a third order latent variable model fit well (RMSEA = .080, χ2/df = 2487/1090 = 2.28; Figure 1), thus allow- ing us to use the one latent variable (attachment) as the mediator. We chose past home football attendance as the one IV for each mediated model for two reasons: consistency across models, and use in prior research. We then used each of the five DVs separately in each of the five models.

Model 1 had adequate model fit (RMSEA = .075) and indicated prior home football game attendance sig- nificantly predicted future home football game attendance intentions directly (β = .495) and indirectly through points of attachment (β = .400 from attendance to attachment, and β = .240 from attachment to attendance intentions). The higher direct path coefficient indicated that there was minimal, if any, mediation in this model.

Model 2 had adequate model fit (RMSEA = .076) and indicated that prior home football game attendance significantly predicted future intentions to support spon- sors directly, although the path coefficient was negative (β = -.143), and indirectly through points of attachment

Figure 1 — Third order latent model of points of attachment.

137

Ta b

le 3

B

iv a ri

a te

C o

rr e la

ti o

n T

a b

le f

o r

a ll P

a s t

B e h

a v io

rs , P

o in

ts o

f A

tt a c h

m e n

t, a

n d

B e h

a v io

ra l

In te

n ti

o n

s

M ea

su re

1 2

3 4

5 6

7 8

9 10

11 12

13 14

15 16

17 18

19 20

21 22

23 24

25 1.

G am

es

A tt

en de

d 1.

0

2. R

ad io

.2 57

1. 0

3. T

el ev

is io

n .2

54 .4

63 1.

0

4. W

eb si

te .1

99 .3

40 .3

94 1.

0

5. W

eb

B ro

ad ca

st .1

44 .3

36 .2

23 .3

02 1.

0

6. F

ac eb

oo k

-. 11

9 .1

89 .2

27 .2

10 .1

70 1.

0

7. T

w it

te r

-. 01

5 .2

46 .1

10 .1

26 .1

67 .4

20 1.

0

8. P

ri nt

M ed

ia .3

66 .4

09 .4

88 .2

93 .2

16 -.

01 1

.0 21

1. 0

9. T

ai lg

at in

g .1

54 .1

20 .1

67 .1

58 .1

05 .0

48 .0

20 .1

76 1.

0

10 . M

er ch

an di

se

P ur

ch as

ed .2

67 .2

34 .1

72 .1

94 .0

87 -.

06 3

.0 49

.2 08

.1 43

1. 0

11 . B

oo st

er

C lu

b .2

55 .2

22 .1

00 .0

98 .2

17 -.

18 5

-. 07

0 .2

66 .1

13 .2

43 1.

0

12 . C

oa ch

.3 77

.4 15

.4 76

.3 91

.2 22

.0 48

.0 46

.4 87

.2 34

25 0

.1 90

1. 0

13 . C

om m

un it

y .1

69 .2

32 .2

37 .1

56 .1

36 .1

02 .1

17 .2

72 .2

06 .0

65 .0

99 .3

00 1.

0

14 . L

ev el

.1 76

.2 18

.2 51

.1 29

.1 57

-. 02

1 .0

19 .2

74 .1

61 .1

16 .0

66 .2

53 .1

85 1.

0

15 . P

la ye

r .2

45 .3

87 .3

71 .3

82 .2

29 .1

98 .1

48 .3

32 .1

32 .1

30 .1

05 .4

85 .2

42 .1

99 1.

0

16 . S

po rt

.2 06

.2 26

.2 80

.2 44

.0 92

.0 53

.0 53

.2 27

.1 58

.1 52

-. 06

3 .3

14 .1

54 .5

74 .2

98 1.

0

17 . T

ea m

.4 30

.3 50

.4 86

.4 77

.2 53

.1 00

.0 29

.5 21

.2 84

.2 71

.1 70

.6 68

.2 81

.2 64

.4 84

.4 15

1. 0

18 . U

ni ve

rs it

y .1

77 .2

22 .3

47 .3

89 .2

80 .1

86 .1

06 .2

90 .2

54 .1

32 .1

47 .4

12 .5

75 .2

12 .3

80 .2

83 .5

04 1.

0

19 . G

en F

an .2

09 .2

52 .2

88 .2

47 .1

76 .0

63 .0

50 .3

50 .2

19 .1

04 .0

91 .3

02 .2

76 .5

56 .3

20 .4

10 .3

65 .4

05 1.

0

20 . A

th le

ti c

D ep

t. .2

73 .3

75 .4

28 .4

01 .3

31 .1

56 .1

11 .4

53 .2

94 .1

92 .2

63 .5

03 .4

78 .2

68 .4

37 .2

43 .6

13 .6

55 .5

44 1.

0

21 . S

up po

rt

S po

ns or

s .1

67 .3

55 .3

47 .3

18 21

6 .1

64 .1

10 .3

16 .1

77 .2

01 .1

02 .3

97 .6

30 .2

58 .3

02 .2

64 .4

28 .4

95 .3

40 .4

81 1.

0

22 . M

er ch

an di

se

In te

nt .2

24 .2

25 .1

69 .1

86 .0

97 -.

05 4

.0 57

.1 88

.1 26

.8 65

.2 15

.2 29

.0 83

.1 07

.1 06

.1 39

.2 48

.1 08

.0 93

.1 71

.2 01

1. 0

23 . A

tt en

da nc

e in

te nt

io ns

(F

oo tb

al l)

.6 08

.1 93

.2 49

.2 08

.1 07

-. 03

9 .0

24 .2

74 .0

75 .1

97 .1

53 .3

58 .1

23 .1

53 .2

72 .2

38 .4

42 .1

91 .1

48 .2

58 .2

02 .1

87 1.

0

24 . M

en ’s

B

as ke

tb al

l .1

84 .2

85 .2

42 .1

90 .2

53 .0

89 .0

80 .2

23 .2

09 .0

78 .2

56 .2

13 .2

73 .2

07 .2

09 .0

99 .2

64 .3

75 .4

45 .5

58 .2

70 .0

71 .1

65 1.

0

25 . W

om en

’s

B as

ke tb

al l

.0 40

.1 51

.1 58

.0 91

.1 20

.1 42

.1 15

.1 00

.0 01

-. 02

7 -.

00 6

.0 80

.2 39

.0 18

.1 00

.0 12

.0 97

.2 75

.1 98

.3 24

.2 46

-. 04

1 .0

96 .3

68 1.

0

138 Shapiro, Ridinger, and Trail

(β = .444 from attendance to attachment, and β = .698 from attachment to attendance intentions). The path coefficients in this model indicated that there is at least partial mediation in this instance.

Model 3 also had adequate model fit (RMSEA = .076) and indicated that prior home football game atten- dance significantly predicted future intentions to purchase team related merchandise in the future. However, the direct path coefficient was small (β = .125) indicating that partial mediation existed as the indirect paths through points of attachment (β = .464) and from attachment to attendance intentions (β = .214) were larger.

Model 4 had adequate model fit (RMSEA = .081) and indicated that attachment fully mediated the prior home football game attendance to future Men’s basket- ball game attendance intentions. The path coefficient between home football attendance and Men’s basketball attendance was not significant (β = -.037). However the indirect paths were significant (β = .400 from attendance to attachment, and β = .554 from attachment to Men’s basketball attendance intentions).

Model 5 also had adequate model fit (RMSEA = .079) and indicated that attachment fully mediated the prior home football game attendance on future Women’s basketball game attendance intentions. The path coef- ficient between home football attendance and Women’s basketball attendance intentions was not significant (β = -.093). However, the indirect paths were significant (β = .449 from attendance to attachment, and β = .294 from attachment to Women’s basketball attendance intentions).

Regressions

To test the effects of the past behaviors on the behavioral intentions while controlling for the points of attachment, a hierarchical multiple regression technique was con- ducted. That is, all 9 of the points of attachment were entered as the first block in the regression to examine their effects on behavioral intentions. Then, all 11 past behaviors were entered as the second block. This process helped determine how much additional variance the past behaviors would explain in behavioral intentions after the points of attachment had entered into the equation. Due to the interest in differentiating among the effects of the independent variables (11 past sport behaviors) and the potential mediators (9 points of attachment) on each of the dependent variables (5 sport consumer intentions), a step- wise technique was used within each hierarchical block to test the effects of past behaviors and points of attach- ment on each behavioral intention separately. Stepwise regression is a viable tool when doing “exploratory model building” (Field, 2009, p. 213), which was the case in this instance. The stepwise procedure also determined which variable had the greatest impact on the dependent variable as the first IV entered would subsume the shared variance explained by all IVs. Thus, five stepwise, hierarchical, regressions, one for each of the dependent variables were conducted. A standard for practical meaningfulness was set at a minimum of 6% of the variance explained in the outcome variable (Cohen, 1992).

Multicollinearity was not an issue across any of the regressions. The highest VIF (Variance Inflation Factor) values did not exceed 2.25 (in the Women’s Basketball regression), well below the value of 10 that Field (2009) suggests is worrisome. The tolerance statistic values all exceeded .452 (Women’s Basketball regression), which was considerably higher than the .2 threshold indicated by Field.

Future Football Game Attendance

In the stepwise hierarchical regression, Team Identifica- tion was the only point of attachment to explain variance (19.2%) in Intentions to Attend Football Games (Table 4). After controlling for points of attachment, Past Atten- dance at Football Games explained an additional 20.8% of the variance from the second block, but no other past behavior was significant.

Supporting Sponsorship

In the stepwise hierarchical regression for Support- ing Sponsorship (Table 5), when points of attachment were entered first in a block, University Identification explained 24.4% of the variance in Supporting Sponsors. No other points of attachment explained more than 4.6%. After the points of attachment had entered the equation, Radio Listenership explained an additional 2.4% in Sup- porting Sponsors, which did not meet the preestablished standard of 6%, and no other past behavior explained a significant amount of variance.

Purchasing Merchandise

In the stepwise hierarchical regression for Intentions to Purchase Team Merchandise (Table 6), Team Identifica- tion was the only point of attachment to explain a signifi- cant amount of variance (6.1%) in Intentions to Purchase Team Merchandise. After the points of attachment were entered into the equation, Past Merchandise Purchases explained an additional 68.7% of the variance.

Future Men’s Basketball Game Attendance

In the stepwise hierarchical regression (Table 7), Athletic Department Identification explained 30.2% of the vari- ance in Intentions to Attend Men’s Basketball Games. No other points of attachment explained more than 2.9%. After controlling for points of attachment, no past behavior met the 6% standard.

Future Women’s Basketball Game Attendance

In the stepwise hierarchical regression (Table 8), Athletic Department Identification explained 10.6% of the vari- ance in Intentions to Attend Women’s Basketball. No other points of attachment explained at least 6% of the variance. After controlling for points of attachment, no past behavior explained a minimum of 6% of the variance.

139

Ta b

le 4

M

u lt

ip le

R e g

re s s io

n A

n a ly

s is

P re

d ic

ti n

g I n

te n

t to

A tt

e n

d (

U n

iv e rs

it y ’s

) F

o o

tb a ll G

a m

e s —

S te

p w

is e H

ie ra

rc h

ic a l

R e g

re s s io

n

(P a s t

B e h

a v

io rs

a ft

e r

C o

n tr

o ll in

g f

o r

P o

in ts

o f

A tt

a c h

m e n

t)

M o

d el

U n

st an

d ar

d iz

ed C

o ef

fic ie

n ts

S ta

n d

ar d

iz ed

C

o ef

fic ie

n ts

t p

95 .0

% C

o n

fid en

ce In

te rv

al fo

r B

B S

td . E

rr o

r B

e ta

L o

w er

B o

u n

d U

p p

er B

o u

n d

S te

p 1

F oo

tb al

l A tt

en da

nc e

In te

nt io

ns (C

on st

an t)

2. 84

5 .2

63 10

.8 32

< .0

01 2.

33 0

3. 36

1

ID T

ea m

.6 81

.0 43

.4 38

15 .8

89 <

.0 01

.5 97

.7 65

S te

p 2

(C on

st an

t) 1.

82 8

.2 33

7. 86

1 <

.0 01

1. 37

2 2.

28 4

ID T

ea m

.3 46

.0 41

.2 23

8. 48

1 <

.0 01

.2 66

.4 27

P as

t F

oo tb

al l A

tt en

da nc

e .4

96 .0

26 .5

04 19

.1 99

< .0

01 .4

45 .5

47

N ot

e: R

2 =

.1 92

f or

S te

p 1,

Δ R

2 =

.2 08

f or

S te

p 2

(p <

.0 01

).

Ta b

le 5

M

u lt

ip le

R e g

re s s io

n A

n a ly

s is

P re

d ic

ti n

g S

p o

n s o

r S

u p

p o

rt —

S te

p w

is e H

ie ra

rc h

ic a l R

e g

re s s io

n (

P a s t

B e h

a v io

rs a

ft e r

C o

n tr

o ll

in g

fo

r P

o in

ts o

f A

tt a c h

m e n

t)

M o

d el

U n

st an

d ar

d iz

ed C

o ef

fic ie

n ts

S ta

n d

ar d

iz ed

C

o ef

fic ie

n ts

t p

95 .0

% C

o n

fid en

ce In

te rv

al fo

r B

B S

td . E

rr o

r B

e ta

L o

w er

B o

u n

d U

p p

er B

o u

n d

S te

p 1

S up

po rt

in g

S po

ns or

s (C

on st

an t)

.4 22

.2 03

2. 08

1 .0

38 .0

24 .8

21

ID U

ni ve

rs it

y .3

35 .0

40 .2

81 8.

34 4

< .0

01 .2

56 .4

14

ID C

oa ch

.1 63

.0 30

.1 61

5. 43

2 <

.0 01

.1 04

.2 21

ID A

th le

ti c

D ep

t. .2

23 .0

42 .1

88 5.

25 0

< .0

01 .1

39 .3

06

ID L

ev el

.1 05

.0 25

.1 10

4. 15

1 <

.0 01

.0 55

.1 55

S te

p 2

(C on

st an

t) .4

57 .2

00 2.

29 3

.0 22

.0 66

.8 49

ID U

ni ve

rs it

y .3

53 .0

40 .2

96 8.

91 6

< .0

01 .2

75 .4

31

ID C

oa ch

.1 10

.0 31

.1 09

3. 60

8 <

.0 01

.0 50

.1 70

ID A

th le

ti c

D ep

t. .1

71 .0

42 .1

45 4.

02 6

< .0

01 .0

88 .2

54

ID L

ev el

.0 90

.0 25

.0 94

3. 61

2 <

.0 01

.0 41

.1 39

P as

t R

ad io

C on

su m

pt io

n .1

52 .0

24 .1

73 6.

20 7

< .0

01 .1

04 .1

99

N ot

e: R

2 =

.3 22

f or

S te

p 1,

Δ R

2 =

.0 24

f or

S te

p 2

(p <

.0 01

).

140

Ta b

le 6

M

u lt

ip le

R e g

re s s io

n A

n a ly

s is

P re

d ic

ti n

g I n

te n

t to

P u

rc h

a s e T

e a m

L ic

e n

s e d

M e rc

h a n

d is

e —

S te

p w

is e H

ie ra

rc h

ic a l

R e g

re s s io

n

(P a s t

B e h

a v

io rs

a ft

e r

C o

n tr

o ll in

g f

o r

P o

in ts

o f

A tt

a c h

m e n

t)

M o

d el

U n

st an

d ar

d iz

ed C

o ef

fic ie

n ts

S ta

n d

ar d

iz ed

C

o ef

fic ie

n ts

t p

95 .0

% C

o n

fid en

ce In

te rv

al fo

r B

B S

td . E

rr o

r B

e ta

L o

w er

B o

u n

d U

p p

er B

o u

n d

S te

p 1

In te

nt t

o P

ur ch

as e

T ea

m

L ic

en se

d M

er ch

an di

se (C

on st

an t)

-1 22

.2 06

30 .7

98 -3

.9 68

< .0

01 -1

82 .6

37 -6

1. 77

4

ID T

ea m

41 .6

97 5.

02 7

.2 47

8. 29

4 <

.0 01

31 .8

32 51

.5 61

S te

p 2

(C on

st an

t) .0

10 16

.1 28

.0 01

.9 99

-3 1.

63 6

31 .6

57

ID T

ea m

2. 18

9 2.

70 8

.0 13

.8 08

.4 19

-3 .1

26 7.

50 3

P as

t M

er ch

an di

se P

ur ch

as es

.8 45

.0 16

.8 61

53 .6

11 <

.0 01

.8 14

.8 76

N ot

e: R

2 =

.0 61

f or

S te

p 1,

Δ R

2 =

.6 87

f or

S te

p 2

(p <

.0 01

).

Ta b

le 7

M

u lt

ip le

R e g

re s s io

n A

n a ly

s is

P re

d ic

ti n

g I n

te n

t to

A tt

e n

d (

U n

iv e rs

it y ’s

) M

e n

’s B

a s k e tb

a ll

G a m

e s —

S te

p w

is e H

ie ra

rc h

ic a l

R e g

re s s io

n (

P a s t

B e h

a v io

rs a

ft e r

C o

n tr

o ll in

g f

o r

P o

in ts

o f

A tt

a c h

m e n

t)

M o

d el

U n

st an

d ar

d iz

ed C

o ef

fic ie

n ts

S ta

n d

ar d

iz ed

C

o ef

fic ie

n ts

t p

95 .0

% C

o n

fid en

ce In

te rv

al fo

r B

B S

td . E

rr o

r B

e ta

L o

w er

B o

u n

d U

p p

er B

o u

n d

S te

p 1

M en

’s B

as ke

tb al

l A

tt en

da nc

e In

te nt

io ns

(C on

st an

t) 1.

56 4

.2 34

6. 67

3 <

.0 01

1. 10

4 2.

02 3

ID A

th le

ti c

D ep

t. .6

84 .0

45 .5

32 15

.1 94

< .0

01 .5

95 .7

72

G en

er al

S po

rt F

an .2

57 .0

37 .2

07 7.

03 1

< .0

01 .1

85 .3

28

ID T

ea m

-. 20

4 .0

43 -.

14 8

-4 .6

98 <

.0 01

-. 29

0 -.

11 9

S te

p 2

(C on

st an

t) 1.

65 2

.2 32

7. 12

0 <

.0 01

1. 19

7 2.

10 8

ID A

th le

ti c

D ep

t. .6

35 .0

45 .4

94 13

.9 73

< .0

01 .5

46 .7

24

G en

er al

S po

rt F

an .2

69 .0

36 .2

17 7.

43 5

< .0

01 .1

98 .3

40

ID T

ea m

-. 20

9 .0

43 -.

15 1

-4 .8

58 <

.0 01

-. 29

3 -.

12 4

A nn

ua l

D on

or C

lu b

.1 27

.0 24

.1 33

5. 23

7 <

.0 01

.0 80

.1 75

N ot

e: R

2 =

.3 47

f or

S te

p 1,

Δ R

2 =

.0 16

f or

S te

p 2

(p <

.0 01

).

141

Ta b

le 8

M

u lt

ip le

R e g

re s s io

n A

n a ly

s is

P re

d ic

ti n

g I n

te n

t to

A tt

e n

d (

U n

iv e rs

it y ’s

) W o

m e n

’s B

a s k e tb

a ll

G a m

e s —

S te

p w

is e H

ie ra

rc h

ic a l

R e g

re s s io

n (

P a s t

B e h

a v io

rs a

ft e r

C o

n tr

o ll in

g f

o r

P o

in ts

o f

A tt

a c h

m e n

t)

M o

d el

U n

st an

d ar

d iz

ed C

o ef

fic ie

n ts

S ta

n d

ar d

iz ed

C

o ef

fic ie

n ts

t p

95 .0

% C

o n

fid en

ce In

te rv

al fo

r B

B S

td . E

rr o

r B

e ta

L o

w er

B o

u n

d U

p p

er B

o u

n d

S te

p 1

In te

nt t

o A

tt en

d (U

ni ve

rs it

y’ s)

W

om en

’s B

as ke

tb al

l G

am es

(C on

st an

t) 1.

96 3

.3 02

6. 50

5 <

.0 01

1. 37

1 2.

55 5

ID A

th le

ti c

D ep

t. .5

26 .0

63 .3

52 8.

37 4

< .0

01 .4

03 .6

49

ID T

ea m

-. 30

5 .0

59 -.

19 0

-5 .1

71 <

.0 01

-. 42

1 -.

18 9

ID U

ni ve

rs it

y .2

10 .0

58 .1

40 3.

64 2

< .0

01 .0

97 .3

23

S te

p 2

(C on

st an

t) 2.

38 9

.3 26

7. 33

8 <

.0 01

1. 75

0 3.

02 7

ID A

th le

ti c

D ep

t. .5

51 .0

63 .3

69 8.

75 1

< .0

01 .4

27 .6

74

ID T

ea m

-. 28

4 .0

59 -.

17 7

-4 .8

16 <

.0 01

-. 40

0 -.

16 8

ID U

ni ve

rs it

y .2

22 .0

57 .1

48 3.

86 0

< .0

01 .1

09 .3

35

T ai

lg at

in g

-. 13

3 .0

39 -.

10 1

-3 .3

89 .0

01 -.

21 0

-. 05

6

N ot

e: R

2 =

.1 34

f or

S te

p 1,

Δ R

2 =

.0 09

f or

S te

p 2

(p <

.0 1)

.

142 Shapiro, Ridinger, and Trail

Discussion Based on identity theory and previous literature on identification and spectator consumption behavior, the current study examined the differential effects of past sport consumer behaviors on various college sport consumer intentions that are considered primary rev- enue sources. Furthermore, this investigation helped determine how much variance past behaviors explained in behavioral intentions after controlling for multiple points of attachment. This provided an evaluation of the amount of variance explained by the points of attachment and whether the points of attachment incorporated the variance explained by past behaviors. In addition, this investigation focused on the first year of a new college football program, which supplements the underdeveloped literature examining identification and consumption behaviors with new sport programs.

To examine more specific relationships among past behaviors, points of attachment and behavioral intentions, the current study empirically examined whether points of attachment mediated the relationship between past behavior and behavioral intentions, as a mediated rela- tionship is consistent with previous literature on identity theory. Five different potentially mediating models were tested; one on each behavioral intention variable. Two of the models indicated full mediation (intentions to attend men’s basketball games and intentions to attend women’s basketball games), two of the models indicated partial mediation (intentions to purchase sponsors products and intentions to purchase football team’s merchandise), and one of the models indicated no mediation (intentions to attend future football games). These varied results indi- cate that one should not assume points of attachment will always mediate the relationship between past behaviors and behavioral intentions. In addition, it provides further evidence that the hierarchical stepwise regression analysis used to examine the specific effects of the points of attach- ment and the past behaviors was appropriate.

In general, the findings showed past behavior pre- dicted future intentions. However, the amount of variance explained varied dramatically depending on the type of behavioral intention and the mediating effect of points of attachment. The significant direct effects support the findings of many researchers who have noted that past behaviors predict future behaviors (Ouellette & Wood, 1998; Smith & McSweeney, 2007; Smith, et al., 2008; Watson, Douglas, Berkley, Madapulli, & Zeng, 2009).

The results also showed identification (points of attachment) typically have a significant and meaning- ful mediating influence on behavioral intentions. These results support the theoretical premises in identity theory which suggest as people create roles (e.g., sport fan, fan of the team, etc.) for themselves and value those roles, they are more likely to intend to behave in ways that represent those roles (Stryker & Burke, 2000). This reflects the results regarding involvement in university organizations by students (Serpe & Stryker, 1987) and commitment to religious activities (Stryker & Serpe,

1982); as the importance of a particular role increases, so do concomitant behaviors. These results also support the findings of several sport scholars (Laverie & Arnett, 2000; Madrigal, 1995; Trail et al., 2003, Trail et al., 2005; Wakefield, 1995; Wann & Branscombe, 1993), all of whom found team identification had a relationship with either behavior or behavioral intentions.

Predicting Future Intentions

Listening to football games on the radio predicted sup- porting sponsors, but the amount of variance explained was only 2.4% after points of attachment were included in the model. These findings indicate exposure to the sup- porting partners during the radio broadcasts likely had a minor influence on intentions to buy sponsors’ products in the future. Attachment to the university explained a much larger amount of variance (24.4%) in supporting business sponsors of the team. This finding suggests attachment to the university is more important than radio consump- tion. Interestingly, as fans become more identified with the university as a whole, rather than the team or athletic department specifically, they appear to be more likely to buy products of the supporting sponsors. This finding may be due to the fact that people perhaps incorrectly assume corporate partners sponsor the university as a whole rather than just the football team. For example, beverage sponsors (e.g., Coke, Pepsi) at many schools have agreements that include placing their beverages across campus in all venues. For athletic department sponsors, this is not always the case. In addition, the lack of influence regarding identification with the team or athletic department could be the result of the infancy of the football program. Yet, identification with the uni- versity may have already existed through other means (e.g., current student, alumni, university employee, or member of the local community). Finally, the lack of any past behaviors meaningfully predicting support of sponsors once points of attachment were in the model might be attributed to the lack of inclusion of any past sponsor-specific support in the survey. This should be rectified in the future.

Not surprisingly, past merchandise purchasing behavior explained a large amount of the variance in intentions to buy team related merchandise in the future (69%) even after controlling for points of attachment. Currently, this has not been reported before in the litera- ture; though many researchers have suggested that past behaviors predict future behaviors (Ouellette & Wood, 1998; Smith & McSweeney, 2007; Smith, et al., 2008; Watson et al., 2009). The only point of attachment that explained a meaningful amount of variance in future team merchandise purchases was team identification (6.1%). This was considerably lower than reported in prior research by Kwon, Trail, and Anderson (2006), who found team identification explained 22.5% of the variance in purchasing team merchandise. Once again, this may be due to the lack of opportunity to identify specifically with a brand new football team.

An Analysis of Multiple Consumption Behaviors 143

Team identification predicted approximately 19% of the variance in attending the school’s football games in the future, which was more than what Trail et al. (2006) found (9%). After controlling for points of attachment, past football attendance behavior was the best and only significant predictor of attending the school’s football games in the future (21%). This finding supports the results of Trail et al. (2006) who found a similar amount of variance explained (25%). Taken in concert with the previous data from Trail et al., (2006) this seems to indicate both team identification and past behavior are important in predicting future attendance behavior. Still, there remains an additional 60% of future attendance behavior that is not explained, suggesting additional variables need to be included such as BIRGing (Trail et al., 2005) and constraints (Kim & Trail, 2010).

Attendance at future men’s basketball games was predicted to some extent by belonging to the annual donor club (6.5%; squaring the bivariate correlation). This is not surprising as previous information has shown that there is a cross-over effect between being an avid college football fan and being an avid college basketball fan. The Sports- Business Journal/ESPN Chilton Sports Poll (2000) found that approximate 46% of college football fans are also college men’s basketball fans. However, once the points of attachment were entered into the equation, behaviors were no longer meaningful. Athletic department identi- fication explained 30% of the variance in intentions to attend the men’s basketball games. This makes sense as being attached to the athletic department would make a fan more likely to be a donor.

As anticipated, these relationships were not reflected to the same extent in regard to going to women’s basket- ball games in the future. Belonging to the annual donor club did not predict attendance intentions at women’s games at all. This seems to indicate that many athletic donors support male sports and probably popular male sports only. When the points of attachment were con- trolled for, athletic department identification explained 10% of the attendance intentions. However, this sup- port seemed to come from people who were not annual donors, but still support the athletic department as a whole. A large percentage of women’s game attendance intentions are unexplained by this dataset. However, previous research, which included many additional women’s specific variables, explained more variance (Funk, Ridinger, & Moorman, 2004). The results of the current study indicated there probably is not much cross- over between fans of men’s sports and fans of women’s sports as is noted by the SportBusiness Journal (2000) poll, in which only 5% of college football fans were also women’s college basketball fans.

In sum, these results provide a number of theoretical implications. First, past behaviors do predict behavioral intentions as theorized, but vary by behavior and by inten- tion. Second, points of attachment (role identities) also predict behavioral intentions, which supports prior theory and research. Third, the predictive ability of points of attachment also varies depending on the situation. That is,

different points of attachment predict different behavioral intentions at varying levels. Therefore, the salience of the role identity has an effect on the behavioral intention, again supporting the premises of identity theory.

These results also provide evidence of the impact of previous consumption and identification on future inten- tions within the context of a new sport program. Mullin et al. (2007) stressed the importance direct experiences have on escalating commitment to a program. In addi- tion, James et al. (2002) suggested identification may even exist before the formation of a new team and the direct experience can enhance this identification. The current findings support these assessments as various past behaviors combined with multiple facets of attachment influenced intentions to attend future games (football and other sports), purchase merchandise, and support sponsors. However, the full impact of some facets of attachment such as team, players, and coach may not be known as identification may increase over time. Further investigation is warranted to identify the longitudinal relationship between direct experiences with a new sport program, identification and future consumption intentions.

Practical Implications

Based on the findings from this study, several practical implications can be surmised. The individuals most likely to purchase products from sponsors were those with an attachment to the university. This suggests sponsors may want to consider expanding their partnerships to include multiple university events, not just athletics. In addi- tion, sponsors should emphasize their connection to the university as a whole in advertising messages. The best predictor of future merchandise consumption was past purchase behavior. Thus, team merchandise should be made readily available within the stadium and surround- ing tailgate areas during all home games. The tailgating environment at football games, specifically with a new program, provides considerably more opportunities to sell merchandise compared with other sporting events. In addition, discount coupons designed to drive traffic to the university bookstore and other locations selling team merchandise could be distributed during games. Furthermore, merchandise purchases should be tracked if possible because of the likelihood that people who pur- chase will purchase again. Thus, these individuals should be directly targeted when new merchandise is available or if discounts on merchandise will be forthcoming.

Past attendance behavior and team identification were the only significant indicators of attending games in the future. There was a large demand for tickets during this inaugural football season which resulted in a high percentage of season tickets being sold. The athletic department needs to capitalize on this interest in the football team by securing season ticket commitments from those who purchased them in the past. There was also a high demand for student tickets, but as the novelty of a new team wears off, more effort from the athletic

144 Shapiro, Ridinger, and Trail

department may be needed to fill the student section of the stadium. Once students have experienced the excitement of attending a game, it will influence their identification with the team, which could increase their likelihood of attending games throughout their college years and beyond.

Attachment to the athletic department predicted attendance at both men’s and women’s basketball games at the university, but much more so for the men’s team. The cross-over effect with football and men’s basketball may allow for cross promotions or special incentives for those who support both programs. The small amount of variance explained for attending women’s basketball games suggests that the athletic department may need to take measures to better understand the needs of supporters of women’s sports. Efforts could then be made to offer relevant initiatives for this group and perhaps cultivate more involvement with the annual donor club.

Limitations and Suggestions for Future Research

Some of the behavioral intentions had minimal amounts of variance explained, indicating that additional variables need to be included beyond points of attachment and past behaviors. Some suggestions were noted above. This examination focused on only one football program at one university, which limits the generalizability across programs throughout NCAA Division I football. In addi- tion, this data set may be unique in that the context was a new football team. This information should help other teams that are new, but the specific relationships may vary for existing teams. Another limitation worth noting is the fact that only future intentions were examined, not actual behavior. It is possible that these intentions will not lead to specific behaviors.

Thus, there are multiple areas to expand this line of research. Future studies could include fans of college football teams with long standing traditions. Investi- gations of multiple football programs would enhance the generalizability of these findings. In addition, the behavior and identification of fans of sports other than college football could be examined. It seems reasonable to assume that most professional sport organizations would be interested in a better understanding of factors that can predict sponsor support, merchandise purchases, and attendance. Future research on behavioral and iden- tification factors relevant to fans of women’s sports is also needed. This study was one of the first to examine multiple categories of past behaviors, identification, and future intentions. Past behavior predicted future inten- tions, however the amount of variance explained varied dramatically depending on the type of behavioral inten- tion. In addition, the effects of points of attachment on the past behavior-behavioral intention relationship also varied dramatically. Are these wide variations typical of the types of behaviors measured or did the context of examining fans of a new football program have an impact on results? Further research is needed to gain a

better understanding of the differential effects of various past behaviors on future behavioral intentions and the influence of points of attachment on these relationships.

References Baade, R.A., & Tiehen, L.J. (1990). An analysis of major league

baseball attendance, 1969-1987. Journal of Sport and Social Issues, 14(1), 14–32. doi:10.1177/019372359001400102

Callero, P.L. (1985). Role-identity salience. Social Psychology Quarterly, 48(3), 203–215. doi:10.2307/3033681

Cohen, J. (1992). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Lawrence Erlbaum Associates.

Demnert, H.G. (1974). The economics of professional team sports. Lexington, MA: D.C. Heath & Co., Field, A. (2009). Discovering statistics using SPSS. London: SAGE Publications, Ltd.

Fulks, D.L. (2011). Revenues & Expenses: 2004-2010 NCAA Division I intercollegiate athletic programs report. India- napolis, IN: National Collegiate Athletic Association.

Funk, D.C., Ridinger, L.R., & Moorman, A.M. (2004). Explor- ing origins of involvement: Understanding the relation- ship between consumer motives and involvement with professional sports teams. Leisure Sciences, 26, 35–61. doi:10.1080/01490400490272440

Greenstein, T., & Marcum, J. (1981). Factors affecting atten- dance at major league baseball: A within-season analyses. Sociology of Sport Journal, 2(4), 314–322.

Hansen, H., & Gauthier, R. (1989). Factors affecting attendance at professional sporting events. Journal of Sport Manage- ment, 3(1), 115–132.

James, J.D., Kolbe, R.H., & Trail, G.T. (2002). Psychological connection to a new sports team: Building or maintaining the consumer base. Sport Marketing Quarterly, 11(4), 215–225.

James, J.D., & Trail, G.T. (2008). The relevance of team identification to sport consumption behavior intentions. International Journal of Sport Management, 9, 427–440.

Kim, Y.K., & Trail, G.T. (2010). Constraints and motivators: A new model to explain sport consumer behavior. Journal of Sport Management, 24(2), 190–210.

Kwon, H.H., Trail, G.T., & Anderson, D. (2006). Points of attachment (identification) and licensed merchandise con- sumption among American college students. International Journal of Sport Management, 7, 347–360.

Laverie, D.A., & Arnett, D.B. (2000). Factors affecting fan attendance: The influence of identity salience and satisfac- tion. Journal of Leisure Research, 32(2), 225–246.

Lock, D., Darcy, S., & Taylor, T. (2009). Starting with a clean slate: An analysis of member identification with a new sports team. Sport Management Review, 12, 15–25. doi:10.1016/j.smr.2008.09.001

Lock, D., Taylor, T., & Darcy, S. (2011). In the absence of achievement: The formation of new team identification. European Sport Management Quarterly, 11(2), 171–192. doi:10.1080/16184742.2011.559135

Madrigal, R. (1995). Cognitive and affective determinants of fan satisfaction with sporting event attendance. Journal of Leisure Research, 27, 205–227.

McGuire, K. (2011, July 4). 25 college football programs to launch by 2014. Retrieved from http://www.examiner. com/college-football-in-national/25-college-football- programs-to-launch-by-2014

An Analysis of Multiple Consumption Behaviors 145

Mullin, B.J., Hardy, S., & Sutton, W.A. (2007). Sport Marketing (3rd ed.). Champaign, IL: Human Kinetics.

Noll, R.G. (1974). Attendance and price setting. In R.G. Noll (Ed.), Government and the sport business. Washington: The Brookings Institution.

Ouellette, J.A., & Wood, W. (1998). Habit and intention in everyday life: The multiple processes by which past behavior predicts future behavior. Psychological Bulletin, 124(1), 54–74. doi:10.1037/0033-2909.124.1.54

Robinson, M.J., & Trail, G.T. (2005). Relationships among spectator gender, motives, points of attachment, and sport preference. Journal of Sport Management, 19, 58–80.

Robinson, M., Trail, G.T., & Kwon, H.H. (2004). Motives and points of attachment of professional golf spectators. Sport Management Review, 7, 167–192. doi:10.1016/S1441- 3523(04)70049-2

Sanders, L. (2010). At what price football? The Chronicle of Higher Education. Retrieved from http://chronicle.com/ article/article-content/64132/

Serpe, R.T., & Stryker, S. (1987). The construction of self and reconstruction of social relationships. In E. Lawler & B. Markovsky (Eds.), Advances in group processes (pp. 41–66). Greenwich, CT: JAI.

Smith, J.R., & McSweeney, A. (2007). Charitable giving: The effectiveness of a revised theory of planned behaviour model in predicting donating intentions and behaviour. Journal of Community & Applied Social Psychology, 17, 363–386. doi:10.1002/casp.906

Smith, J.R., Terry, D.J., Manstead, A.R., Louis, W.R., Kotter- man, D., & Wolfs, J. (2008). The attitude-behavior relation- ship in consumer conduct: The role of norms, past behav- ior, and self-identity. The Journal of Social Psychology, 148(3), 311–333. doi:10.3200/SOCP.148.3.311-334

SportsBusiness Journal/ESPN Chilton Sports Poll. (2000). Street & Smith’s SportsBusiness Journal, 7(1), 1.

Stryker, S. (1968). Identity salience and role performance: The relevance of symbolic interaction theory for family research. Journal of Marriage and the Family, 30, 558– 564. doi:10.2307/349494

Stryker, S. (1980). Symbolic interactionism: A socio-structural version. Menlo Park, CA: Benjamin/Cummings.

Stryker, S., & Burke, P.J. (2000). The past, present, and future of identity theory. Social Psychology Quarterly, 63(4), 284–297. doi:10.2307/2695840

Stryker, S., & Serpe, R.T. (1982). Commitment, identity salience, and role behavior: A theory and research example. In W. Ickes & E.S. Knowles (Eds.), Personality, roles, and social behavior (pp. 199–218). New York, NY: Springer- Verlag.

Tierney, M. (2009, April 20). Georgia State hoping football builds community in football town. The New York Times. Retrieved from http://www.nytimes.com

Trail, G.T., Anderson, D.F., & Fink, J.S. (2000). A theoretical model of sport spectator consumption behavior. Inter- national Journal of Sport Management, 1, 154–180.

Trail, G.T., Anderson, D.F., & Fink, J.S. (2005). Consumer satisfaction and identity theory: a model of sport spectator conative loyalty. Sport Marketing Quarterly, 14, 98–111.

Trail, G.T., Anderson, D.F., & Lee, D. (2006). Determinants of attendance: The predictive value of team identifica- tion, past attendance, and attendance intentions. Pre- sentation at the Sport Marketing Association Annual Conference.

Trail, G.T., Fink, J.S., & Anderson, D.F. (2003). Sport specta- tor consumption behavior. Sport Marketing Quarterly, 12, 8–17.

Trail, G.T., Robinson, M.J., Dick, R.J., & Gillentine, A.J. (2003). Motives and points of attachment: Fans versus spectators in intercollegiate athletics. Sport Marketing Quarterly, 12(4), 217–227.

Wakefield, K.L. (1995). The pervasive effects of social influence on sporting event attendance. Journal of Sport and Social Issues, 19, 335–351. doi:10.1177/019372395019004002

Wann, D.L., & Branscombe, N.R. (1993). Sports fans: Measur- ing degree of identification with their team. International Journal of Sport Psychology, 24, 1–17.

Watson, G.W., Douglas, T., Berkley, R., Madapulli, R., & Zeng, Y. (2009). Are past normative behaviors predictive of future behavioral intentions? Ethics & Behavior, 19(5), 414–431. doi:10.1080/10508420903035414

Whitney, J.D. (1988). Winning games versus winning championships: The economics of fan interest and team performance. Economic Inquiry, 26(4), 703–724. doi:10.1111/j.1465-7295.1988.tb01524.x

Woo, B., Trail, G.T., Kwon, H.H., & Anderson, D. (2009). Testing models of motives and points of attachment among spectators in college football. Sport Marketing Quarterly, 18, 38–53.

Copyright of Journal of Sport Management is the property of Human Kinetics Publishers, Inc. and its content

may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express

written permission. However, users may print, download, or email articles for individual use.